HP 441877-00B Software And Configuration Manual
HP Net w or k Adapt er So f t w ar e and C onf i gur ati on Guide Part Number 441877-00B March 2007 (Second Edition)
© Copyright 2005-2007 Hewlett-Packard Development Co mpany, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Noth ing herein should be construed as constituting an a dditional w arranty. HP shall not be liable for technical or editor ial errors or omissions contained herein. Confidential computer software. Va lid license from HP required for possession, use or copying. Consistent with FAR 12.211 and 12.212 , Commercial Computer Software, Co mputer Software Documentation, and Technical Da ta for Commercial Items are li censed to the U.S. Government under vendorâs standard commercial license. Microsoft, Windows, and Windows NT are U.S. re gistered trademarks of Microsoft Corporation. Windows Server 2003 is a trademark of Microsoft Corporation. Intel, Pentium, and Itanium are trademarks or registered trademarks of Intel Corporatio n or its subsidiaries in the United States and other countries. UNIX is a registered trademark of The Open Group. Audience assumptions This document is for the person who insta lls, admini sters, and troubleshoots servers and storage syst ems. HP assumes you are qualified in the serv icing of co mputer equipment and trained in recognizing hazards in products with hazardous energy levels.
Contents 3 C on t en ts Overvi ew ..................................................................................................................................... 5 Supported adapters and operating systems....................................................................................... ............ 5 Adapter driver and so ftware info rmation ...................................................................................................... 7 PCI Hot Plug ................................................................................................................... .......................... 9 Duplex support................................................................................................................. ....................... 10 Installa tion ................................................................................................................... .............. 13 Download latest drivers, fi rmware, and do cumentation ........................................................................... ..... 13 Windows Server 2003 and Wi ndows Server 2003 x64 .............................................................................. 13 Windows 2000 an d Windows NT .................................................................................................... ........ 15 Linux.......................................................................................................................... ............................ 20 SCO OpenServer and Unix Ware .................................................................................................... .......... 21 Solaris ........................................................................................................................ ........................... 23 NetWare ........................................................................................................................ ....................... 28 Firmware upgr ade util ity ....................................................................................................... ....... 48 Windows ut ility ................................................................................................................ ....................... 48 c-Class BladeSys tem utility .................................................................................................... .................... 54 Linux ut ility .................................................................................................................. ........................... 54 DOS util ity .................................................................................................................... .......................... 59 Adapter conf igurat ions ......................................................................................................... ....... 62 HP NC-Series Broa dcom adapters ................................................................................................. ............ 62 Keywords for Q5 7 NDIS2 driver .................................................................................................. ... 62 Keywords for B 06 NDIS2 driver .................................................................................................. .... 63 PXE configuration: HP NC-series Broa dcom and Multifun ction adapters ................................................ 64 HP NC-Series Multif unction a dapters............................................................................................ .............. 66 Accelerated iS CSI devices ...................................................................................................... ........ 66 iSCSI boot configuration: Multifunctio n adapters ............................................................................... 67 PXE configuration: HP NC-series Broa dcom and Multifun ction adapters ................................................ 67 HP NC-Series In tel adapters .................................................................................................... .................. 69 Boot Agent ..................................................................................................................... .............. 69 Keywords for N100 NDIS2 driver ................................................................................................. .. 81 Keywords for N10 00 NDIS2 driver ................................................................................................ .8 7 PXE configuration for HP NC-Series Inte l adapters.............................................................................. 89 Windows ........................................................................................................................ ....................... 91 Modifying adapter proper ties using the NCU .................................................................................... 9 1 Modifying HP NC-Series Broadcom and Multifunction adapt er properties ............................................. 98 Modifying HP NC-Series Inte l adapter prop erties ............................................................................. 103 Modifying NC31xx Fast Ethern et adapter pr operties ........................................................................ 106 Modifying adapter properti es in Wind ows NT ................................................................................ 107 Novell Netware ................................................................................................................. ................... 110 Advanced Server Prog ram for NetW are ......................................................................................... 110 NetWare Client 32 driv er installation fo r HP adapters...................................................................... 120 Solaris ................................................................................................................................................. 122 Team prop erties ................................................................................................................ ........ 126 Adapter te aming ................................................................................................................................... 126
Contents 4 Multifunction gigabit se rver adapter teaming ................................................................................... ......... 129 Windows 2000 and Wi ndows NT4 teaming ........................................................................................... .1 3 0 NetWare te aming ................................................................................................................................. 139 VLAN prop erties ................................................................................................................ ....... 150 VLAN info rmation .................................................................................................................................. 150 ProLiant E ssentia ls ............................................................................................................ ......... 151 Networking pack installation ................................................................................................... ............... 151 Intelligent Networking Pack-Windows Edition.................................................................................... ........ 151 Intelligent Networking Pack-Linux Edition ...................................................................................... ............ 152 Accelerated iSCSI for stand-alon e and c-Class BladeSystem Multif unction server adapters .............................. 153 Accelerated iSCSI for embedded HP Multifunction serv er adapters .............................................................. 15 3 Diagnost ics .................................................................................................................... .......... 154 PCI shared interrupts .............................................................................................................................. 154 HP NC-Series Intel ad apter diagnostics ......................................................................................... ........... 154 HP NC-Series Broadcom adapter diagnostics ...................................................................................... ...... 156 HP NC-Series Broadcom Multif unction adapter diagnostics ........................................................................ .1 6 7 Technical support .............................................................................................................. ........ 172 Before you contac t HP ............................................................................................................................ 172 HP contact in format ion ........................................................................................................................... 172 Free automated customer support services ....................................................................................... ......... 172 Acronyms and abbreviati ons ..................................................................................................... . 174 Index .......................................................................................................................... ............. 176
Overview 5 Ov er v ie w In this section Supported adapters and operating systems ....................................................................................... .......... 5 Adapter driver and so ftware info rmation ........................................................................................ ............. 7 PCI Hot Plug ................................................................................................................... ......................... 9 Duplex support................................................................................................................. ...................... 10 Supported adapters and operating systems Supported adapters The Network Configuration Utility (NCU) supports th e following HP Fast Ethernet and Gigabit Server Adapters. Standup adapters: ⢠HP NC110T PCI Express Gigabit Server Adapter ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC310F PCI-X Gigabit Server Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC340T PCI-X Quad Port Gigabit Server Adapter ⢠HP NC360T PCI Express Dual Port Gigabit Server Ad apter ⢠HP NC364T PCI Express Quad Port Gigabit Server Adapter ⢠HP NC370T PCI-X Multifun ction Gigabit Server Adapter ⢠HP NC370F PCI-X Multifun ctio n Gigabit Server Adapter ⢠HP NC373F PCI Express Multifunction Gigabit Server Adapter ⢠HP NC373T PCI Express Multifunction Gigabit Serv er Adapter ⢠HP NC380T PCI Express Dual Port Multifunction Giga bit Server Adapter ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC3123 Fast Ethernet Server Adapter * ⢠HP NC3132 Fast Ethernet Upgrade Module (D ual 10/100) for the NC3134 Server Adapter * ⢠HP NC3133 100FX Upgrade Module f or the NC3134 Server Adapter * ⢠HP NC3134 Fast Ethernet Server Adapter (Dual Ba se 10/100) * ⢠HP NC3135 Fast Ethernet Upgrade Module (Dual 10 /100) for the NC3134 Server Adapter * ⢠HP NC6132 1000 SX Gigabit Upgrade Module f or the NC3134 Server Adapter* ⢠HP NC6133 1000 LX Gigabit Upgrade Modu le for the NC3134 Server Adapter * ⢠HP NC6134 Gigabit Server Adapter (PCI 1000 SX)
Overview 6 ⢠HP NC6136 Gigabit Server Adapter (PCI 1000 SX) ⢠HP NC6170 Dual Port PCI- X Gigabit Server Adapter ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7131 Gigabit Ethernet Server Adapter ⢠HP NC7132 Gigabit Ethernet Upgrade Module for the NC31 34 Server Adapter * ⢠HP NC7170 Dual Port PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC320i PCI Express Gigabit Server Adapter ⢠HP NC324i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC325i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC326i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC370i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC371i PCI-X Multifunction Gigabit S erver Adapter ⢠HP Embedded NC373i Multifunct ion Gigabit Server Adapter ⢠HP NC3161 Fast Ethernet Server Adapter * ⢠HP NC3163 Fast Ethernet Server Adapter * ⢠HP NC7760 PCI-X Gigabit Server Adapter ⢠HP NC7761 PCI-X Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gigabit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter Mezzanine adapters: ⢠HP NC320m PCIe Gigabit Server Adapter ⢠HP NC325m PCI Express Quad Port 1Gb Server Ad apter for c-Class BladeSystem ⢠HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem ⢠HP NC373m PCI Express Dual Port Multifunction Giga bit Server Adapter for c-Class BladeSyst em ⢠HP NC374m PCI Express Dual Port Multifunction Giga bit Server Adapter * Not supported in Windows Server 2003 x64 Supported operating systems These server adapters are supported on the followin g operating systems. Microsoft ⢠Windows Server 2003 x64 Editions ⢠Windows Server 2003
Overview 7 ⢠Windows® 2000 (NDIS5) ⢠Windows NT® 4.0 (NDIS4) (only supported on adapters that shipped befor e January 1, 2003) ⢠DOS NDIS2* NetWare ⢠NetWare 6.5 Server ⢠NetWare 6.6 Server ⢠NetWare Open Enterprise Server Linux 32 ⢠Red Hat Enterprise Linux 5 ⢠Red Hat Enterprise Linux 4 ⢠Red Hat Enterprise Linux 3 ⢠SUSE Linux Enterprise Server 10 Service ⢠SUSE Linux Enterprise Server 9 Service Linux 64 ⢠Red Hat Enterprise Linux 5 for AMD64 and Intel EM64T ⢠Red Hat Enterprise Linux 4 for AMD64 and Intel EM64T ⢠Red Hat Enterprise Linux 3 for AMD 64 and Intel EM64T ⢠SUSE Linux Enterprise Server 10 for AMD64 and Intel EM64T ⢠SUSE Linux Enterprise Server 9 for AMD64 and Intel EM64T UNIX ⢠Unixware 7.x ⢠OpenUnix 8.x ⢠Open Server 5.x/6.x Solaris ⢠Solaris 10 ⢠Solaris 9 * Unattended Install Only. Also, DOS ND IS2 is not supported for NC3133 adapters. Adapter driver and software information To obtain driver and software information for Micro soft® Windows Server 2003 x64, Microsoft® Windows Server 2003, Microsoft ® Windows® 2000, Microsoft® Windows NT® (4.0), Novell NetWare, SCO OpenServer, UnixWare, Linux, and Solaris: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system.
Overview 8 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. Drivers Drivers are found in the following folder s: ⢠Microsoft® Windows Ser ver 2003 \WIN2003 ⢠Microsoft® Windows Ser ver 2003 x64 \WIN2003x64 ⢠Microsoft® Windows® 2 000 \WIN2000 ⢠Microsoft® Windows NT® 4.0 \NT4 ⢠Novell NetWare \NETWARE\SERVER ⢠Linux \LINUX ⢠UnixWare 7.x and OpenServer 6.x \UNIX\UNIXWARE7_OPENSERVER6 ⢠Solaris 9 and Solaris 10 \SOLARIS ⢠Unattended Install \DOS Supplemental applications in \APPS ⢠Boot agent ⢠Diagnostic tests including MFDIAG.exe, N100X.ex e and Q57.exe ⢠Online firmware upgrade utility ⢠Multifunction firmware upgrade utility ⢠c-Class BladeSystem firmware upgrade utility ⢠Linux firmware upgrade utility Device-specific information See \docs\files.txt for a descriptive li st of all files in the SoftPaq. Release notes and other documentation in \docs ⢠HP Accelerated iSCSI for Multifunct ion Network Adapters (AiSCSIUG.pdf) ⢠HP Network Adapter License Util ity for Windows (NALICNSE.pdf) ⢠HP Network Adapter Scripting Utility (NICSCRPT.pdf) ⢠HP Network Configuration Agent (ncuasr.pdf) ⢠HP ProLiant Essentials Intelligent Networking PackâLinux Edition (PEINPL.pdf) ⢠HP ProLiant Essentials Intelligent netw orking PackâWindows Edition (PEINPW.pdf) ⢠HP Virus Throttle for Windows License Utility (VTLICNS E.pdf) ⢠HP NC510x PCIe 10 Gb Software Configuration Guide (NC5 10x PCIe 10 Gb Software Guide.pdf) ⢠Network adapter user guides (\docs\hw) ⢠Release notes (Relnotes.txt) ⢠SetLACState Utility (SetLAC.pdf)
Overview 9 Linux documentation ⢠Linux readme.txt file is located in the \LINUX directory ⢠HP iSCSI Boot User Guide for Linux in the \docs directory PCI Hot Plug PCI Hot Plug is an industry-standard solution that provides increased system availability and non-stop serviceability in business-cri tical computing environm ents. PCI Hot Plug enables PCI networking a dapters to be replaced or added to a server without taking the system down. HP pioneered this technology and has worked closely with the PCI Special Interest Group (SIG) and several key industry partners to accelerate its adoption as an industry standard. PC I Hot Plug technology also has a broad range of industry support from leading operating system suppliers incl uding Microsoft® a nd Novell. HP server adapters support PCI Hot Plug technology. To take advantage of this solution, you must use the adapter in an HP or third-party system that suppor ts PCI Hot Plug. To implement PCI Hot Plug, use the configuration instructions that come with your sy stem. If you use an HP system that supports PCI Hot Plug, refer to the documentation for configuring and insta lling PCI Hot Plug devices in the HP SmartStart application, which ships with HP systems. If you use a third-party system that supports PCI Hot Plug, refer to the documentat ion that came with your system or contact technical support for the vendor. Support for enabling PCI Hot Plug Different operating syst ems support PCI Hot Plug in diff erent ways. The following list describ es how different operating systems enable PCI Hot Plug devices. ⢠Windows® 2000, Windows Server 2003, and Windows Serv er 2003 x64 o Can replace like item for like item o Can add a new adapter to an empty slot o Can add an option module to an existing NC3134 adapter ⢠Windows NT® 4.0 o Can replace like item for like item o Cannot add a new adapter to an empty slot ⢠NetWare 6.x o Can replace like item for like item o Can remove an adapter o Can add a new adapter to an empty slot o Can add an option module to an existing NC3134 adapter (This feature is not supported on an NC3134 adapter that has a pers istent driver installed.) ⢠UnixWare 7.x and OpenServer 6.x Requires hardware that implement s PCI Hot Plug fa ilover and a DDI 8 MDI driver that implements the CFG_SUSPEND and CFG_RESUME subfunctions to th e entry point routine. This lets the system administrator suspend the driver, power down the slot, hot-swap or hot-replace the adapter, power up the slot, and resume the driver, all with out bringing the machine down. For further information about PCI Hot Plug support, refer to your UNIX documentation. ⢠SCO OpenServer 5.x
Overview 10 No support ⢠Linux No support Upgrading the adapter with a persistent driver To upgrade the adapter with a persistent driver: 1. Manually remove (unload) the driver. 2. Power down the slot and remove the adapter. 3. Upgrade the adapter (add the option module). 4. Reinstall the upgraded adapter. 5. Power up the slot to let NetWare automa tically detect the (upgrad ed) adapter. Duplex support Duplexing is a performance option that allows you to choose how the adapter sends and receives packets over the network. Duplex options include : ⢠Auto. (Auto-negotiation) The adapter negotiates with the full-duple x switch to send and receive packets at the highest rate. The switch must suppor t auto-negotiation. If the switch does not support auto-negotiation, the adapter defaults to half-duplex. NOTE: An exception is the HP NC3133 adap ter, which defaults to full-duplex. ⢠Full-duplex. (Requires f ull-duplex switch) The adapter se nds and receives packets simultaneously by using a different wire to perform each function. This feature can improve adapter performance by eliminating collisions. Set duplex mod e to full-duplex ONLY if you have a hub/switch that supports full-duplex. ⢠Half-duplex. The adapter performs one operation at a ti me (because send and receive lines are shared); it either sends or receives. Configuring for full-duplex If you have an auto-negotiating switch, the adapter au tomatically matches the duplex setting at the switch. If you do not have an auto-negotia ting switch, you must manually configur e the adapter for full-duplex support. Configuration is specific to the driver yo u are loading for your network operating system (NOS). To set up the duplex mode, refer to the section below that corresponds to your operating system. Note th at performance may suffer if your switch is not full-dup lex and you configure the adapter to full-duplex. Leave the adapter on half-duplex if you are not sure wh at type of switch to which you are connected. Windows NT® 4.0 To set the duplex options: 1. Double-click the HP Network icon in the Control Panel. Or, right-click on Network Neighborho od and select Properties. Then select the Adapters tab. Double-click on an adapter or team to launch the CPQNTAC utility. 2. Select an adapter from the list and click Properties.
Overview 11 3. Set Speed/Duplex on the adapter to Auto/Auto, 10/Half, 10/Full, 1 00/Half, 100/Full, or 1000/Full. 4. Click OK until prompted to restart. 5. Restart Windows NT® 4.0. Windows® 2000, Windows Server 2 003, and Windows Server 2003 x64 For new installations, Windows® 20 00 automatically installs the adapter drivers. Exception: For the NC150T, NC310F, NC32xx, NC37xx, NC340x, NC1020, NC61 36, NC6170, NC71xx and NC77xx adapters, you must install the appropriat e driver components located in the \WIN2000 directory in the HP SoftPaq. Complete the fo llowing steps to download the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. NOTE: If you need to update existing Windows® 2000 drivers, refer to the instructions in Windows 2000 installation notes (" Windows 2000 and Windows NT " on page 15 ). To set the duplex options: 1. Launch the Network Configuration Utility. 2. Select an adapter from the list in the NCU Main window and click Properties . The Adapter Properties window appears. 3. On the Settings tab, set the Speed/Duplex Setting for the adapter to Auto/Auto, 10/Half, 10/Full, 100/Half, 100/Full, or 1000/Full . 4. Click OK twice to accept the configuration changes. NetWare Server To set the duplex options, use the following syntax for FORCEDUPLE X and SPEED. NOTE: The NC340T, NC1020, NC71xx, and NC77xx adapters support full-duplex and half- duplex. The fiber Gigabit adapters support fu ll-duplex mode only and cannot be changed. ⢠Syntax: FORCEDUPLEX=n Where n = 0âauto-negotiation (You must have an auto-negot iating switch/hub (an Nway* swit ch) to get full- duplex support with the FORCEDUPLEX par ameter set to 0 (auto-negotiation). 1âhalf-duplex 2âfull-duplex NOTE: To set the adapter to half- or full-duplex, yo u must set the SPEED parameter to either 10 or 100. Default = auto-n egotiate
Overview 12 Examples: o 100 Mbps full-duplex: FORCEDUPLEX=2 SPEED=100 o 10 Mbps full-duplex: FORC EDUPLEX=2 SPEED=10 ⢠Syntax: SPEED= n (n = 10 or 100 or 1000) Specifies the speed the driver uses. If you do not use this parameter, the driver aut omatically detects the network speed. If unable to detect the network sp eed, the driver defaults to 10 Mbps. If you use this parameter, the driver operates at the specified speed instead of auto-d etecting network speed. NOTE: The NC3133 adapter supports 100 Mbps on ly. It does not default to 10 Mbps. You must set the SPEED parameter to either 10 or 100 if you want to set the FORCEDUPLEX parameter to either half or full. The SPEED parameter is for N100.LAN driver and N1000.LAN. You cannot change the speed for the Gigabit (fiber) ad apters. You can, however, change the speed for the NC7131 and NC7132 Gigabit (copper) adapters to 10 Mbps or 100 Mbps. Default = The adapter automatically senses speed. Other NDIS 2.x-based clients To set the duplex options in other NDIS 2.x-base d clients, edit the PROTOCOL.INI file. Add the FORCEDUPLEX and SPEED parameters to the N100 sect ion, or the SPEEDUPLEX parameter to the N1000 section. For information on these parameters, refer to K eywords for N100 NDIS2 driver (on page 81 ) , Keywords for N1000 NDIS2 driver (on p age 87 ), or Keywords for Q57 NDIS2 driver (on page 62 ).
Installation 13 Inst allati on In this section Download latest drivers, fi rmware, and documentation ........................................................................... .... 13 Windows Server 2003 and Wi ndows Server 2003 x64 ............................................................................ 13 Windows 2000 an d Windows NT .................................................................................................... ....... 15 Linux .......................................................................................................................... ........................... 20 SCO OpenServer and UnixWare .................................................................................................... ......... 21 Solaris ........................................................................................................................ .......................... 23 NetWare ........................................................................................................................ ...................... 28 Download latest drivers, firmware, and documentation HP updates networ king softwar e freq uently to include new functionality and features. For the latest driver, firmware, and documentation updates go to the HP website ( http://h18004.www1.hp.com/products/ servers/networking/ index.html ). Windows Server 2003 and Windows Server 2003 x64 The following provides inform ation about Windows Server 200 3 and Windows Server 2003 x64 support for HP NC-series network adapters. Windows Server 2003 The following Windows Server 2003 network adap ter drivers are used in this release: ⢠N100325.sys for NC31xx Fast Ethernet server adapters ⢠N1000325.sys for NC310xx, NC340x, NC61xx, an d NC71xx server adapters ⢠Q57XP32.sys for NC150x, NC32xx, NC10xx, NC 67xx, and NC77xx server adapters ⢠N1E5132.sys for NC110T, NC360x, NC364x server adapters ⢠BXND51X.sys for all Multifunction server adapters ⢠NXP2NIC.sys for NC510x 10 Gb server adapters Components for these drivers are located in the \W IN2003 directory in the HP SoftPaq. Complete the following steps to downloa d the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system.
Installation 14 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. Windows Server 2003 x64 IMPORTANT: NC31xx, NC6132, NC6133, and NC7132 adapters are not supported in Windows Server 2003 x64 . The following Windows Serv er 2003 x64 network adapter drivers are used in this relea se: ⢠N1G5132E.sys for NC310F, NC340T, NC61xx, and NC71xx server adapters ⢠Q57AMD64.s ys for NC150T, NC32x, NC1020, NC6 7xx, and NC77xx server adapters ⢠BXVBDA.sys for Multifunction server adapters ⢠N1E5132E.sy s for NC110T, NC360T, NC364x server adapters ⢠NXP2NIC.sys for NC510x 10 Gb server adapters Components for these drivers are located in the \WIN2003x 64 directory in the HP SoftPaq. Complete the following steps to do wnload the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. Installing network drivers (new installations) When you install the adapter in the server for the first ti me: 1. Connect the cable from the adapter to your network. 2. Provide power to your server. 3. Start Windows Serve r 2003 or Windows Se rver 2003 x64 . o For the NC150T, NC310F, NC32xx, NC340x, NC3 7xx, NC380x, NC1020, NC6136, NC6170, NC6770, NC71xx, and NC77xx server ad apters, you must install the appropriate driver component located in the \WIN2003 or \WIN2003x 64 directory in the HP SoftPaq file. o For all other adapters, Windows® automatically de tects and installs a driver for the adapter from its own library of drivers. Updating network drivers To update the network driver, follow these instructions: Locate the driver on the CD, diskette, or directory on your hard driv e where you copied it from an updated source. See Technical Support for a list of online sources for software and driver updates.
Installation 15 1. Run the appropriate Driver Component Package (lis ted as CPxxxxxx.exe) in a directory on your hard drive. The component is a self-extracting execut able with a file name based on the component number. 2. Click the Install button to proceed with the installation. 3. After the drivers have been installed, you may delete the Component Package that you copied. Removing the driver software To remove the driver software: 1. Before physically removing an adapter from your system, first remove the adapter driver software. 2. Start Windows Server 2003 or Windows Server 2003 x64 and log in. You must have Netwo rk Administrator privileges to re move the driver software. 3. Open the Control Panel and double-click the System icon. In the System Properties dialog box, select the Hardware tab. 4. Click Device Manager . The Device Manager window appears. 5. Click the plus ( ) sign to expand the Network adapter s. All network adapters appear. 6. Right-click the adapter to be removed and choose Uninstall . NOTE: Not all driver files are removed as part of this procedure. The driver and adapter can be removed via Hot Plug, if supported. Windows 2000 and Windows NT The following provides inform ation about Windows 2000 support for HP NC-series network adapters. The following Windows 2000 netw ork adapter drivers are used in this release: ⢠N100NT5.sys for NC31xx Fast Ethernet serv er adapters ⢠N1000NT5.sys for NC310x, NC340x, NC61xx, and NC71xx Gigabit Ethernet server adapters ⢠Q57W2K.sys for NC150x, NC32xx, NC10xx, NC67xx, and NC77xx Gigabit Ethernet server adapters ⢠BXVBDX.sys for Multifunction Gigabit server adapters ⢠N1E5032.sys for NC360x and NC364T Gigabit Ether net server adapters Components for these drivers are located in the \WIN2000 dir ectory in the HP SoftPaq file. Installing network drivers (new installations) When you install the adapter in the server for the first ti me: 1. Connect the cable from the adapter to your network. 2. Provide power to your server. 3. Start Windows Serve r 2003 or Windows Se rver 2003 x64 . o For the NC150T, NC310F, NC32xx, NC340x, NC3 7xx, NC380x, NC1020, NC6136, NC6170, NC6770, NC71xx, and NC77xx server ad apters, you must install the appropriate driver component located in the \WIN2003 or \WIN2003x 64 directory in the HP SoftPaq file. o For all other adapters, Windows® automatically de tects and installs a driver for the adapter from its own library of drivers. Updating network drivers from disk
Installation 16 To update the network driver, follow these instructions: Locate the driver on the CD, diskette, or directory on your hard driv e where you copied it from an updated source. See Technical Support for a list of online sources for software and driver updates. 1. Run the appropriate Driver Component Package (lis ted as CPxxxxxx.exe) in a directory on your hard drive. The component is a self-extracting execut able with a file name based on the component number. 2. Click the Install button to proceed with the installation. 3. After the drivers have been installed, you may delete the Component Package that you copied. Windows NT4 unattended install (push) The following provides inform ation about Windows NT support for HP NC-series network adapters. Overview The primary objective of the Push Installation is to get the client properly connected to the Windows NT® server before the installation begi ns using the NDIS2 driver, and to keep the same client connected using the NDIS4 driver after the Windows NT® server is installed without any end-user action. The difficulty is that the list of supported adapter s on the Windows NT® 4.0 CD-ROM is r estricted and special steps must be performed to update the installati on source files and modify the scripts to automate the process for the adapters not on the list. Main steps of the process 1. Create a Network Installation Startup disk from the Windows NT® server and modify it for your adapter. 2. Create a shared directory for the Windows NT® server installation files on the Wi ndows NT® server. 3. Modify the UNATTEND.TXT installation script to automate the procedure and add the proper adapter installation and configuration information. The sample UNATTEND.TXT file, provided in Item 4 of the "Samples and Templates" section, indicates how to override default registry paramet ers for the adapter in the [CPQN ic1ParamSection] section. The override parameters sho uld not be used unless the def ault parameters will not work properly for a specific installation. Th e sample parameters are commented out. 4. Integrate the adapter driver files into th e Windows NT® installa tion source files. 5. Perform the Unattended Installation bootup with th e prepared startup disk from the client system. Performing an unattended installation on Windows NT 4.0 1. Prepare a Windows NT® 4.0 server: a. Create a folder on the server hard disk for the installation files (for example, NT40.INS). Make this folder shared (for example, with shar ed name NT40.INS). b. Create a user (for example, USERID with a password password) and grant read and view rights for the NT40.INS folder. 2. Use a CD-ROM drive at the Windows NT® server an d copy the i386 folder with Windows NT® 4.0 installation files to the shared folder NT40.INS on this server (for ex ample, using NT Explorer). 3. Create or modify the installation answer file UNA TTEND.TXT as required by your specific syst em and desired Windows NT® config urat ions. UNATTEND.TXT may be re ad-only and may need to be made write-able before editing (for example, a ttrib -r unattend.txt). Refer to the sample UNATTEND.TXT file with comments. 4. Create the \CPQANC direc tory: NT40.INS\i386\DRVLIB.NIC\CPQANC.
Installation 17 5. Copy driver installation files: Xcopy the HP Configuration and Drivers Disk to the proper location under Adapter Device Driver folder (NT40. INS\i386\DRVLIB.NIC\CPQANC fol der on the Windows NT® server). This enables the installati on program to find the same file structure as on the Drivers disk. Be sure to use th e /s /e switches on xcopy. 6. On the Windows NT® server, prepare the Microso ft® Client Installation disk using the Windows NT® Network Client Administrator: a. Prepare a system diskette (for example, DOS 6. 22: format a: /s). Do not use the Windows NT® system disk. b. Copy UNATTEND.TXT from \NT40.INS\i386 to th e diskette as specified in item 3 above (preferably to the root (a:\)). The best location to pu t this file is in the root of your install disk, but the only requirement is to specify the full path to this file in the AUTOEXEC.BAT for the WINNT command. c. Use the Windows NT® Administrat ive Tools (Co mmon), start Network Client Administrator and select Make Network Installation Startup Disk. d. Set "Existing Path" to use the previously install ed software for MS Cl ient or, if it is the first time, create the shared directory on the server' s ha rd drive, using the Windows NT® 4.0 Server CD- ROM as a source (for example, by copying from \client on the CD-ROM to c:\client). e. Choose Network Client v3.0 fo r MS-DOS and Windows. f. Choo se any of the Network Adapters from the list (for example, NE2000 compatible). g. Set Computer name, User name, Domain and Network protocol when prompted. h. Select OK and wait while files are copied to the disk. i. Copy the proper ndis2 dos driver (N100.DOS, N1000.DOS, or Q57.DOS) to this disk in the \NET directory. N100.DOS, N1000.DOS, Q57.DO S can be found in \DOS\NDIS2\ in the HP SoftPaq file. j. Modify A:\NET\SYSTEM.INI : netcard=N100.DOS (or N1000.DOS or Q57.DOS). k. Modify A:\NET\PROTOCOL .INI : drivername =N100$ (or N1 000$ or Q57$ ). l. Add a line: CACHEFLUSH=1 just after the "drivername = N100 $" line. m. Verify or modify A:\AUTOEXEC.BAT to have the following commands: NET USE W: \\ServerName\NT40.INS W:\WINNT /s:W:\ /u:a:\unattend.txt Note that W is an example for the logical drive mapped to the share on the Windows NT® server. n. Remove the invocation of setu p.exe from A:\AUTOEXEC.BAT. 7. Insert the startup disk and boot up your DOS clie nt. After connecting to the network. do one of the following: o Input USERID as a User name when prompted. Input password as a password when prompted. o Modify the NET START line in the AUTOEXEC .BAT to NET LOGON USERID PASSWORD /YES. Samples and templates ⢠PROTOCOL .INI [network.setup] version=0x3110 netcard=ms$ne2clone,1,MS$NE2CLONE,1 ;transport=ms$ndishlp,MS$NDISHLP
Installation 18 ;transport=ms$netbeui,MS$NETBEUI ;lana0=ms$ne2clone,1,ms$netbeui ;lana1=ms$ne2clone,1,ms$ndishlp transport=tcpip,TCPIP lana0=ms$ne2clone,1,tcpip [ms$ne2clone] drivername=N100$ CACHEFLUSH=1 ; CACHEFLUSH line is required ONLY IF NDIS2 does not shut down ; properly ;INTERRUPT=3 ;IOBASE=0x300 ;SlotNumber=1 [protman] drivername=PROTMAN$ PRIORITY=MS$NDISHLP ;[MS$NDISHLP] ;drivername=ndishlp$ ;BINDINGS=ms$ne2clone ;[ms$netbeui] ;drivername=netbeui$ ;SESSIONS=10 ;NCBS=12 ;BINDINGS=ms$ne2clone ;LANABASE=0 [tcpip] NBSessions=6 DefaultGateway0= SubNetMask0=255 255 0 0 IPAddress0=137 65 10 12 DisableDHCP=1 DriverName=TCPIP$ BINDINGS=ms$ne2clone LANABASE=0 ⢠SYSTEM.INI [network] filesharing=no printsharing=no autologon=yes computername=COMPUTERNAME lanroot=A:\NET username=Administrator workgroup=DOMAIN reconnect=no directhost=no dospophotkey=N lmlogon=0 logondomain=DOMAIN preferredredir=full autostart=full maxconnections=8
Installation 19 [network drivers] netcard=N100.DOS ;transport=ndishlp.sys,*netbeui transport=tcpdrv.dos,nemm.dos devdir=A:\NET LoadRMDrivers=yes [Password Lists] *Shares=a:\net\Share000.PWL USERID=A:\NET\USERID.PWL ⢠AUTOEXEC.BAT path=a:\net a:\net\net start rem a:\net\net logon userid password /yes net use W: \\ServerName\nt40.ins W:winnt /s:W:\ /u:a:\unattend.txt ⢠UNATTEND.TXT [Unattended] OemPreinstall = yes NoWaitAfterTextMode = 1 NoWaitAfterGUIMode = 1 FileSystem = LeaveAlone ExtendOEMPartition = 0 ConfirmHardware = no NtUpgrade = no Win31Upgrade = no TargetPath = winnt OverwriteOemFilesOnUpgrade = no OemSkipEULA = yes [GuiUnattended] OemSkipWelcome = 1 OEMBlankAdminPassword = 1 TimeZone = "(GMT-08:00) Pacific Time (US & Canada)" [UserData] FullName = "Userid" OrgName = "Organization" ComputerName = COMPUTER ProductId = "xxx-xxxxxx" [Display] ConfigureAtLogon = 0 BitsPerPel = 8 XResolution = 640 YResolution = 480 VRefresh = 60 AutoConfirm = 1 [Network] InstallAdapters = CPQAdapters InstallProtocols = ProtocolsSection
Installation 20 InstallServices = ServicesSection JoinWorkgroup = workgroup [ProtocolsSection] TC = TCParamSection [TCParamSection] DHCP = yes [CPQAdapters] CPQANC.1 = CPQNic1ParamSection, \i386\drvlib.nic\CPQANC ;CPQANC.2 = CPQNic2ParamSection, \i386\drvlib.nic\CPQANC ;CPQANC.3 = CPQNic3ParamSection, \i386\drvlib.nic\CPQANC ;CPQANC.4 = CPQNic4ParamSection, \i386\drvlib.nic\CPQANC [CPQNic1ParamSection] ;NetAddress = "" ;SpeedDuplex = 0 ; 0 means autonegotiation ;MaxFrameSize = 0 ; 0 is use card default (1514) [ServicesSection] References ⢠Microsoft® Window s NT® Resource KitâMic rosoft® Press 19NT ⢠Microsoft® TechNet CD-ROM, April 1997 ⢠Microsoft® support web page at http://support .microsoft.com/support Article ID : Q155197, Q156795 Linux The following Linux drivers are available for HP adapters: ⢠tg3âSupports NC150x, NC320x, NC324x, NC32 5x, NC326x, NC1020, NC67xx, and NC77xx server adapters ⢠e1000âSupports NC1 10T, NC310F, NC340x, NC 360x, NC364T, NC61xx, and NC71xx server adapters ⢠hp-pelâSupports ProLiant Essentials Licen sing package ⢠hp-vtâSupports Virus Throttle for NCxxxx adapters ⢠iSCSI-bootâSupports the HP iSCSI boot f eature that allows booting a server over iSCSI ⢠unmâSupports HP NC510x PCIe 10 Gigabit Adapter driver The drivers are distributed in source RPM format and are available in the HP SoftPaq file in the \Linux directory. Complete the following steps to download the HP SoftPa q. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number.
Installation 21 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. In addition, the Linux (open source) drivers can be found at the HP software and drivers website ( http://h18000.www1.hp.com/s upport/files/index.html ). HP ROMPaqs for specific s ervers can be found at the HP support website ( http://www.hp.com/support ). The source RPM is suitable f or use on any system that has the RPM utility installed and a properly configured kernel source tree. SCO OpenServer and UnixWare SCO OpenServer 5.x, 6.x, and UnixWare 7.x driv ers are provided in the following locations: ⢠HP SoftPaq file at the HP website ( http://www.hp.com ). ⢠HP EFS 7.64A for UnixWare 7.x and OpenServer 6.x, and HP EFS 5.64A for SCO OpenServer 5.x at the HP website ( http://www.hp.com ). HP Extended Feature Supplement (EFS) is a colle ction of device drivers that provide maximum performance for HP servers. T he EFS also includes several utilit ies to monitor and increase the performance of your HP server. SCO OpenServer 5.x This procedure describes the installation of the SCO O p enServer driver. This driver is released as a media image file containing the driver package. The media image file can be copied to the target ma chine directly for installation or from an in stallation diskette that you can create. Creating a diskette 1. Copy the file VOL.000.000 to a Calder a system. 2. Create a diskette using: dd if =VOL.000. 000 of=/dev/rfd0135ds18. Installing the driver To install the driver: 1. Use custom or scoadmin software to instal l the SCO OpenServer driver from the media image or from the diskette created above. 2. Use netconfig to add the new network adapter. 3. Modify the hardware configuration in Advanced Options to change the Line Speed and Flow Control if desired. The settings for th ese parameter s are listed below: o Line Speed AutoNegotiate (defa ult) HalfDuplex10 FullDuplex10 HalfDuplex100 FullDuplex100 FullDuplex1000 o Flow Control Off (default) AutoNegotiate (Symmetric Paus e advertised)
Installation 22 RxPause TxPause RxPause/TxPause NOTE: 1000 Mbps (1 Gbps) full-duplex fixed speed is only valid for fiber connections. For copper, twisted-pair connections , 1 Gbps can only be set through auto-negotiation with a 1 Gbps partner. NOTE: Auto-negotiation of Flow Control is only valid when the Line Speed is set to AutoNegotiate. NOTE: A kernel relink and reboot is required before the new configuration will take effect. For more detailed information and tuning par ameters, at the command prompt enter: man bcme Jumbo Frames and other advanced parameters Jumbo MTU sizes and other advanced, tunable paramet ers for the controller are located in the file space.c in the directory /etc/conf/pack.d/bcme. A description for each parameter is contained in space.c. Modify the desired parame ter in space.c, rebuild the kernel, and reboot the syst em. The MTU sizes can be individually set for each adapter in the system (Jumbo Frames is not supported on the NC1020). All other parameters apply globally to all adapters. UnixWare 7.x and OpenServer 6.x The following procedure describe s the installation of the UnixWare driver. The driver is release d as a PKG file. Installing the driver locally 1. Copy the file bcme-<version>.pkg to a UnixWare 7.x system. 2. Enter the following command: pkgadd -d <path><filename> where <path> is the directory to wh ich the PKG file was copied. Creating a diskette 1. Copy the file bcme-<version> .pkg to a Caldera system. 2. Create a diskette using the following command: dd if=bcme-<vers ion>.pkg of=/d ev/fd0 Installing the driver from a diskette 1. Use pkgadd -d diskette1 to install the bcme package. 2. Use netcfg or scoadmin network to add the new network adapter. 3. When prompted, select the Line Speed and then select Advanced Option for Flow Control, MAC Address, and Jumbo MTU Size settings, if desir ed. The settings for these parameters are li sted below: o Line Speed Auto Ne gotiati on (de fault ) 10 Mbps half-duplex 10 Mbps full-duplex 100 Mbps half-duplex
Installation 23 100 Mbps full-duplex 1000 Mbps full-duplex o Flow Control Disabled (default) Auto Neg otiation ( Symmetr ic Paus e advertised) Receive Pause Transmit Pause Receive & Transmit Pause o MAC Address No Override (default) - a user-administered M AC address entered with a colon separating each hexadecimal byte (for example, 12:34:56:78:9a:bc). o Jumbo MTU Size 1500â9000 (default i s 1500) For more detailed information, at the command prompt, enter: man bcme NOTE: 1000 Mbps (1 Gbps) full-duplex fixed speed is only valid for fiber connections. For copper, twisted-pair connections , 1 Gbps can only be set through auto-negotiation with a 1 Gbps partner. NOTE: Auto-negotiation of Flow Control is only valid when the Line Speed is set to AutoNegotiate. Solaris The following describes the Solaris driv er for HP 10/100/1000 Mbps Netwo rk Server Adapters. The driver is released in two formats: ⢠BRCMbcme.pkg: Datastream format ⢠bcmedu.img: Driver Update (DU) binary imag e which is used to create DU diskette. Installing the driver 1. Change directory to wher e BRCMbcme.pkg resid es. 2. pkgadd -d BRCMbcme.pkg 3. Execute prtconf to determine instance number of the adapter. 4. ifconfig bcme [instance_number] plumb 5. ifconfig bcme [instance_nu mber] ip_address netmask .... To make these changes permanent: 1. Use your favorite text editor (for ex ample, VI), and create a file named hostname.bcme[instance_number] in /etc directory. Add the IP address of the interface to this file, save and exit. 2. Add a proper subnet mask to the file /etc/netmasks.
Installation 24 In Solaris 7.0 (Intel platform), the operating system only allocates 36 pages of 4 K physically contiguous memory. The driver needs approximately 130K of physic ally contiguous mem ory per adapter. In order to use more than one adapter, the operating system must allocate more memory. This can be done by setting the operating system variable, "lomempages" in /etc/ system. For example, if four adapters are installed in a So laris 7 system, physically contiguous memory is calculated as follows: 4 adapters * 130K = 520K 130 pages of 4K is required. Since this memory might be used by another driver in the system, 200 of 4K memory is allocated. Add the following line in the /etc/system file: set lomempages=200 Uninstalling the driver To uninstall the driver: 1. ifconfig bcme [instance_number] down 2. ifconfig bcme [instance_number] unplumb 3. pkgrm BRCMbcme Driver update (DU) diskette instruction (Intel® Platform Only) This area contains the Solaris diskette image files, and the instruction s to create diskettes from these image files. 1. Insert a blank diskette into the di skette drive of your system, and type one of the following commands to format it: o If you are using DOS, type: format A: o If you are using the Solaris operating environment, type: fdformat -Ud 2. Check to see if Volume Management is running: a. Do one of the following: If you are using DOS, type: dd filename A: If you are using the Solaris operating environment, type: volcheck ls -l /vol/dev/aliases/floppy0 b. If you see a message similar to this: lrwxrwxrwx 1 root 34 Jan 21 17:28 /vol/dev/aliases/floppy0 -> /vol/dev/rdiskette0/unnamed_floppy Type: dd if=bcmedu.img of=/vol/dev/aliases/floppy0 bs=36k eject floppy0 c. If you see this message: /vol/dev/aliases/floppy0 not found Type:
Installation 25 dd if=bcmedu.img of=/dev/rdiskette bs=36k Installing Solaris DU dskettes You can use Solaris DU diskettes in the following ways: ⢠To use new drivers to install or upgrade t he Solaris operating system environmen t on a system with new hardware. ⢠To add new drivers to support new hardware on an already installed and booted system. Installing Solaris using DU diskettes To install Solaris (Intel Platform Edition) using driver s on the DU diskette: 1. Insert the appropriate configuration Assistant diskette , made from the included bcmedu.img file, into your system's disk drive. Also, insert the S olari s Installation CD-ROM or, for network installation, verify with your system administrator that the Sola ris network installation im age is available on your network. 2. Power on your system. 3. When the Configuration Assistant scr een is displayed, select the F4 option (on ver sion 2.6, it is labeled F4_Driver Update; on Solaris 7, it is la beled F4_Add Driver). Th e message, "Enumerating buses...." is displayed. 4. Remove the Configuration Assistant diskette from the disk drive and insert the first Solaris DU diskette. 5. Select F2_Continue. The Select Solaris System Ve rsion screen is displayed. 6. Select the appropriate Solaris OS and select F2_Continue. The Loading Driver Update Software screen is disp layed, along with a prog ress bar that indicates the percentage of drivers that have been extracte d from the diskette. Drivers are read into mem ory and survive long enough for the system to successf ully boot to its installation program. When all the new drivers on the diskette have been processed, th e Contin ue Driver Update Installation screen is displayed. 7. If you have additional DU diskettes to install, remove the DU diskette from the dis k drive, insert the next DU diskette. 8. Select F2_Continue. Again, the Loading Driver Update Software screen is displayed, along with a progress bar that shows the percentage of drivers that have been extracted from the diskette. Drivers are read into memory and survive long enough for the syst em to successfully boot to its installati on program. When all the new drivers on the diskette have been processed, the Contin ue Driver Update Installation screen is displayed. 9. Repeat step 7 and 8 until all desired DU diskettes are i nstalled. 10. When all the drivers are processed, remove the DU diskette from the disk drive and reinsert the Configuration Assistant diskette. IMPORTANT: Do not remove the Configuration Assistan t diskette from the disk drive until you see the following message display: "If you want to bypass the device configuration and boot screens when the system reboots, eject the Device Configuration Assistant/Boot diskette now." 11. Select F2_Continue. The Solaris Device Configuration Assistant scr een is displayed. 12. Select F2_Continue. The message, "Enumerating buses...." is displayed, followed by the Scanning Devices screen. System devices are scan ned. When scanning is complete, the Identified Devices screen is displayed.
Installation 26 13. Select F2_Continue. The message "Loading driv er...." is displayed, followed by messages about the drivers that are required to boot your system. After a few seconds, th e Boot Solaris screen is displayed. 14. At the Boot Solaris screen, select the device controller attached to the device th at contains your install medium. 15. Select F2_Continue. Drivers display for the device controller you sele cted. Your system boots to run the install program. The install program starts and your system begins booting the complete Solaris operating system environment. Then the follow ing messages are displayed, prompting you to in sert each of the Solaris DU diskettes required to install on your system: Installing unbundled device driver support Extracting driver list from tree.. <DU diskette name> driver-name... Please insert the Driver Update diskette labeled <DU diskette name> driver-name. Press <ENTER> when ready. 16. Remove the Configuration Assistant diskette and rein sert the first DU diskette into the disk drive. 17. Press the Enter key. Packages, patches, or both that cont ain the new drivers are installed from the diskette onto your system. Me ssages about each inst allation are displayed. If drivers on other D U diskettes are required for your sy stem, the following prompt is displa yed. Please insert the Driver Update diskette labeled <DU diskette name> Press <ENTER> when ready. Otherwise, this prompt is displayed: If you have additional Update diskettes to install (such as video), please insert diskette now. Additional Update diskettes to install? (y/n) [y] 18. If drivers on other DU diskettes are requir ed, remove the DU diskette from the diskette drive, insert the next DU diskette you are prompted to insert. Press t he Enter key. Repeat until all the drivers you need are installed. 19. Press the Enter key. When installation is complete, the messag e "Installation complete" is displayed. 20. Remove the diskette from the disk drive. 21. Reboot the system. When the Solaris operating system is fini shed booting and running, the new devices whose drive rs you installed are available for use. Adding a DU diskette to an existing Solaris system Before adding new or updated drivers, the newly su pported hardware devices should be installed and configured according to the instructions in the corre sponding Device Reference Page, if any. Refer to the Solaris (Intel® Platform Edition) Device Configuration Guide. When the Solaris (Intel® Platform Edition ) software is already installed, the simplest way to add new or updated drivers is to install the DU diskett es as patches on your system, as follow s: 1. Become the superuser on your system. 2. Check to see if Volume Man a gement is running on the system you are updating: ps -ef | grep vold For more information about managing diskette and drives, refer to the System Ad ministration Guide.
Installation 27 3. If Volume Management is running, temporarily stop it by typing: # /etc/init.d/volmgt stop 4. Insert the DU diskette into the disk drive. 5. Mount the DU diskette at the /mnt mount point: # mount -F pcfs /dev/diskette /mnt At this point, you must mount the DU diskette in the file structure to up date your system successfully. 6. Execute the install script on the diskett e, using th e appropriate Solaris release directory (currently sol_26 for Solaris 2.6, sol_27 for Solar is 7, and so on). For example: # /mnt/DU/sol_27/i86pc/ Tools/install.sh -i The install.sh script searches for all new or updat ed drivers on the diskette. When a new or updated driver is found, the following prompt is displayed: Unconditionally installing DUs <DU driver names> Install patch driver-name? [y] 7. If the driver is the one you want to install, at the prompt, type y for yes or press th e Enter key. If the driver is not the one you want to install, type n for no. If you select "yes," then the install.sh script inst alls the driver you selected, as well as the bootmod and bootbin patches. 8. When you have finished, and the install.sh script exits, unmount th e diskette: # cd / # umount /mnt 9. Remove the DU diskette from the disk drive. 10. Reboot the system. # touch /reconfigure # reboot 11. Power down your system, add the new hardware, and power up the system. 12. When the autoboot sequence prompt is displayed, quickly press the Esc key. The au toboot sequence is interrupted, and the Configuration Assistant screen is displayed. 13. Select F2_Continue. The message "Enumerating buses ..." is displayed, followed by the Scanning Devices screen. System devices are scan ned. When scanning is complete, the Identified Devices screen is displayed. 14. Select F2_Continue. The message "Loading driver com.bef ..." is displayed, followed by the Boot Solaris screen. 15. On the Boot Solaris screen, select the device controller attached to the devi ce that contains your install medium, in this case the main system disk. T he /etc/bootrc script is displayed. 16. At the prompt, type: b -r Your system boots. You can now use your new hardware.
Installation 28 NetWare A network device driver must be in stalled before the Gigabit Ethernet adapter can be used with your Novell NetWare system. Before you c an successfully in stall the adapter driver for Novell NetW are, the adapter card must be physically installed in the serv er and, typically, NetWare OS software must already be running on the server. Make sure that your server meets the hardware and operating system software requirements. For an adapter installation with an existing NetW are server, NetWare will automatically detect the new adapter and attempt to load the appropriate driver. Ensu re that your CD-ROM is mounted and select the proper driver. To enable the Gigabit Ethernet adapter to function corr ectly, you need to install the latest support pack files. The latest support pack can be found at the Novell website ( http://www.novell.com ). Netware install program A commonly used method to install a driver on a NetWare server is through NWCONF IG. The following drivers are supported. ⢠Driver names: N1000e.LAN, N1000.LAN, and N100.LAN ⢠HP N1000e.LAN driver requirements o Use with an HP NC110T, NC360T, an d NC364T Gigabit Server Adapter o Always install the latest Novell support pack o Always install the latest HP NSSD o SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) o SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) o SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) ⢠HP N1000.LAN driver requirements o Use with any HP NC310F, NC340T, NC61 xx, and NC71xx Gigabit Server Adapter o Always install the latest Novell support pack o Always install the latest HP NSSD o SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) o SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) o SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) ⢠HP N100.LAN driver requirements o Use with any HP NC31xx Fast Ethernet Server Adapter o Always install the latest Novell support pack o Always install the latest HP NSSD o SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) o SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) o SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) Using NWCONFIG or HDETECT to install the adapter
Installation 29 NOTE: Your NetWare server software must be th e latest version of the support pack before installing the HP driver. The latest support pack is available from the Novell support site: http://support.novell.com/misc/patlst.htm. Before installing, use the HP SoftPaq to create a floppy disk on a different computer. Refer to the README.TXT file located in the HP SoftPaq for det ails on creating a floppy disk. Use the NetWare Install programs to install the adapter driver. See the Novell Installation Guide for specific instructions on installing device drivers. Packet Receive Buffers Setting is Too Low error message When a server is first installed or has been upgraded, the number of MAXIMU M PACKET RECEIVE BUFFERS and MINIMUM PACKET RECEIVE BUFFERS are usually too low. The adapter requires a minimum of 64 Packet Receive Buffers per ad apter port during driver initialization. This initial load of the driver may fail because of the two above mentioned se ttings. The followin g error message is displayed: *** Packet Rece ive Buffers setting is too low. *** To correct: 1. Load MONITOR. 2. Select SERVER PARAMETE RS. 3. Select COMMUNICATIONS. 4. Set MAXIMUM PACKET RE CEIVE BUFFERS = 10000 (or larger). 5. Set MINIMUM PACKET RE CEIVE BUFFERS = 2000 (or larger). 6. Set MAXIMUM PHYSICAL RECEIVE P ACKET SIZE = 2048 (or larger). 7. Restart NetWare to complete. This can also be set by using the NWCONFIG an d selecting the Startup.NCF file. See the Netware Installation Guide for specific instructions. The maximum and minimum settings of 10000 and 2000 respectiv ely are suggested values only. You may need to request a larger number for MAXI MUM PACKET RECEI VE BUFFE RS depending on the number of supported adapter ports in the system. Th is message is also displayed on the console if the driver cannot allocate the required number of Receive Buf fers during driver load. Driver configuration parameters Parameter Options Description CheckSum = Default = ON Selections are: ON, OFF, Tx, Rx Enables or disables the transmit and receive checksum off-loading feature. Checksum is supported under NetWare 5.x only. If you want to enable the Ch eckSum parameter, you need to load it on the first instance. Frame = type Valid types are: Ethernet_802.2, Ethernet_802.3, Ethernet_II, Ethernet_SNAP Defines the frame type being used by this load instance. Ethernet_802.2 and Ethernet_II are the default values. node = NNNNNNNNNNNN Specifies a no de address in this field to override the default Media Access Controller (MAC) address (also known as the Locally Administered Address) name = text Displays the name assigned to this adapter
Installation 30 Parameter Options Description PDriver = Default = OFF Selections are: OFF, ON Allows driver to operate in persistent driver mode. Persistent driver mode is supported under NetWare 5.x only. Use only if adapter is placed in a Hot Plug PCI slot and only if required to swap with an exact board. RxBuffers = Default = 200 Recommended Min = 32 Max = 512 Min = 1 when used with DOS Client32 and when Keywrod P3-1 or 2. Pre-allocates ECB resources on the adapter for receiving packets TxDescriptors = Default = 200 Recommended Min = 100 Max = 512 Min = 1 when used with DOS Client32 and when Keywrod P3-1 or 2. Pre-allocates ECB resources on the adapter for transmitting packets. RxFlow = Default = OFF Selections are: ON, OFF Allows enabling/disabling of RxFlow control. TxFlow = Default = OFF Selections are: ON, OFF Allows enabling/disabling of TxFlow control. Slot = n Identifies the slot number for the specific adapter currently being configured. This parameter is not necessary if only a single adapter is installed. Speed = n If link negotiation has been disabled, specifies port speed to be either Auto, 10HD or 10FD, 100HD or 100FD. Jumbo = Set maximum physical receive packet Size = 18000 in the STARTUP.NCF. Choices are Jumbo = 1514â9000. T his keyword is only supported on NetWare 6.x. Enables/disables Jumbo Frame support. When enabled, jumbo packets of up to 9000 bytes are supported. Not supported on NC1020 adapters. Link= Default=FORCE Selections are: AUTO, FORCE Only used to allow the adapter to negotiate a specific or forced line speed with a switch that is not forced, but instead setup for auto- negotiation. It is best to allow for auto- negotiation of the card and switch by not setting this keyword or the speed keyword. Only use this keyword if the speed keyword is set to something other than AUTO. RxTicks= Default = 360 Min = 0, disabled Max = 5000000, 5 second s Units are in micro seconds Enables the use of batching receives within a specific time period.
Installation 31 Parameter Options Description TXPacketsPer Default = 64 Min = 0, disabled Max = 100 Enables the use of batching transmits to a specific amount of packets. NOTE: With Jumbo Frames, the first frame must be Ethernet_ii. Post installation After NetWare 6.x has been successfully installed, set the minimum packet receive buffers param eter in the startup.ncf file to 1500 for each adapter in th e system. Set the maximum packet receive buffers to three times the minimum packet receive buff ers. Typica lly one MB of RAM is required per 1000 receive buffers. For more information, refer to the "Configuring Driver Param eters" section. In the autoexec.ncf file, delete the packet receive buffers param eter (RxBuffers=32) in the l oad statement for this adapter. Deleting the receive b uffers phrase from the load statement res ets the receive buffers parameter to the default value of 200 fo r this adapter. NOTE: You must reboot the server for the new configuration. Example: The default maximum number of rec eive buffers for the system is 500; the default minimum is 128. Edit the startup.ncf file to have the following entries. Th e actual numbers will be a func tion of the number of adapters in the system. ⢠set maximum packet receive buffers = 30000 ⢠set minimum packet receive buffers = 1 0000 ⢠set maximum physical receive packet size = 2048 Verifying or modifying adapter properties When an adapter configuration is saved, the NetWare install program adds load and bind statements to the autoexec.ncf file. By accessing this file , you can verify the parameters configured for each adapter, modify them, or enter additional parameters. NOTE: The Novell monitor program and the CONFIG command are also us eful for verifying driver configuration. For information on how to use these programs, ref er to the utilities reference in your Novell NetW are online documentation. The parameters that can be defined in the load statem ents are described in NetWare server dri ver LOAD line parameters for HP server adapters below. A valid au toexec.ncf file is shown below. One set of load and bind commands is added for each frame type the adapter is configured to support. Valid Autoexec.ncf file Set Time Zone = PST8PDT set Daylight Savings Time Offset = 1 set Start Of Daylight Savings Time = (APRIL SUNDAY FIRST 2:00:00 AM) set End Of Daylight Savings Time = (OCTOBER SUNDAY LAST 2:00:00 AM) set Default Time Server Type = SINGLE set Bindery Context = O=LAN
Installation 32 # WARNING!! file server name NOVELLSERVER51 # WARNING!! # If you change the name of this server, you must update # all the licenses that are assigned to this server. Using # NWAdmin, double-click on a license object and click on # the Certificate Assignments button. If the old name of # this server appears, you must delete it and then add the # new server name. Do this for all license objects. ServerID 1C8EE2C LOAD ODINEB.NLM LOAD TCPIP LOAD Q57 SLOT=2 FRAME=Ethernet_802.2 NAME=Q57_1_E82 BIND IPX Q57_1_E82 NET=FAFD3D25 LOAD Q57 SLOT=2 FRAME=Ethernet_802.3 NAME=Q57_1_E83 BIND IPX Q57_1_E83 NET=5A2D8D6D LOAD Q57 SLOT=2 FRAME=Ethernet_SNAP NAME=Q57_1_ESP BIND IPX Q57_1_ESP NET=477A35BD LOAD Q57 SLOT=2 FRAME=Ethernet_II NAME=Q57_1_EII BIND IPX Q57_1_EII NET=C3C8F2E4 BIND IP Q57_1_EII ADDR=172.16.1.1 MASK=ff.ff.0.0 mount all SEARCH ADD SYS:\JAVA\BIN SEARCH ADD SYS:\JAVA\NWGFX IMPORTANT: If you modify any adapte r parameters, you must reboot the system before the changes will take effect. If you make changes and do no t reboot, you may experience configuration problems. Removing drivers from Autoexec.ncf To remove the drivers from the Autoexec.ncf, locate the LOAD and BIND command lines as sociated with the driver and remark them out by inserting the poun d (#) symbol at the beginning of each command line. Example: # LOAD Q57 SLOT=2 FRAME=Ethernet_802.2 NAME=Q57_1_E82 # BIND IPX Q57_1_E82 NET=FAFD3D25 # LOAD Q57 SLOT=2 FRAME=Ethernet_802.3 NAME=Q57_1_E83 # BIND IPX Q57_1_E83 NET=5A2D8D6D # LOAD Q57 SLOT=2 FRAME=Ethernet_SNAP NAME=Q57_1_ESP # BIND IPX Q57_1_ESP NET=477A35BD # LOAD Q57 SLOT=2 FRAME=Ethernet_II NAME=Q57_1_EII # BIND IPX Q57_1_EII NET=C3C8F2E4 # BIND IP Q57_1_EII ADDR=172.16.1.1 MASK=ff.ff.0.0 NetWare server driver LOAD line parameters This following lists the N etWare server driver LO AD line parameter settings for HP serve r adapters for N100, N1000, and N1000e. ⢠FORCEDUPLEX. This keyword specifies one of the following dupl ex modes: o Auto-negotiate. The adapter negotiates with the switch and hub how to send and receive packets at either full- or half-duplex speed. If uns uccess ful at negotiating the duplex, HP server adapters default to half-duplex. Yo u must have an auto-n egot iating switch/hub (an Nway switch) to get full-duplex support with the FORCEDUPLEX parameter set to 0 (auto-negotiation).
Installation 33 o Full-duplex. The adapter sends and receives packets at the same time. This improves the performance of your adapter. Set duplex mode to full-duplex ONLY if you have a hub/switch that supports full-duplex. o Half-duplex. The adapter performs one operation at a time. It either transmits or receives. NOTE: The HP NC31xx, 340T, and NC71xx Server Adapters support full-duplex and half- duplex. The HP NC61xx fiber Gigabit Server Ad apters support full-duplex mode only and cannot be changed. NOTE: The NC3133 default to full-duple x. NOTE: To set the adapter to half- or full-duplex, yo u must set the SPEED parameter to either 10 or 100. o Syntax: FORCEDUPLEX=n Where n = 0âauto-negotiate 1âhalf-duplex 2âfull-duplex o Default = auto-negotiate o Examples: 100 Mbps full-duplex FORCEDUPLEX=2 SPEED=100 10 Mbps full-duplex FORCEDUPLEX=2 S PEED=10 ⢠SLOT. (Required) For PCI adapters, SL OT is derived from bus number and device location as defined by the PCI specification. One way to determine th e slo t number is to load the driver from the command line. You will be prompted with valid de vice number(s) for the adap ter(s). Select one of them. o Syntax: SLOT=n Where n = 1, 2, 3, 4,... ⢠SPEED. This keyword specifies the speed the driver use s. If you do not use this parameter, the driver automatically detects the network spee d. If unable to detect, the driver defaults to 10 Mbps. If you use this parameter, the driver operates at the specified speed instead of auto detecting netw ork speed. NOTE: The NC3133 adapter supports 100 Mbps on ly. It does not default to 10 Mbps. You must set the SPEED parameter to either 10 or 100 if you want to set the FORCEDUPLEX parameter to either half or full. The SPEED parameter is for N100.LAN, N1000.LAN, and N1000e.LAN drivers. You cannot change the spee d for the Gigabit (fiber) adapters. You can, however, change the speed for the NC7 170, NC7131, and NC7132 Gigabit (copper) adapters to 10 Mbps or 100 Mbps. o Syntax: SPEED=n Where n = 10 or 100 or 1000 o Default = The adapter automatically detects network speed.
Installation 34 ⢠NODE. This keyword specifies a locally administered address (LAA) unique to each adapter. Use this option to provide your own unique node address for the adapter. The node address is a 12-digit hexadecimal number. The second digit must be one of the followin g digits: 2, 6, A, E. o Syntax: NODE=xnxxxxxxxxxx Where n = 2, 6, A, E Where x = hexadecimal number o Default = The adapter's assigned address ⢠FRAME. This keyword indicates one of fo ur valid frame types the system is transmitting and rec eiving. o Syntax: FRAME=n Where n = Ethernet_802.2, Ethernet_802.3, Ethernet_II, Ethernet_SNAP o Default = Ethernet_802.2 ⢠POLL. This keyword is used to enable polling mode in the driver with interrupt backup. If polling is enabled, interrupts will be reduced, allowing th e processor to spen d more time performing other functions. In polling mode, interr upts will occur only when receive re sources have been reduced to less than half. If polling is no t enabled, the driver will perfor m in traditional interrupt mode. o Syntax: POLL=n o Where n = 0 (interrupt mode), 1 (polling mode with int errupt backup) o Default = 0 (interrupt mode) NOTE: The POLL parameter is for the N100.LA N, N1000.LAN, and N1000e.LAN drivers. ⢠SPURIOUS. This keyword is used to reduce the number of spurious interrupts reported by the OS on the driver interrupt line. This automati cally loads for 10, 100, and 1000 adapters. o Syntax: SPURIOUS=n Where n = 0 or 1 0âDriver operates in norm al mode 1âDriver operates to reduce the number of spur ious interrupts o Default = 1 NOTE: The minimum threshold for acceptable number of spurious int errupts is 200 interrupts per second, which is the default value of the environmental variable "set display spurious interrupt alerts threshold." ⢠RXCHECKSUM. This keyword is used to enable or disable the offlo ad of Receive Packet Checksum verification to the adapter. This au tomatically loads for 10/100 adapters. o Syntax: RXCHECKSUM=n Where n = 0 or 1 0âDoes not offload Rx Checksum ver ification to the adapter 1âOffloads Rx Checksum verification to the adapter o Default = 0 NetWare custom statistics NetWare custom statistics appear when you run monitor NLM and load the LAN driver. One custom statistic counter is the duplex mode, where 1 = half-duplex and 2 = full-dupl ex.
Installation 35 The following describes the adapter Status Counters and Software Counters an d their acceptable ranges. N100 custom counters ⢠Link Status This counter contains the link status, as read from the PHY, or general status register of the ASIC. It is updated once a second. A value of 0 is link up. 1 is link down. ⢠Duplex This counter holds the current duplex as read directly from the PHY, or general status register of the ASIC. A value of 0 is not supported or unknown, 1 is half-duplex, and 2 is full-duplex. Unsupported is usually due to an unsupported chip type, or link not being present. ⢠Current Speed This counter holds the current speed as read directly from the PHY, or general status register of the ASIC. A value of 0 is not supported or not available, 10 for 10 Mbps, 100 for 100 Mbps, and 1000 for 1000 Mbps. Unsupported is usually due to an unsupported chip type, or link not being present. ⢠Hardware Status A counter of non-zero indicates failure. ⢠Tx Good Frames This counter contains the number of frames that we re transmitted properly on the link. It is updated only after the actual transmission on the link is completed, and not when the frame was read from memory as is done for the TxCB status. This value va ries in accordance to the amount of frames sent, and can reach billions very quickly. ⢠Tx Maximum Collisions This counter contains the number of frames that were not transmitted because they encountered the configured maximum number of collisions. This counter increments when the net work is heavily saturated with traffic. If this va lue is consiste ntly growing, you may want to segment your network with switches to change the collision domain, so that more time is spent sending data instead of collisions. ⢠Tx Late Collisions This counter contains the number of frames that were not transmitted because they encountered a collision outside of the normal collision window. No rmally, late collision events should occur only rarely, if at all. A high number of late collisio ns is of ten a sign that there is some problem with the network topology. Check the topology to be sure that all cables m eet IEEE specifications (Category 5 for 100 Mbps TX), and that all distance requirements for cabling are met. Also, check to be su re that all of the network hubs and switches are configured for the proper duplex mode. ⢠Tx DMA Underruns This counter contains the number of frames that we re either not transmitted or re-transmitted due to a Tx DMA under-run. If the device is configured to retransmit on under-run, this counter may be updated multiple times for a single frame. Un der-runs occur because a lack of PCI bandwidth resulted in the device's internal transmit FIFO running dr y during the transmission of a frame. The adaptive algorithms within the software keep this nu mber under control. If a high number of under- runs are seen, you may want to load the driver with TxThreshold=200 on the command line. Example: Load N100 slot=2 Frame=Ethernet_802.2 TxThreshold=200 For additional information about the adaptive algori thm , refer to the "CurrentThreshold" section.
Installation 36 ⢠Tx Lost Carrier Sense This counter contains the number of frames that were transmitted by the device despite the fact that it detected the de-assertion of Carrier Sense durin g the transmission. ⢠Tx Frames Deferred This counter contains the number of frames that were deferred before transmission due to activity on the link. This is a normal condition in busy co nditions, and the number can grow quite large. ⢠Tx OK with Single Collisions This counter contains the number of transmitted fram es that encountered one, and only one, collision before being successfully sent. This is a normal condition in busy conditions. If sin gle collisions accumulate quickly, you may want to consider segm enting your network more . To combat collis ions, the software can adjust the interframe ga p to a llow more room between fr ames. More room between frames allows more frames on av erage. Refer to the "CurrentIFSValue" section for more details. ⢠Tx OK with Multiple Collisions This counter contains the number of transmitted frames that encountered more than one collision. This only gets incremented once, even though one frame m ay have had several collisions before being successfully s ent. This is a normal condition in busy conditions. If mu lti-collisions accumulate quickly, you may want to consider segmenting your netw ork more. To find the average number of multi- collisions a frame had to endure before be ing successful, use the following formul a: (TxTotalCollisions - TxSingle Collisions - TxMaxCo llisions) / TxMultiple Col lisions If the average is close to 16 (the Ethernet max co llisions before a fra me will no longer be retried), then you will need to segment your network to avoid losing frames and network efficiency to collisions. ⢠Tx Total Collisions This counter contains the total number of collisi ons that were encountered while attempting to transmit. This count includes late collisions and collisions from frames that encountered MAXCOL(16). If this counter is gro wing quickl y, you need to segment your network more. ⢠Rx Good Frames This counter contains the number of frames that we re received properly from the link. It is updated only after the actual reception from the link is completed and all the data bytes are stored in memory. ⢠Rx Crc Errors This counter contains the number of aligned frames discarded because of a CRC error. The CRC_ERRORS counter i s mutually exclusive to the ALIGNMENT_ERRORS and SHORT_FRAMES counters. CRC errors are usually ca used by late collisions whe re another station has sent data in the middle of a frame. This number should be very low. ⢠Rx Alignment Errors This counter contains the number of frames that are both mis-aligned (not an even length) and contain a CRC error. The ALIGNMENT_ERRO RS counter is mutually exclusive to the CR C_ERRORS and SHORT_FRAM ES counters. ⢠Rx No Resource Errors This counter contains the number of good fram es discarded because there were no resourc es available. Frames intended for a host whose Receive Unit is in the " No Resources" state fall int o this category. The number of resource errors can be re duced by increasing the number of receive buffers
Installation 37 that software uses, or increasing the perf ormance of the host machine's proce ssing power (which would allow for faster processing of the r esources). ⢠Rx DMA Overrun Errors This counter contains the number of frames known to be lost because the device's internal receive FIFO overflowed. This can occur if the device is unable to get the necessary bandwidth on the system (PCI) bus. If the overflow condition persists for more than one frame, the frames that follow th e first could also be lost. Howeve r, because there is no lost frame indicator they may not be counted. A frame that was counted as an overrun will not be counted in other error counters (SHORT_FRAMES, CRC_ERRORS , or ALIGNMENT _ERRORS). To c ombat th is, a faster bus machine, o r fewer devices on the bus will be needed. ⢠Rx Collisions Detect Errors This counter contains the number of frames that encountered collisions during frame reception. ⢠Rx Short Frame Errors This counter contains the number of received fram es that are shorter than the minimum frame length. The SHORT_FRAMES coun ter is mutually exclusive to the ALIGNMENT_ERR ORS and CRC_ERRORS counters and has a higher priority (a short frame will always increment only the SHORT_FRAMES counter). ⢠Flow Control TX Pause This counter contains Xon and Xoff frames transmitted. ⢠Flow Control RX Pause This counter contains Xon and Xoff frames received. ⢠Flow Control Unsupported This counter contains the MAC Control frames re ceived that are not Flow Control Pause frames. These frames contain the predefined MA X control type value but contain an unsupported opcode. ⢠TCO Rx (or Tx) Frames This counter is incremented at each received/transmitted TCO packet. ⢠Tx Timeouts This is a debug timer and not currently used. It should be zero. ⢠Tx Dynamic Chaining Count This is a performance evaluation counter an d not currently used. It should be zero. ⢠Current IFS Value This is the current Inter Frame Spacing (IFS) value that the driver is using. IFS is an Ethernet specification term for clock time spaces between fr ames sent by an individual card in a burst. Normally a card sends frames back to back with only the minimum frame spacing between each frame. Under low load conditions, this produc es maximum throughput. But under heavy load conditions, this means less time for other stations to send safely; thus, more colli sions and less throughput. By adaptively controlling the IFS, more stations can safely send, lo wering collisions and helping to keep throughput high. The larger this number is, the more collisions your n etwork is experiencing. ⢠Current Threshold As under-runs occur, the software can change the amount of time the hardware will wait before starting to transfer data onto the adapter's hardware buffers. By doing so, it can slow or stop the
Installation 38 amount of under-runs. The higher th is number, the more time the adapter is waiting befor e starting to transmit data onto the wire. Poor bus performance, large amounts of traffic or a large numb er of devices on the same bus can cause this to rise. If this number is always above 170, you may want to consider loading the driver with the TxThreshold parameter set to a higher value, such as 200. ⢠RstartCnt Re-Start Count, also known as RSC, is a support counter to see how often the restart effort occurs. Under some high traffic, or high utilization condit ions, the receiving unit does not start normally and needs a restart. RSC has no impact on the end user. It is intended for softwar e development use only. It is normal for this to be non-zero. ⢠REC Under some busy conditions, the receiving unit stalls and needs to be restarte d. This counter tracks the number of times it needs to be restart ed. RSC is a subset of REC. Most REC events restart normally, but the few that do not are REC events. Th is counter is intended for software developm ent use only and may be disregarded by the end user. It is normal for this to be non-zero. ⢠Checksum Good This counter indicates the number of received packe ts that the hardware has verified, the checksum, and if the checksum is valid. ⢠Checksum Bad This counter indicates the number of received packe ts that the hardware has verified, the checksum, and if the checksum is valid. ⢠Stalled Ints SW Resets This counter indicates the number of times the driver issued a complete reset (to the driver and to the adapter) as a result of a stalled in terrupt situation. A stalled interr upt situ ation is detected by the driver if there are no interrupts for a long time, and a SW interrupt issued by the driver fail s to arrive. ⢠Carrier Loss PHY Resets This counter indicates the number of times the driver reset the PHY a s a result of carrier sense loss. N1000 and N1000e custom counters ⢠Duplex Mode (1=Half, 2=Full) This counter indicates the current link duplex conf iguration. A value of 1 indicates the adapter is configured to half-duplex mode. A 2 indicates ful l-d uplex mode configuration. This counter exists in v1.12 or higher of the NetWare server driver. ⢠Line Speed (Mbps) This counter displays the cu rrent line speed in Mbps. ⢠Configured Duplex Mode (0=Auto, 1=Half, 2=Full) ⢠Configured Speed (Mbps; 0=Auto) ⢠PCI Mode (0=PCI, 1=PCI-x, 2=PCI-Expres s) ⢠Bus Clock Rate (Mhz) (0=Unavailable) ⢠Link Status (0=Up, 1=Down) This counter describes the current link status of the adapter. A value of 0 indicates the link is up and a value of 1 indicates that the link is down. This counter exists in v1.12 or higher of the NetWare server driver.
Installation 39 ⢠Link Status Change Count This counter indicates the number of ti mes the link status has been changed. ⢠Rx CRC Error Count This counter indicates the number of receive packe ts with CRC errors. Packets less than 64 bytes are not counted in this statistic. ⢠Alignment Errors This counter holds the number of receive packets wi th alignment errors. In order for a packet to be counted, it must pass address filtering and must be 64 bytes or greater in length. If receives are not enabled, then this counter does not increment. This counter is valid only in MII mode during 10/100 Mbps operation. ⢠Rx Symbol Error Count This counter indicates the number of symbol erro rs between reads. The coun t increases for every bad symbol received, whether or not a packet is currently being received and whether or not the link is up. This counter will increment along with the Rx Se quence Error Count when eith er the fiber cable is disconnected or the connector is not seated completely. Also, during driver load, this counter will increments minimally. ⢠Rx Error Count This counter indicates the number of packets receiv ed in which I_RX_ER was as serted by the PHY. In order for a packet to be counted, it must pass a ddress filtering and must be 64 bytes or greater (from <Destination Address> through <CRC>, inclusivel y) in length. If receives are not enabled, then this counter does not increment. In TBI mode, this counter increments on the reception of /V/ codes. ⢠Rx Missed Packets Count This counter indicates the number of missed packets. Packets are mi ssed when the hardware receive FIFO has insufficient space to stor e the incoming pa cket. This could be the result of having too few buffers allocated or because there is insuffici ent ban dwidth on the IO bus. These packets are also be counted in the Total Packets Received counte r as well as in Tota l Octets Received. ⢠Tx Single Collision Count This counter indicates the number of times that a transmitted packet encounte red a single collision. This counter increments only if th e device is in half-duplex mode. ⢠Tx Excessive Collisions Count This counter indicates when 16 or more collisi ons have occurred on a packet. Thi s counter increments only if the device is in half-d uplex mode. ⢠Tx Multiple Collision This counter indicates the number of times that a transmitted packet encountered more than one collision but less than 16. This counter incremen ts only if the device is in half-duplex mode. ⢠Tx Late Collisions Count This counter indicates collisions that occur after 512 bit (64 byte) tim es. This counter increments only if the device is in half-duplex mode. ⢠Tx TOTAL Collision This counter indicates the total number of collisions seen by the transmitter. This counter increm ents only if the device is in half-duplex mode. ⢠Tx DMA Underruns
Installation 40 This counter contains the number of frames that we re either not transmitted or re-transmitted due to a Tx DMA under-run. If the device is configured to retransmit on under-run, this counter may be updated multiple times for a single frame. Un der-runs occur because a lack of PCI bandwidth resulted in the device's internal transmit FIFO running dry during the tr ansmission of a frame. The adaptive algorithms within the soft ware keep this number under control. To see how the adaptive algorithm is doing, refer to the "Current Threshold" section. ⢠Tx Defer Count This counter indicates transmit defer events. A defer event occurs wh en the transmitter cannot immediately send a packet due to the medium being busy, either because another device is transmitting or the link is not up. ⢠Tx with No CRS This counter counts the number of su ccessful packe t transmissions in which the Carrier Sense input from the PHY was not asserted within one slot time of start of transmission. The PHY should as sert Carrier Sense during every transmissi on. Failure to do so may indicate that the link has failed, or the PHY has an incorrect link configuration. This counte r is not valid in TBI mode, and is only valid when the device is operating at full-duplex. ⢠Rx Sequence Error Count This counter indicates receive sequence error even ts. This counter increments along with the Rx Symbol Error Count when either the fiber cable is disconnected or the connector is not seated completely. ⢠Rx Length Error Count This counter indicates receive length error events. A length error occurs if an incoming packet is undersized (less than 64 bytes) or oversized (greater than 1522). ⢠Rx DMA Too Early This counter counts the total number of times that the device attempted to perform an Early DMA operation (by speculatively requesting the host bus prior to reception of th e entire packet) and ran out of data to DMA before the device was forced to relinquish owners hip of the bus. This condition is really not an "error" condition, bu t is nevertheless counte d for diagnostic purpos es. This statistic is valid only if Early Receives are enabled. ⢠XON Received Count This counter indicates the number of XON packe ts received. XON packets can occur when Flow Control is enabled and both the adapter and its link partner are configured to use it. ⢠XON Transmitted Count This counter indicates the number of XON packets tra nsmitted. XON packets can occur when Flow Control is enabled and both the adapter and it s link partner are configured to use it. ⢠XOFF Received Count This counter indicates the number of XOFF packe ts received. XOFF packets can occur when Flow Control is enabled and both the adapter and it s link partner are configured to use it. ⢠XOFF Transmitted Count This counter indicates the number of XOFF packets transmitted. XOFF packets can occur when Flow Control is enabled and both the adapter and it s link partner are configured to use it. ⢠Flow Control Received Unsupported Count
Installation 41 This counter indicates the number of unsupported fl ow control frames that are received. This counter increments when a flow control packet is received which does not match the adapter definition of a flow control packet. This counter is not expected to increment. ⢠64 Byte Packets Received This counter indicates the number of good packets received that are exactly 64 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠65â127 Byte Packets Received This counter indicates the number of good packe ts received that are 65â127 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠128â255 Byte Packets Re ceived This counter indicates the number of good pa ckets received that are 128â255 bytes in len gth, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠256â511 Byte Packets Re ceived This counter indicates the number of good pa ckets received that are 256â511 bytes in len gth, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠512â1023 Byte Packets Re ceived This counter indicates the number of good packe ts received that are 512â1023 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠>1024 Byte Pa ckets Received This counter indicates the number of good packe ts received that are 1024â1522 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠Good Packets Received Count (Any Length) This counter indicates the number of good packets received of any legal length. This cou nter does not include received flow control packets. ⢠Broadcast Packets Received Count This counter indicates the number of good broadcast p ackets received. ⢠Multicast Packets Received Count This counter indicates the number of good multicas t packets received. This counter does not include received flow control packets. ⢠Good Packets Transmitted Count (Any Length) This counter indicates the number of good packets transmitted of any legal le ngth. This cou nter does not include received flow control packets. ⢠Good Octets Received Count LO This counter, in conjunction with the Good Octets Received Count HI, makes up a 64-bit counter that indicates the number of good octets received. This statistic reflects the lower 32 bits of the 64-bit counter. This counter does not incl ude received flow control packets.
Installation 42 ⢠Good Octets Received Count HI This counter, in conjunction with the Good Octe ts Received Count LO, makes up a 64-bit counter that indicates the number of good octets received. Th is counter reflects the uppe r 32 bits of the 64-bit counter and is incremented each time the Good Oc tets Received Count LO has reached it maximum value and wraps to zero. This counter does not include received flow control packets. ⢠Good Octets Transmitted Count LO This counter, in conjunction with the Good Octe ts Transmitted Count HI, makes up a 64-bit counter that indicates the number of good octets transmitted. This counter reflects the lower 32 bits of the 64- bit counter. This counter does not in clude received flow control packets. ⢠Good Octets Transmitted Count HI This counter, in conjunction with the Good Octe ts Transmitted Count LO, makes up a 64-bit counter that indicates the number of good octets transmitte d. This counter reflects the upper 32 bits of the 64- bit counter and is incremented each time the G ood Octets Transmitted Count LO has reached it maximum value and wraps to zero. This counter do es not include received flow control packets. ⢠Receive No Buffers Count This counter indicates the number of times that pa ckets were received when there were no available buffers in host memory to store tho se packets. The packet is still received if there is space in the hardware FIFO. This counter does not include received flow control packets. Increasing Rx Descriptors on the command line duri ng load could result in this co unter incrementing less frequently. ⢠Receive Undersize Count This counter indicates the number of received packe ts that were less than minimum size (64 bytes) and had a valid CRC. ⢠Receive Fragment Count This counter indicates the number of received packe ts that were less than minimum size (64 bytes) but had a bad CRC. ⢠Receive Oversize Count This counter indicates the number of received pa ckets that were greater than maximum size (1522 bytes) and had a valid CRC. ⢠Receive Jabber Count This counter indicates the number of received pa ckets that were greater than maximum size (1522 bytes) but had a bad CRC. ⢠Total Octets Received LO This counter, in conjunction with the Total Octe ts Received HI, makes up a 64-bit counter that indicates the total number of octets received. This counter reflects the lower 32 bits of the 64-bit counter. ⢠Total Octets Received HI This counter, in conjunction with the Total Octe ts Received LO, makes up a 64-bit counter that indicates the total number of octets received. This counter reflects the upper 32 bits of the 64-bit counter and is incremented each time the Total Oc tets Received LO has reached it maximum value and wraps to zero. ⢠Total Octets Transmitted LO
Installation 43 This counter, in conjunction with the Total Octe ts Transmitted HI, makes up a 64-bit counter that indicates the total number of octets transmitted. This counter reflects the lower 32 bits of the 64-bit counter. ⢠Total Octets Transmitted HI This counter, in conjunction with the Total Octe ts Transmitted LO, makes up a 64-bit counter that indicates the total number of octets transmitted. This counter reflects the upper 32 bits of the 64-bit counter and is incremented each time the Total Oc tets Transmitted LO has reached it maximum value and wraps to zero. ⢠Total Packets Received This counter indicates the total nu mber of packets received. All pack ets are counted here, regardless of their length, whether they are erred, or whether they are flow contr ol packets. ⢠Total Packets Transmitted This counter indicates the total nu mber of packets transmitted. All packets transmitted will be counted here, regardless of their length or wh ether they are flow control packets. ⢠64 Byte Packets Transmitted This counter indicates the number of good packets tran smitted that are exactly 64 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠65â127 Byte Packets Transmitted This counter indicates the number of good packets tran smitted that are 65â127 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠128â255 Byte Packets Trans mitted This counter indicates the number of good packe ts tran smitted that are 128â255 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠256â511 Byte Packets Trans mitted This counter indicates the number of good packe ts tran smitted that are 256â511 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠512â1023 Byte Packets Transmitted This counter indicates the number of good packe ts transmitted that are 512â1023 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠>1024 Byte Packets Transmitted This counter indicates the number of good packets tran smitted that are 1024â1522 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠Broadcast Packets Transmitted Count This counter indicates the number of good broadcast packets transmitted. ⢠Multicast Packets Transmitted Count This counter indicates the number of good multic ast packets transmitted. This cou nter does not include transmitted flow control packets. ⢠Rx Checksum Good Pkts This counter indicates the number of receive packets with good checks ums calculated. This counter increments only when RxTcpChecksum has been enab led and packets with good checksums are received. ⢠Rx Checksum Error Pkts
Installation 44 This counter indicates the number of receive packe ts with CRC errors. Packets less than 64 bytes are not counted in this counter. ⢠Tx Threshold This counter indicates the current value of the Early Transmit Threshold. The adapter starts transmitting the packet when the specified number of quad-words has been DMAed in from memory. ⢠Current IFS Value This counter indicates the current value of Inte r Frame Spacing between back-to-back transmits. ⢠Jumbo Frames Received This counter indicates the number of jumbo packets received (packets greater than 1514 in len gth). ⢠Jumbo Frames Transmitted This counter indicates the number of jumbo packets transmitted (pa ckets that are greater than 1514 in length). ⢠PCI-X mode (0=PCI, 1=PCI-X) This counter indicates whether the adapter is in a PCI. CPQANS custom counters ⢠Number of Tx ECBs Queued in Members This counter indicates the number of packets (Event Control Blocks) that are queued in team members for transmission. ⢠Number of Tx ECBs Queued in Unavailable ECBs Queue This counter indicates the number of transmit packets (Event Control Blocks) that are currently in the CPQANS transmit queue. ⢠Total Enqueued Tx Due to No Available ECBs This counter indicates the number of transmit packe ts that were queued du e to no available ECBs. ⢠Number of Probes Sent by All Team Members This counter indicates the number of pr obe packets sent by all team members. ⢠Number of Probes Received by All Team Members This counter indicates the number of prob e packets received by all team members. Advanced Network Services help - supported keywords ⢠AGG_SELECTION o Syntax: AGG_SELECTION=[BANDWID TH | COUNT] o Description: Sets active aggregator selection mode by bandwidth or count ⢠BALANCE_INTERVAL o Syntax: balance_interval=nnn o Description: Changes Balance interval, nnn counts 1/18 sec ⢠BALANCE_SET_DEFAULT o Syntax: balance_set_default o Description: Restores Balance interval to factory settin gs ⢠COMMIT o Syntax: commit [team=nnn]
Installation 45 o Description: Sets a certain mode. Use after binding to base drivers with Team=nnn. ⢠DELAY o Syntax: delay=nnn o Description: Delays the commit in nnn seconds ⢠FRAME o Syntax: frame=[ETHERNET_802.2 | ETHERNET _802. 3 | ETHERNET_II | ETHERNET_SNAP] ⢠HELP o Syntax: -H o Description: Displays Help o Example: CPQANS -H ⢠JOIN_INDIVIDUALS o Syntax: join_individuals=[yes | no] o Description: Joins all individual links to one aggregator ⢠LBN o Syntax: Reset LBN=nnn o Example: cpqans reset lbn=nnn o Description: Resets the logical board number (LBN) of a bound adapter. Supported only after commit command. Default value = 0 ⢠MAX_TX_QUEUE o Syntax: max_tx_queue=nnn o Description: Sets the Max number of TX ECBs queued for send ⢠MODE o Syntax: mode=[NFT | ALB | FEC | GEC | 802.3AD] ⢠NAME o Syntax: name=[any unique name] o Description: Sets a unique name ⢠PRIMARY o Syntax: primary | secondary o Description: Identifies the primary a dapter. Supported only in BIND command ⢠PROBE_BURST_SIZE o Syntax: probe_burst_size=nnn o Description: Changes number of probes to send in a retry ⢠PROBE_CHECK_INTERVAL o Syntax: probe_check_interval=nnn o Description: Changes probes check interval ⢠PROBE_RECHECK_INTERVAL o Syntax: probe_recheck_interval=nnn o Description: Changes probe retries check interval
Installation 46 ⢠PROBE_RETRY_COUNT o Syntax: probe_retry_count=nnn o Description: Changes probes retry count ⢠PROBE_SEND_INTERVAL o Syntax: probe_send_interval=nnn o Description: Changes probes send interval, nnn counts 1/18 sec ⢠PROBE_SET_DEFAULT o Syntax: probe_set_default o Description: Restores probes settings to f actory settings ⢠PROBES o Syntax: probes=[on | off] o Description: Enables /disables probes ⢠PROBES o Syntax: probes=[BROADCAST|MULTICAST] o Description: Changes probes addressing ⢠RECOMMIT o Syntax: recommit [team=nnn] o Description: Resets a certai n mode. Use after hot binding to base drivers with Team=nnn ⢠REMOVETEAM o Syntax: removeteam [team=nnn] o Description: Removes a team. Use after hot binding to base drivers with Team=nnn ⢠REMOVEVLANID o Syntax: RemoveVlanID=nnn o Description: Removes the selected VLAN ⢠RESET o Syntax: RESET LBN=nnn o Description: Supported on ly after Commit command ⢠SECONDARY o Syntax: primary | secondary o Description: Identifies the secondary a dapter. Supported only in BIND command. ⢠SMPMODE o Syntax: SMPMODE=[SMP | NONSMP] o Description: Enables /disables SMP aware ⢠STATUS o Syntax: status [team=nnn] o Description: Prints CPQANS status ⢠TEAM o Syntax: team=nnnn
Installation 47 o Description: Identifies the team. nnnn=DecimalNumber. ⢠TX_ECBS_TO _USE o Syntax: tx_ecbs_to_use=nnn o Description: Sets number of TX EC Bs to allocate per virtual adapter ⢠VLANID o Syntax: VlanID=nnn o Description: Sets team to VLAN mode. Creates MLID edge
Firmware upgrade utility 48 F ir m w ar e upgr ade utilit y In this section Windows ut ility ................................................................................................................ ...................... 48 c-Class BladeSys tem utility .................................................................................................... ................... 54 Linux utility .................................................................................................................. .......................... 54 DOS utility .................................................................................................................... ......................... 59 Windows utility These notes describe how to us e the Gigabit Server Adapter Firmware Upgrade Utility to flash server adapter firmware. Flashing consists of using software to replace the current server adapter firmware on a target server with a new firmware image. Server adapter firmware is typically flashed for the following reasons: ⢠To support new features ⢠To correct problems in a previous server adapter firmware version Customers should up grade their servers to the latest server adapter firmware version: ⢠When HP releases a new serv er adapter firmware version ⢠When installing a new server adapter ⢠After purchasing a new server ⢠When installing a SoftPaq that specifies th e need for the latest version of firmware The Gigabit Server Adapter Firmwa re Upgrade Utility automatically checks for driver, hardware, and operating system dependencies, and installs only the correct s erver adapter firmware upgrades required by each target server. Supported operating systems The upgrade utility support s the following operating system s: ⢠Microsoft® Windows® 2 000 Server ⢠Microsoft® Windows® 2000 Advanced Server ⢠Microsoft® Windows® 2000 Data Center ⢠Microsoft® Windows® 2 000 Terminal Server ⢠Microsoft® Windows 20 03 Server ⢠Microsoft® Windows 20 03 (32-bit) Enterprise Serv er ⢠Microsoft® Windows 20 03 (32-bit) Data Center ⢠Microsoft® Windows 20 03 (x64) Enterprise Server Supported Gigabit Server Adapters The upgrade utility provides support for t he following adapters:
Firmware upgrade utility 49 Standup adapters: ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC364T PCI Express Quad Port Gigabit Server Adapter ⢠HP NC370T PCI-X Multifunction Gigabit Server Adapter ⢠HP NC370F PCI-X Multifun ctio n Gigabit Server Adapter ⢠HP NC373F PCI Express Multifunc tion Gigabit Server Adapter ⢠HP NC373T PCI Express Multifunction Gigabit Serv er Adapter ⢠HP NC380T PCI Express Dual Port Multifunction Giga bit Server Adapter ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC320i Integrated PCI Express G igabit Server Adapter ⢠HP NC324i Integrated Dual Port PC I Express Gigabit Server Adapter ⢠HP NC325i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC326i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC370i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC371i Integrated PCI-X Mult ifunction Gigabit Server Adapter ⢠HP NC373i PCI Express Multifunction Gigabit Server Adapter ⢠HP NC7760 PCI-X Gigabit Server Adapter ⢠HP NC7761 PCI Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gigabit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter Mezzanine Adapters: ⢠HP NC320m PCI Express Dual Port 1Gb Server Adapter ⢠HP NC325m PCI Express Quad Port 1Gb Server Ad apter for c-Class BladeSystem ⢠HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem ⢠HP NC373m PCI Express Dual Port Multifunction Giga bit Server Adapter for c-Class BladeSyst em ⢠HP NC374m PCI Express Dual Port Multifunction Giga bit Server Adapter IMPORTANT: For each server adapter in the target server, the driver for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated.
Firmware upgrade utility 50 IMPORTANT: Before updating NC6770 or NC7770 server adapter s, move jumper J1 to the Enable position. After updating each server adapter, move jumper J1 to the Disable position. NOTE: NC6770, NC7770, and NC7780 server adapters require about 15 mi nutes each to update both the PXE and Boot Code. NOTE: To update the PXE code for embedded and mezzanine adapters, update the system BIOS. Drivers For each server adapter in the target server, the driv er for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. Installing the firmware utility The HP Network Adapter Online Firmware Upgrade Ut ility is located in the \apps\fwupgrade\windo ws directory obtained by extracting the HP SoftPaq fi le. Complete the following step s to download and extract the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. The upgrade utility installation is launched by clicking the Smart Component (CPXXXXXX.exe) file (interactive mode) or by using the command line pr ompt. The Smart Component for the upgrade utility is launched using the command line prompt to allow yo u to pass arguments to alter installation behavior. Command line syntax The general command line syntax for the Smart Component installation includes: CPXXXXXX [/S[ILENT]] [/F[ORCE]] [/R[EBOOT]] [/H[ELP]] [/?] CPXXXXXX is the command that launches the Smart Compon ent on the target server. The six X's represent the serialized number assigned to the component. Command line arguments The Smart Component recognizes the following comm and line arguments. Smart Components are not case-sensitive. NOTE: All arguments and information en closed in brackets are optional.
Firmware upgrade utility 51 Command line argument Description /S[ILENT] Supresses console messages. If this argument is omitted from the command line, console messages are displayed. /F[ORCE] Specifies whether an installation will be forced, regardless of the server adapter firmware version that is detected on the target server. The argument changes the Smart Component installation in the following ways: ⢠If the server adapter firmware on the target server is current, the firmware reinstalls itself, and the installed version number remains the same. ⢠If a newer version of the server adapter firmware is already installed on the target server, the component installs itself and downgrades the server adapter firmware to the older version number. If this argument is omitted from the command line, the installation is not forced. /R[EBOOT] Specifies whether the target server will reboot if the Smart Component requests a reboot. If this argument is omitted from the command line and the Smart Component requests a reboot, the server must be rebooted manually for th e server adapter firmware upgrade to take effect. The Windows server adapter upgrade component will NOT request a reboot. /H[ELP] Displays command line Help information. /? Displays command line Help information. Thi s argument is identical to the /H[ELP] argument. Command Line Examples The following are examples of how to use the CPXXXXXX.exe command. Command line input Result CPXXXXXX Deploys the CPXXXXXX.EX E Smart Component on the target server using the defaults of the component. CPXXXXXX /F Deploys the CPXXXXXX. E XE Smart Component on the target server, forcing the installation of the server adapter firmware. CPXXXXXX /S Deploys the CPXXXXXX. E XE Smart Component on the target server using the defaults of the component. All console messages are suppressed. CPXXXXXX /S /F /R Deploys the CPXXXXXX.EXE Smart Component on the target server, forcing the installation of the server adapter firmware and allowing the server to reboot automatically. All console messages are suppressed. Common installation log file
Firmware upgrade utility 52 The upgrade installation activity (including erro rs) is written by the Smart Component to a common installation log file called CPQSETUP.LOG in the \CPQSYSTEM\LOG subdirector y on the target server. Information regarding subsequent installation activity is appended to the same log file, which provides a chronological history of all component inst allation activity on the target server. Using the NicFwUpg.exe command line interface It is possible to launch the HP Online Firmware Up grade Utility by clicking the executable file d uring installation or by using the command line interface. The following in formation describes how to launch the utility using the command line interface. Command line syntax The utility supports the follo wing commands: Command Line Input Result nicfwupg.exe -c Displays the cu rrent firmware version of all HP NC-Series Broadcom adapters in the system nicfwupg.exe -l Displays the latest version of HP NC-Series Broadcom adapters in the system as determined by versions shipped with the component. nicfwupg.exe -a Retuns the saved image version of HP NC- Series Broadcom adapter s in the system nicfwupg.exe -s Suppresses console output nicfwupg.exe -f Forces downgr ades or restores as needed nicfwupg.exe -p Requests a restore on all adapters nicfwupg.exe -e [MAC Address] Requ ests a restore on one adapter. Supply the MAC address for the adapter. nicfwupg.exe -o [MAC Address] Performs an upgrade of boot code for one adapter. Supply the MAC address for the adapter. nicfwupg.exe -x [MAC Address] Performs an upgrade of PXE for one adapter. Supply the MAC address for the adapter. nicfwupg.exe -b [MAC Address] Performs an upgrade of CLP for one adapter. Supply the MAC address for the adapter. nicfwupg.exe -i [filename] Performs an upgrade and stores the image file to be used during a restore. Supply the filename. Using the nicfwup_mf.exe command line interface It is possible to launch the HP Online Firmware Up grade Utility for Multifunction adapters by clicking the executable file during installation or by using the command line interface. The fo llowing information describes how to launch the utility for Multifunction network adapters using the command line interface. Command line syntax The utility supports the follo wing commands:
Firmware upgrade utility 53 Command Line Input Result nicfwupg_mf.exe -c Displays the current firmware version of all HP NC-Series Broadcom adapters in the system nicfwupg_mf.exe -l Displays the latest version of HP NC-Series Broadcom adapters in the system as determined by versions shipped with the component. nicfwupg_mf.exe -a Retuns the saved image version of HP NC- Series Broadcom adapter s in the system nicfwupg_mf.exe -s Suppresses console output nicfwupg_mf.exe -f Forces down grades or restores as needed nicfwupg_mf.exe -p Requests a restore on all adapters nicfwupg_mf.exe -e [MAC Address] Requests a restore on one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -o [MAC Address] Performs an upgrade of boot code for one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -x [MAC Address] Performs an upgrade of PXE for one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -b [MAC Address] Performs an upgrade of CLP for one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -i [filename] Perfor ms an upgrade and stores the image file to be used during a restore. Supply the filename. Return codes When the upgrade utility has fini shed running, the component reports a return code to the operating system or calling application. These return codes ar e processed by the %error level% operating system variable and are used to determine the stat us of a firmwa re upgrade proc edure. The following summarizes the ret urn codes for the Sm art Component installer and HP Online Firmware Upgrade Utility commands. Error level Meaning 0 Upgrade component failed to install 1 Upgrade component installed successfully 3 Upgrade component did not de tect any supported network adapters 4 Firmware upgrade was recommended, but the upgrade component was instructed to not update the firmware 5 All firmware versions are current and an upgrade is not recommended 6 At least one network adapter has firmware installed that is newer than the available firmware.
Firmware upgrade utility 54 c-Class BladeSystem utility The HP Gigabit Server Adapter Firmware Upgrade Ut ility for c-Class BladeSystem uses two separate utilities (ccfwupg1.bat and ccfwupg2.bat) to upgr ade firmware on the c-Cl ass BladeSystem adapters. The ccfwupg1.bat command upgrades the Boot Code an d Option ROM fi rmware image on the following c-Class Multifunction Gigabit Server Adapters: ⢠HP NC373i Multifunction Gigabit Server Adapter ⢠HP NC373m Multifunction Gigabit Server Adapter The ccfwupg2.bat command upgrades the Boot Code an d Option ROM fi rmware image on the following c-Class Gigabit Server Adapters: ⢠HP NC326i PCIe Dual Port Gigabit Server Adapter ⢠HP NC326m PCIe Dual Port Gigabit Server Adapter ⢠HP NC325m PCIe Quad Port Gigabit Server Adapter Upgrading firmware To upgrade the boot code and Option ROM im age on the adapter, complete the following: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. 8. Copy all files and folders from the \apps\ccfwupg\ccfwupg1 dir ectory or the \apps\ccfwupg\ccfwupg2 directory to a bootable diskette or USB drive. 9. Boot to DOS using the diskette for your a dapter. 10. Type ccfwupg1.bat or ccfwupg2.bat and press the Enter key. A confirmation message displays when the upgrade is complete. Linux utility The following information describes how to use th e Linux Gigabit Server A dapter Firmware Upgrade Utility to flash server adapter firmware. Supported operating systems The firmware upgrade utility support s the following Linux operating syst ems: Linux 32 ⢠Red Hat Enterprise Linux 4 ⢠Red Hat Enterprise Linux 5 ⢠SUSE Linux Enterprise Server 10 x86
Firmware upgrade utility 55 ⢠SUSE Linux Enterprise Server 9 x86 Linux 64 ⢠Red Hat Enterprise Linux 4 for AMD64 and Intel EM64T ⢠Red Hat Enterprise Linux 5 for AMD64 and Intel EM64T ⢠SUSE Linux Enterprise Server 10 for AMD64 and Intel EM64T ⢠SUSE Linux Enterprise Server 9 for AMD64 and Intel EM64T For a complete listing refer to the Linux Readme fil e located in the HP SoftPaq file.Complete the follo wing steps to download and extract the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable with a file name based on the SoftPaq number. 7. The hpnicfwupg Linux online firmware upgrade utilit y is located in the apps \fwupgrade\linux folder. Supported Gigabit Server Adapters The upgrade utility provides support for t he following adapters: Standup adapters: ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC370T PCI-X Multifunction Gigabit Server Adapter ⢠HP NC370F PCI-X Multifun ctio n Gigabit Server Adapter ⢠HP NC373F PCI Express Multifunc tion Gigabit Server Adapter ⢠HP NC373T PCI Express Multifunction Gigabit Serv er Adapter ⢠HP NC380T PCI Express Dual Port Multifunction ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC320i PCI Express Gigabit Server Adapter ⢠HP NC324i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC326i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC370i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC371i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC373i PCI Express Multifunction Gigabit Server Adapter
Firmware upgrade utility 56 ⢠HP NC7760 PCI-X Gigabit Server Adapter ⢠HP NC7761 PCI Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gig abit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter Mezzanine Adapters: ⢠HP NC320m PCI Express Dual Port Gigabit Server Ad apter ⢠HP NC325m PCI Express Quad Port 1Gb Server Ad apter for c-Class BladeSystem ⢠HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem ⢠HP NC373m PCI Express Dual Port Multifunction Giga bit Server Adapter for c-Class BladeSyst em ⢠HP NC374m PCI Express Dual Port Multifunction Giga bit Server Adapter IMPORTANT: For each server adapter in the target server, the driver for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. IMPORTANT: Before updating NC6770 or NC7770 server adapter s, move jumper J1 to the Enable position. After updating each server adapter, move jumper J1 to the Disable position. NOTE: NC6770, NC7770, and NC7780 server adapters require about 15 mi nutes each to update both the PXE and Boot Code. NOTE: To update the PXE code for embedded and mezzanine adapters, update the system BIOS. Drivers For each server adapter in the target server, the driv er for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. To download the latest Linux drivers: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Open the Linux folder to select drivers. Upgrading firmware This firmware upgrade utility updat es Boot code, PXE code, and CLP upgrades for c-Class BladeSystem (NC325m, NC326m and NC373m) serv er adapter images. The utility is released in RPM fo rmat. The file
Firmware upgrade utility 57 names for the packages are hpnicfwupg-<vers ion>.i38 6.rpm for Linux 32 and hpnicfwupg- <version>.x86_64.rpm for Linux 64. The inst ructions b elow describe the x32 version. 1. Before the RPM package is installed, check for th e existence of a previous version by issuing the rpm -q hpnicfwupg command. 2. If a previous version is found, remove it using the rpm -e hpnicfwupg command. 3. Install the RPM package using the hpnicfwupgâ<version>.rpm command. 4. Launch the adapter interface by entering ifup ethx or ifconfig ethx up . 5. Launch the upgrade uti lity in interactive mode b y entering hpnicfwupg . 6. Before the upgrade is attempted, a copy of the firmware image is saved. If the firmware upgrade fails due to an unrecoverable error, you can restor e your original fi rmware interactively by entering /usr/sbin/hpnicfwupg -p . If errors have occurred, use the hp nicfwupg utility to restore the server adapter fir mware from a backup copy. Command line syntax Use the following command line syntax. To Use command Verify previous version of hpnicfwupg package exists # rpm -q hpnicfwupg Remove an old version of hpnicfwupg package # rpm âe hpnicfwupg Install the RPM package # rpm -ivh hpnicfwupg-<version>.rpm Launch the firmware upgrade utility # user/sbin/hpnicfwupg Restore original firmware if errors occur # /usr/sbin/hpnicfwupg -p Uninstall the RPM package # rpm -e hpnicfwupg Command line arguments The upgrade utility recognizes the foll owing command line arguments. Command line argument Description ây Yes. Executes the utility with no user interaction. Default action is implied for all prompts. Messages are displayed to the screen and to the log. If ây is used with âs, the behavior is the same as if âs had been specified. âs Silent. Executes the utility silently with no user interaction. Default action is implied for all prompts. Messages are logged but are not displayed o n the screen. If the ây option is used with âs, ây is ignored. âf Force. Forces the firmwa re upgrade even if the firmware installed on the system is newer than the firmware in the RPM. This opti on is only valid with âs or ây. If neither are specified, the option is ignored.
Firmware upgrade utility 58 Command line argument Description âp Previous. Interactively restores previous firmware images that were automatically backed up prior to the firmware update. This option is useful for recovering from firmware upgrade failures. This option can be used with any other option. âh Help. Displays the usage mess age. If this option is used with any other option, the other options are ig nored. âc Displays server adapters and current firmware version without performing an updates. This option cannot be used with any other option. Restoring firmware image from backup Before the server adapter firmware image on the server is updated, the current image is saved in a file called <MAC_ADDRESS>.bc<Boot Code version>.p<PXE version>.bin . The backup file is saved in the /var/opt/hp/nicfw_backup directory of the target server. Use the hpnicfwupg command with the âp option to restore the original firmware image after an upgrade or after an upgrade failure. This option ca n only be used interactively. The firmware backup directory is examined for saved images and a me ssage is displayed showing the curr ent firmware versions and the versions from the saved image. You will be prompted be fore the restore is attempted. If more than one saved image is found for a particula r server adapter, you will be prompted for each image. The following is a sample screen sh ot of a typical restore. [root@rdple root]# /usr/sbin/hpnicfwupg -p Found HP NC7760 Gigabit Server Adapter MAC: 000BCD5035DE *** WARNING *** - Installed firmware is the same version as the selected firmware. Restore Boot Code 2.35 to 2.35 - y/n/q (n): q Firmware upgrade cancelled by User. Firmware upgrade log file The firmware upgrade installation stat us (including e rrors) is written by the Li nux server adapter upgrade package to a firmware upgrade log file called nicfwupg.log in the /var/opt/hp subdirectory on the target server. Information regarding subsequent inst allation activity is append ed to the same log file, providing a chronological history of all server adapte r firmware upgrade activity on the target server. Return codes When the Linux server adapter firmware upgrade packa ge has finished running, it report s the following return codes to the operating sy stem or calling application. Error level Meaning 0 Hpnicfwupg installed all selected firmware successfully. For options âc and âi , 0 means the information was displayed correctly.
Firmware upgrade utility 59 Error level Meaning 2 Hpnicfwupg did not attempt firmware upgrade because the installed firmware was up-to-date, the target hardware was not present, the command was invalid, the user responded no to all prompts, or the user selected qu it in response to a prompt. For options âc and âi, a 2 means that the information could not be displayed. 3 Hpnicfwupg attempted to upgr ade firmware but one or more upgrades failed due to an unrecoverable error. This return code is used even if some of the server adapters were upgraded successfully. DOS utility These notes describe how to us e the Gigabit Server Adapter Firmware Upgrade Utility to flash server adapter firmware. Flashing consists of using software to replace the current server adapter firmware on a target server with a new firmware image. Server adapter firmware is typically flashed for the following reasons: ⢠To support new features ⢠To correct problems in a previous server adapter firmware version Customers should up grade their servers to th e latest server adapter firmware ve rsion: ⢠When HP releases a new serv er adapter firmware version ⢠When installing a new server adapter ⢠After purchasing a new server ⢠When installing a SoftPaq that specifies th e need for the latest version of firmware The Gigabit Server Adapter Firmware Upgrade Utilit y automatically checks for driver, hardware, and operating system dependencies, and installs only the correct server adapter firmware upgrades required by each target server. Supported Gigabit Server Adapters The upgrade utility provides support for t he following adapters: Standup adapters: ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC7760 PCI-X Gigabit Server Adapter
Firmware upgrade utility 60 ⢠HP NC7761 PCI Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gigabit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter IMPORTANT: For each server adapter in the target server, the driver for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. IMPORTANT: Before updating NC6770 or NC7770 server adapter s, move jumper J1 to the Enable position. After updating each server adapter, move jumper J1 to the Disable position. NOTE: NC6770, NC7770, and NC7780 server adapters require about 15 mi nutes each to update both the PXE and Boot Code. NOTE: To update the PXE code for embedde d adapters update the system BIOS. Drivers For each server adapter in the target server, the driv er for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. Upgrading firmware 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. 8. Click the SoftPaq file (SP31728) located in the \apps\fwupgrade\dos directo ry. 9. Execute the downloaded file and follow the on-scr een instructions. After the files ha ve been unpacked, you may delete the Soft Paq file that you downloaded. 10. Copy these files to a bootable DOS diskette. 7781V235 BIN 150TV318 BIN 6770V219 BIN 7760V237 BIN 7761V318 BIN 7770V219 BIN 7771V235 BIN 7780V219 BIN 1020V318 BIN
Firmware upgrade utility 61 7782327B BIN NICFWUPG BAT PT5704C2 36 PXE_V322 BIN Q57UDIAG EXE SP29044 TXT PXE_V778 BIN PXE_V839 BIN 320TV355 BIN 11. Boot the system using the DOS diskette. 12. Run NICFWUPG.BAT from a command prompt.
Adapter configurations 62 A dapte r conf igur ations In this section HP NC-Series Br oadcom adapters ................................................................................................. ........... 62 HP NC-Series Multif unction adapters ............................................................................................ ............ 66 HP NC-Series In tel adapters.................................................................................................... ................. 69 Windows ........................................................................................................................ ...................... 91 Novell Ne tware ................................................................................................................. .................. 110 Solaris ........................................................................................................................ ........................ 122 HP NC-Series Broadcom adapters Keywords for Q57 NDIS2 driver This document describes the keywords for th e Q57 NDIS 2 driver version 7.65 (Q57.DOS). The following custom keywords are used in PROTOCOL.INI: ⢠BUSNUM ⢠DEVNUM ⢠FUNCNUM or PORTNUM ⢠LineSpeed ⢠Duplex ⢠NodeAddress ⢠FixCheckSumOff Descriptions ⢠BUSNUM and DEVNUMâThese two keywords are used concurrently and have been included for manufacturin g purposes. Do not use the m unless yo u are familiar with PCI device configuration. These two keywords are needed if multiple NC 1020, NC67xx, and/or NC77xx boards are in a system and a specific adapter (or ad apters) needs to be loaded in sp ecific order. BU SNUM specifies the PCI bus number, and DEVNUM specifies the devi ce number giv en to the board when configured by the PCI BIOS. The values are in decimal. ⢠LineSpeedâThis decimal parameter, 10 or 100, spec i fies the speed of the network connection. According to IEEE specifications, line speed of 100 0 cannot be forced and its only achievable by auto negotiation. ⢠DuplexâThis string parameter, HALF o r FULL, specif ies duplex mode on the Ethern et controller. The LineSpeed parameter must be set when this keyword is used. If neither the Duple x nor the LineSpeed parameters are specified the Ethernet controlle r will default to autonegotiate mode. ⢠NodeAddressâThis string parameter sp ecifies the network address us ed by the Ethernet controller. If Multicast Address or Broad cast Address was spec ified , the default MAC Address will be used.
Adapter configurations 63 ⢠FixCheckSumOffâThis strin g parameter turns off the driver work-around for the T CP/IP stack to recognize the 1's complimented version of the checksum. The following are examples of the use of these keyword s in the protocol.ini. [Q57] DriverName = "Q57$" BusNum = 3 DevNum = 14 PortNum = 2 LineSpeed = 100 Duplex = Full NodeAddress = "001020304050" To add more adapters (up to four), repeat the entry below, where n can be from 1 to 3. [Q57_n] DriverName = "Q57$" BusNum = DevNum = The following examples use additional entries to lo ad more than one adapter in the protocol.ini. ⢠[Q57] DriverName = "Q57$" BusNum = 3 DevNum = 10 ⢠[Q57_1] DriverName = "Q57$" BusNum = 3 DevNum = 11 ⢠[Q57_2] DriverName = "Q57$" BusNum = 3 DevNum = 12 ⢠[Q57_3] DriverName = "Q57$" BusNum = 3 DevNum = 13 Keywords for B06 NDIS2 driver This document describes the keywords fo r the B06 BNDIS 2 driver (bxnd20x.DOS). The following keywords are supported through the PROTOCOL.INI fil e. If a keyword is not present in the PROTOCOL.INI, the default value is chosen.
Adapter configurations 64 ⢠DRIVERNAME o Possible value: BXND20X$ o This field is case sensitive ⢠LINESPEED o Possible values: 10, 100, or 1000 o Default: AUTONEG ⢠BUSNUM o Possible values: the bus slot numb er the NC37xx/N C380x device is on o Default: the bus slot number occupied by one of NC37xx/NC380x devices in the system ⢠DEVNUM o Possible value: the de vice number the NC37xx/NC380x device is on o Default: the device number occupied by one of NC37xx/NC380x devices in the system ⢠FUNCNUM o Possible values: the fu nction number the NC37x/NC380x device is on o Default: the function number occupied by one of NC37x/NC380x devices in the system ⢠DUPLEX o Possible values: half or full (LineSpeed is required when this keywor d is used) o Default: full ⢠NODEADDRESS o Possible values: 6 hex byte s looking like "0A0B0C0D0E0F" o Default: the MAC address for the selected NC37xx/NC380x device Known limitations: For Windows Server 2003 or Window Server 2003 x 64 se lf-installation, a smartdrv driver needs to be loaded. Otherwise, bad performance or even incorrect I/O write errors might occur. PXE configuration: HP NC-series Broadcom and Multifunction adapters NOTE: PXE is not supported on the NC150T adapter. The Pre-boot Execution Environment (PXE) of HP adap ters has been upgraded to allow PXE functions to work with the HP ProLiant servers. The HP NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters have PXE support. PXE is a software protocol that allows your networked computer to boot with the images provided by remote servers across the network. PXE operates in a client/server environment. A network consists of one or more boot serv ers that provide boot images to multiple computers through the network. The PXE driver complies with the PXE-2.1 specification. The PXE driver is released with both mo nolithic and split binary images. This feature provides flexibi l ity to users in different environments in whic h the motherboard may or may not have built-in base-code.
Adapter configurations 65 HP PXE implementation is supported in the followin g environments: ⢠Linux Red Hat PXE Server. HP PXE clients are able to remote-boot and utilize network resources (for example, NFS mount) and to do Linux installation. In the case of a remote boot, the Linux u niversal driver binds seamlessly with the HP UNDI (Unive rsal Network Driver Interface) and provides a network interface in the Linux remote-booted client en vironment. ⢠Microsoft® Windows® 2000, Windo ws 2003, and Windows 2003x64 Remote Installation Services (RIS). HP PXE clients are able to install from the network. ⢠Intel APITEST. The HP PXE driver passes all API compliance test suites. ⢠DOS UNDI. Universal NDIS seamlessly binds with HP UNDI to provide NDIS2 interface to the upper- layer protocol stack. This allows computers to co nnect to network resources in a DOS environment. ⢠3Com® Boot Server. Client setup Setting up PXE client for NC32xx, NC37xx, NC 380x, NC1020, NC6770, and NC77xx Gig abit Server Adapters involves enabling/disablin g the PXE feature and setting up the BIOS for the boot order. Boot method (NC32xx, NC37xx , NC380x, NC1020, NC6770 , NC7770, and NC7771 only) By default, PXE detects if the BIOS supports BBS (BIOS Boot Specification). If the BIOS supports BBS, then PXE uses BBS as the boot method. If the BIOS does not support BBS, then PXE uses Int18h instea d. However, for some machines equipped with an old BI OS, this process may not work. In case of an old BIOS, users must force PXE to use Int18h, Int19h, or BBS as the boot method. To specify the boot method, press the following keys wh en you see the startup banner. Broadcom Gigabit Ethernet Boot Agent 2.2.6 Copyright (c) 2000 Broadcom Corporation All rights reserved ⢠For Int18h: Press the Ctrl F8 keys ⢠For Int19h: Press the Ctrl F9 keys ⢠For BBS: Press the Ctrl F10 keys Setup BIOS To boot from the network with PXE, make PXE the first bootable device under BIOS. This procedure depends on server BIOS implementation. Refer to the user manual for the server for additional information. Server setup Refer to the following instr uctions for your operating system. Windows® 2000 The current version of Windows® 2000 does not include a network driver for th e NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters. To perform remote installation with PXE, include a network driver for these adapters as part of the client's installation image on the server. Refer to Microsoft® Article ID Q246184, "How to Add Third-Party OEM Network Adap ters to RIS Installations." DOS UNDI/APITEST
Adapter configurations 66 To boot in DOS and connect to a network for the DO S environment, download the Intel® PXE PDK from the Intel® web site. This PXE PDK comes with a TFTP/ProxyDHCP/Boot server. The PXE PDK can be downloaded from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). Red Hat Linux Red Hat Linux 7.2 (or higher) distribution has PXE serv er support. It allows users to perform a complete Linux installation over the network. Distribution also comes with boot imagesâboo t kernel (vmlinux) and initial ram disk (initrd). These two images can be found in the HP SoftPaq file. ⢠/misc/src/trees/boot/vmlinux ⢠/misc/src/trees/initrd-netw ork.img Refer to Red Hat documentation for in structio ns on installing PXE ser ver on Linux. Initrd.img, distributed with Red Ha t 7.2, does not have a Linux netw ork driver for the NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC32 6m, NC10 20, NC67xx, and NC77xx Gigabit Server Adapters. ⢠/images/pxeboot/vmlinux ⢠/images/pxeboot/initrd.img Intel® has also created patches to the Linux PXE serv er code to allow clients to perform a remote boot. You can download these patches from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). A remote boot does not require a standard Linux network driver for t he HP NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC326m, NC102 0, NC67 xx, and NC77xx Gigabit Server Adapters. After the PXE client downloads the Linux kernel and in itial ram disk, the Linux universal driv er, that came with the Linux distribution, will bind with the UNDI code of the PXE to form a Linux network driver. HP NC-Series Multifunction adapters Accelerated iSCSI devices The iSCSI protocol transports block-level storage requests over TCP connection s. Using the iSCSI protocol, systems can connect to rem ote storage and use it as a physical disk (although the remote storage provider, or "target," may actually be providing "virtual" physical disks). HP Multifunction Gigabit Server Adapters can be en abled as virtual iSCSI host bus adapters (HBAs), which provide the ability to run iSCSI protocol over offloaded TCP connections . By offloading the TCP connection processing to a Multifunction Server Adapt er the CPU utilization is reduced, which provides additional processing capacity for applications. iSCSI is enabled and network settings are config ured for the iSCSI device through the Network Configuration Utility (NCU). Configuration is limited to IP configuration (including DHCP support), user- specified MAC address, and VL AN ID (only one allowed for the iSCSI device). Prerequisites To use the Multifunction Server Adapter iSCSI accelera ting feature, Microsoft® iS CSI Software Initiator version 2.0 or later must be installed on your sy stem . This software is required to configure the iS CSI protocol, including discovery portals, iSCSI targets, authentication, and so on. The command line utility iscsicli.exe , which is part of Microsoft iSCSI Software In itiator can also be used to configure iSCSI
Adapter configurations 67 protocol. See the Microsoft website ( http://www.microsoft.com ) to d ownload the Microsoft iSCSI Software Initiator installation wizard software an d to obtain installa tion and usage information. The accelerating iSCSI feature comes standard wi th standup HP Multifunction Server Adapters. Accelerating iSCSI capabili ties are available for embedded HP Mu ltifunction Gigabit Server Adapters through the purchase and installation of an HP ProLiant Essentials Accelerated iSCSI Pack license (" Accelerated iSCSI for embedded HP Multifunction server adapters " on page 153 ). Configuring advanced settings parameters If iSCSI is enabled on a Multifunction adapter port, you can adjust the SRB (iSC SI Request Block) Timeout Delta setting and create an iSCSI VLAN for the adap ter. Increasing this value may help prevent I/O timeouts if your network is highly congested and packets are being dropped frequently. iSCSI adapter information includes the following: o Current MAC Address. The current MAC address for the iSCSI device. o iSCSI Driver Name. The name of the driver for the iSCSI device o iSCSI Driver Version. The version of the driver for the iSCSI device. o Default Gateway. The default route for the iSCSI device. o Subnet Mask. The subnet mask (dotted decimal) for the iSCSI device. o iSCSI VLAN Id. The identification number (if any) for the virtual iSCSI VLAN device. o iSCSI Packet Priority. The packet priority for the iSCSI device. o SRB Timeout Delta. The increase to the timeout value for I/O operat ions. o Page File Created. Indicates whether or not the Windows page file was created on an iSCSI drive that is attached to the adapter. The values are Yes and No. o DHCP (Dynamic Host Configuration Protocol). Enabled indicates that DHCP is used to acquire IP configuration; Disabled indicates that the stat ically configured IP conf iguration is used. o IP Address. The IPv4 address (dotted decimal) for the iSCSI device. iSCSI boot configuration: Multifunction adapters HP iSCSI boot is currently supported on Linux operating syst ems only. The HP iSCSI boot feature allows you to b oot from a remote disk (known as the iSCSI target) on a Storage Area Network (SAN) without having to directly attach a boot disk. Booting from a remote disk on a SAN allows you to centralize the boot process and consolidate equipment resources. Unlike other implem entations, iS CSI boot does not require a separate DHCP server or a PXE server. For more information about HP iSCSI Boot refer to the HP iSCSI Boot for Linux User Guide . PXE configuration: HP NC-series Broadcom and Multifunction adapters NOTE: PXE is not supported on the NC150T adapter. The Pre-boot Execution Environment (PXE) of HP adap ters has been upgraded to allow PXE functions to work with the HP ProLiant servers.
Adapter configurations 68 The HP NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters have PXE support. PXE is a software protocol that allows your networked computer to boot with the images provided by remote servers across the network. PXE operates in a client/server environment. A network consists of one or more boot serv ers that provide boot images to multiple computers through the network. The PXE driver complies with the PXE-2.1 specification. The PXE driver is released with both mo nolithic and split binary images. This feature provides flexibi l ity to users in different environments in whic h the motherboard may or may not have built-in base-code. HP PXE implementation is supported in the followin g environments: ⢠Linux Red Hat PXE Server. HP PXE clients are able to remote-boot and utilize network resources (for example, NFS mount) and to do Linux installation. In the case of a remote boot, the Linux u niversal driver binds seamlessly with the HP UNDI (Unive rsal Network Driver Interface) and provides a network interface in the Linux remote-booted client en vironment. ⢠Microsoft® Windows® 2 000, Windows 2003, and Windows 2003x64 Remote Installation Services (RIS). HP PXE clients are able to install from the network. ⢠Intel APITEST. The HP PXE driver pas ses all API compliance test suites. ⢠DOS UNDI. Universal NDIS seamlessly binds with HP UNDI to provide NDIS2 interface to the upper- layer protocol stack. This allows computers to co nnect to network resources in a DOS environment. ⢠3Com® Boot Server. Client setup Setting up PXE client for NC32xx, NC37xx, NC 380x, NC1020, NC6770, and NC77xx Gig abit Server Adapters involves enabling/disablin g the PXE feature and setting up the BIOS for the boot order. Boot method (NC32xx, NC37xx , NC380x, NC1020, NC6770 , NC7770, and NC7771 only) By default, PXE detects if the BIOS supports BBS (BIO S Boot Specification). If the BIOS supports BBS, then PXE uses BBS as the boot method. If the BIOS does not support BBS, then PXE uses Int18h instea d. However, for some machines equipped with an old BI OS, this process may not work. In case of an old BIOS, users must force PXE to use Int18h, Int19h, or BBS as the boot method. To specify the boot method, press the following keys wh en you see the startup banner. Broadcom Gigabit Ethernet Boot Agent 2.2.6 Copyright (c) 2000 Broadcom Corporation All rights reserved ⢠For Int18h: Press the Ctrl F8 keys ⢠For Int19h: Press the Ctrl F9 keys ⢠For BBS: Press the Ctrl F10 keys Setup BIOS To boot from the network with PXE, make PXE the first bootable device under BIOS. This procedure depends on server BIOS implementation. Refer to the user manual for the serv er for additional information. Server setup Refer to the following instr uctions for your operating system. Windows® 2000
Adapter configurations 69 The current version of Windows® 2000 does not include a network driver for th e NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters. To perform remote installation with PXE, include a network driver for these adapters as part of the client's installation image on the server. Refer to Microsoft® Article ID Q246184, "How to Add Third-Party OEM Network Adap ters to RIS Installations." DOS UNDI/APITEST To boot in DOS and connect to a network for the DO S environment, download the Intel® PXE PDK from the Intel® web site. This PXE PDK comes with a TFTP/ProxyDHCP/Boot server. The PXE PDK can be downloaded from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). Red Hat Linux Red Hat Linux 7.2 (or higher) distribution has PXE serv er support. It allows users to perform a complete Linux installation over the network. Distribution al so comes with boot images âboot kernel (vmlin ux) and initial ram disk (initrd). These two images can be found in the HP SoftPaq file: ⢠/misc/src/trees/boot/vmlinux ⢠/misc/src/trees/initrd-netw ork.img Refer to Red Hat documentation for in structio ns on installing PXE ser ver on Linux. Initrd.img, distributed with Red Ha t 7.2, does not have a Linux netw ork driver for the NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC32 6m, NC10 20, NC67xx, and NC77xx Gigabit Server Adapters. ⢠/images/pxeboot/vmlinux ⢠/images/pxeboot/initrd.img Intel® has also created patches to the Linux PXE serv er code to allow clients to perform a remote boot. You can download these patches from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). A remote boot does not require a standard Linux network driver for t he HP NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC326m, NC102 0, NC67 xx, and NC77xx Gigabit Server Adapters. After the PXE client downloads the Linux kernel and in itial ram disk, the Linux universal driv er, that came with the Linux distribution, will bind with the UNDI code of the PXE to form a Linux network driver. HP NC-Series Intel adapters Boot Agent The Intel® Boot Agent is a software product that allo ws your networked client computer to boot using a program code image supplie d by a remote server. Implementation support for the Boot Agen t includes: ⢠Multi-protocol boot ROM code designed for re mote booting from Gigabit Ethernet Adapters ⢠Compliance with the Pre-boot eXec ution Env ironment (PXE) Version 2.1 Specification ⢠Remote Program Load (RPL) runtim e and loader software included wi th some versions of the Boot Agent Intel® Boot Agent offers:
Adapter configurations 70 ⢠Compliance with the Wired for Manage ment Baseline (WfM) 2.0 specification ⢠Compatibility with legacy boot agent environments that use BOOTP protocol ⢠Customization in pre-boot, Wi ndows®, and DOS environments Operating environment The Boot Agent operates in a client/server environment. Often, in this envir onment, one or more servers provide remote boot services to a large number of client computers thro ugh a common network. The computer system where the Intel® B oot Agent is loaded is considered to be a client with respect to the remote boot capability even if that system acts as a server after the syst em has finished booting. Configuration options The Intel® Boot Agent software provides configuration options that allow you to custom ize the behavior of the Boot Agent software. You can configure the Boot Agent in a pre-boot environment (before the operating system is loaded). Configuring the Boot Agent in a pre-boot PXE or RPL environment You can customize the behavior of the B oot Agent software through a pre-boot (operating system independent) configuration setup program contained within the adapter's flash ROM. A single user interface allows you to configure PXE and RPL protocols on HP Gigabit adapters. You can access this pre- boot configuration setup program each time the client computer cycles throug h the boot process. The boot process is triggered whenever any of the following boot events occur: ⢠Power on ⢠Hard reset ( Reset bu tton on system, if available) ⢠Soft reset ( Ctrl Alt Del ) ⢠Operating system or application-initiated system restart When the boot process begins, the screen clears and the computer begins it s Power On Self-Test (POST) sequence. Shortly after completion of the POST, the Boot Agent soft ware stored in flash ROM execut es. The Boot Agent then displays an initialization message, similar to the one below , indicating that it is active: Initializing Intel(R) Boot Agent Version X.X.XX PXE 2.0 Build 083 (WfM 2.0) Press Ctrl S to enter the Setup Menu. This display may be hidden by the manufacturer's splash screen. Refer to your manufactur er's documentation for details. To customize the behavior of the Boot A gent software in a pre-boot PXE or RPL environment complete the following steps: 1. Press the Ctrl S keys immediately after the initializati on message appears. A configuration setup menu appears allowing you to set configuration values for the Boot Agent. The configuration set up menu is the same for bo th kinds of adapters. If you do not press the Ctrl S keys, the Boot Agent software wil l proceed with the boot process eventually bringing up the operating syst em. If you miss your opportunity to p ress the Ctrl S keys within the allowed number of seconds, reboot the co mputer to try again. If you select a Setup Menu Wait Time setting of zero or a Show Setup Prompt setting of Disa bled, you will not be prompted to press the Ctrl S keys even though yo u can still enter the configuration setup menu using this key combination.
Adapter configurations 71 The configuration setup menu sho ws a list of configuration setting s on the left and their corresponding values on the right. Key descriptio ns near the bottom of the menu indicate how to change values for the configuration setti ngs. For each selected setting, a brief "mini-Help" description of its function appears just above the key descriptions. 2. Select the setting you need to change by using the arrow keys. 3. After you have accessed the settin g yo u want to change, press the Spacebar until the desired value appears. 4. If you want to change additional settings, repeat steps 2 and 3. 5. After you have completed your changes, press the F4 key to update the adapter with the new values. Any changed configuration values are appl ied as the boot process resumes. Boot Agent configuration settings Configuration settings Possible values Description Network Boot Protocol ⢠PXE (Preboot eXecution Environment) ⢠RPL (Remote Program Load) Controls whether the RPL or PXE boot protocol will be used. Select PXE for use with WfM-compatible network management programs, such as Intel® LANDesk® Manage ment Suite, Windows 2000 RIS, and Linux®. Select RPL for legacy-style remote booting, as well as for Novell® Netware® remote boot solutions. Depending on the configuration of the Boot Agent, this parameter may not be changeable. Boot Order ⢠Use BIOS Setup Boot Order ⢠Try network first, then local drives ⢠Try local drives first, then network ⢠Try network only ⢠Try local drives only Sets the boot order in which devic es are selected during boot up if the computer does not have its own control method. If your client computer's BIOS supports the BIOS Boot Specification (BBS), or allows PnP-compliant selection of the boot order in the BIOS setup program, then this setting will always be Us e BIOS Set up Boot Order and cannot be changed. In this ca se, refer to the BIOS setup manual specific to your client computer to set up boot options. If your client computer does not have a BBS- or PnP-compliant BIOS, you can select any one of the othe r possible values listed for this setting except for Use BIOS Setup Boot Order. Show Setup Prompt ⢠Enabled ⢠Disabled Controls whether or not the Boot Agent displays the Ctrl S prompt after POST. If you select Enabled, the Ctrl S prompt appears after POST so that you can press Ctrl S within the number of seconds allowed by the Setup Menu Wait Time setting to display the configuration setup menu. If you select Disabled and/or the Setup Menu Wait Time setting is zero, the Ctrl S setup prompt will not appear during the boot process. However, you can display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedly pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again.
Adapter configurations 72 Configuration settings Possible values Description Setup Menu Wait Time ⢠0 seconds ⢠2 seconds ⢠3 seconds ⢠5 seconds Controls the number of seconds the Boot Agent waits for you to press the Ctrl S keys, so as to suspend the boot process and then configure the way the Boot Agent operates. If you select a Setup Menu Wait Time setting of zero and/or a Show Setup Prompt setting of Disabled, the Ctrl S setup prompt will not appear during the boot process. Ho wever, you can still display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedl y pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again. If during PXE or RPL boot more than one adapter is in stalled in a computer and you want to boot from the boot ROM located on a specific adapter, you can do so by removing the adapter from the BIOS Boot Order or disabling the flash by running IBAUTIL -FlashDisable on the desired adapter. To display diagnostics information: Anytime the configuration setup men u is displayed, you may press the D key to display diagnostics information in the lower half of the screen . This in formation can be helpful during interaction with HP Customer Support personn el or your IT team members. After you press the D key, the informat ion displayed remains until you leav e the configuration setup screen (either by pressing the F4 key, pressing the Esc key, or by rebooting the computer). For more information about how to interpret the in formation displayed, refer to the "Diagnostics infor mation for pre-boot PXE or RPL environments" section. Auxiliary DOS utilities The IBAUtil allows you to install and configure the Bo ot Agent using the DOS environment. IBAU til is a utility program that provides an alternative means for modifying the adapter co nfiguration settings. Use IBAUtil to: ⢠Change the default settings of your HP NC 310F, NC3123, NC617 0, or NC7170 adapter. ⢠Enable/disable the Wake-on-LAN (WOL) an d Intel® Boot Agent capabilitie s. ⢠Allow in-the-field upgrades to the image stor ed in the flash component of the adapter. Wake-On-LAN and Intel® Boot Agent in a DOS environment Use IBAUtil to enable or disable WOL or Boot Agent fe atures. To obtain this utility: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. The IBAUtil file is located in the APPS\BOOTAGNT\N 100X directory. ⢠Wake-On-LAN. When enabled, the adapter can reac t to special "wake up" packets and power up the computer without end user intervention. However, there i s a higher power draw when the system
Adapter configurations 73 is in the suspended state when this i s enabled. It is strongly recommended that no more than two adapters should have this feature enabled per computer. ⢠Intel Boot Agent. When enabled, the computer can initiate PXE/RPL boot if a valid flash image is present on the adapter. Running IBAUtil in DOS IBAUtil (for DOS) must be run with the computer boot ed to DOS only. This utility cannot be run in a DOS window in any Microsoft® Windows® product or in DOS compatibility mode in IBM OS/2. The syntax for issuing IBAUtil command line parameters in DOS is: IBAUTIL [-option]... DOS command line parameters IBAUtil accepts up to 16 simultaneous parameters and does not perfor m both operations if orthogonal parameters (like the commands -FLASHENABLE -FLASHDI SABLE used together) are supplied. If opposing parameters are specified, IBAU til exits with an error. The parameters are grouped with their orthogonal pairs, where applicable. -H ELP, -EXITCODES and - IMAGEVERSION have special overridi ng characteristics. Unless noted, all other parameters may be used in conjunction with all other parameters. If you enter IBAUTIL without any command line options, the utility displays a listing of all of the Intel® network adapters found in your system. Valid parameters are listed below. All other parameters generate an error. Parameter Description -HELP or -? Displays command line help and exits. When -HELP is set, all other parameters are ignored. -EXITCODES Displays exit code help. When -EXITCODES are set, all other parameters except -HELP are ignored. -IMAGEVERSION or -IV Displays the PXE vers ions stored within the IBAUtil. When - IMAGEVERSION is set, all other parameters except -EXITCODES and -HELP are ignored. Adapter Selection Options -ALL Selects all adapters found in system, works for both EEPROM and Flash activity. -NIC=XX Selects a specific adapter (1â16). -BLINK Blinks the LED on the selected adapter for 10 sec onds to provide a method for identifying an adapter. Flash Programming Options -AUTO or -QUIET Runs IBAUtil without asking for user intervention. It does NOT create a backup image. This paramete r must be used with -NIC or - ALL if there is more than one adapter in the system. It runs without - ALL or -NIC=xx if there is only one adapter in the system. -UPGRADE or -UP Downloads the Boot Agent stored within IBAUtil to the adapters in the system. -FLASHENABLE or -FE Enables Boot ROM. Saves -FLASHDISABLE or -FD Disables Boot ROM.
Adapter configurations 74 Parameter Description -SAVE Saves existing Intel Boot Agen t and EEPROM settings to a file. The file is named based on the PCI Vendor and Device ID of the adapter. -RESTORE Restores previously saved Inte l Boot Agent image. IBAUtil looks for a file name based on the PCI Vendor and Device ID of the adapter. If -RESTORE is used with -UPGRADE, an error is generated. Power Management Options -WOLDISABLE or -WOLD Disables WOL bit on Fast Ethernet and disables APMPME and APME EEPROM bits on gigabit. -WOLENABLE or -WOLE Enables WOL bit on Fast Ethernet and enables APMPME and APME EEPROM bits on gigabit. -LWSENABLE OR -LWSE Enables changing the legacy OS Wakeup Support Option on 10/100 adapters. -LWSDISABLE OR -LWSD Disables changing the legacy OS Wakeup Support Option on 10/100 adapters. PXE/RPL Configuration Options -SETUPENABLE or -STE Enables Setup Menu. -SETUPDISABLE OR -STD Disables Setup Menu. -DEFAULTCONFIG or -DEFCFG Resets configurat ion and control words in the EEPROM back to default settings. -TITLEENABLE OR -TLE Enables initial title message. -TITLEDISABLE OR -TLD Disables initial title message. -PROTOCOLENABLE OR -PRO E Enables changing the boot protocol option. -PROTOCOLDISABLE OR -PROD Disables changing the boot protocol option. -ORDERENABLE OR -ORDE Enables changing the boot order option. -ORDERDISABLE OR -ORDD Disables ch anging the boot order option. -SETWAITTIME=X or -SWT=X Sets the setup delay time for the ada pter to X seconds. Valid delay times are 0, 2, 3, and 5. Only the first instance of this parameter is valid. All other instances of the parameter generate an error. -RPL BOOT Sets the network boot protocol to RPL if an RPL image is present in the adapter's flash. -PXE BOOT Sets the network boot protocol to PXE if a PXE image is present in the adapter's flash. DOS error codes IBAUtil returns codes to the DOS co mmand line when an error occurs. ⢠0âSuccess ⢠1âOperator termination ⢠2âInvalid adapter ⢠3âBad command line parameter ⢠4âEEPROM checksum failure
Adapter configurations 75 ⢠5âEEPROM read failure ⢠6âEEPROM write fail ure ⢠7âEEPROM dump failure ⢠8â(Not used) ⢠9âNo memory ⢠10âNo adapters found Examples of IBAUtil command lines The following are examples of some typical IBAUtil command lines. ⢠Disables WOL on all HP network adapters: IBAUTIL -ALL -WOLDISABLE ⢠Enables WOL on the second HP netw ork adapter found in your system: IBAUTIL -NIC=2 -WOLENABLE ⢠Updates the Boot Agent image on all HP network adapters: IBAUTIL -ALL -UPGRADE ⢠Enables the Boot ROM on the fir st HP Network Adapter found in your system: IBAUTIL -NIC=1 -FE Boot Agent messages The following error and information messages may be di splayed during initializati on. Possible causes and how to avoid an error condition are provided where applicable. ⢠Flash device wrong size This error message can occur when using IBAUtil. The full version of Boot Agent so ftware cannot be loaded into a flash ROM device smaller than 64 KB. If a 32 KB (or smaller) flash ROM device is detected in the adapter, the flash update software returns this message. Assuming the device is socketed, remove the "too small" device and install a blank 64 KB (or larger) flash ROM device. ⢠Invalid PMM function number PMM is not installed or is not work ing correctly. Try updating the BIOS. ⢠PMM allocation error PMM could not or did not allocate the reques ted amount of memory for driver usage. ⢠Press Ctrl S to enter the Setup Menu. Instructs you how to enter the configuration setup menu. ⢠PXE-E00: This system does not have enough free conventional memory. The Boot Agent cannot continue. System does not have enough free memory to run PXE image. The Boot Ag ent was unable to find enough free base memory (below 6 40K) to install the PXE client software. The system cannot boot via PXE in its current configuration. The error retu rns control to the BIOS and the system does not attempt to remote boot. If this error persists, try updating your system's BIOS to the most-recent version. Contact your system adm inistrator or HP customer support to resolve the problem. ⢠PXE-E01: PCI Vendor and Device IDs do not match! Image vendor and device ID do not match those loca ted on the card. Be sure the correct flash image is installed on the adapter. ⢠PXE-E04: Error reading PCI configuration space. The Boot Agent cannot continue.
Adapter configurations 76 PCI configuration space could not be read. Machin e is probably not PCI compliant. The Boot Agent was unable to read one or more of the adapter's PCI configuration registers. The adapter may be misconfigured, or the wrong Boot A gent image may be installed on the adapter. The Boot Agent returns control to the BIOS and not attempt to remote boot. Try to update the flash image. If this does not solve the problem, contact your system administrat or or HP Customer Support. ⢠PXE-E05: The LAN adapter's configuration is corrupted or has not been initialized. The Boot Agent cannot continue. The adapter's EEPROM is corrupted. The Boot Agent determ ined that the adapter EEPROM checksum is incorrect. The agent will return control to the BIOS and not attempt to remote boot. Try to update the flash image. If this does not solve the problem, contact your system administrat or or HP Customer Support. ⢠PXE-E06: Option ROM requires DDIM support. The system BIOS does not support DDIM. The BIOS does not support the mapping of the PCI expansion ROMs into upper memory as required by the PCI specification. The Boot Ag ent cannot function in this system. The Boot Agent returns cont rol to the BIOS and does not attempt to remote boot. You may be able to resolve the problem by updating the BIOS on your system. If updating your system's BIOS does not solve the pro blem, contact your system administrator o r HP customer support to resolve the problem. ⢠PXE-E07: PCI BIOS calls not supported. BIOS-level PCI services not available. Machine is prob ably not PCI compliant. ⢠PXE-E09: Unexpected UNDI loader error. Status == xx The UNDI loader returned an unknown e rror status. xx is the status returned. ⢠PXE-E20: BIOS extended memory copy error. BIOS could not move the image into extended memory. ⢠PXE-E20: BIOS extended memo ry copy error. AH == xx Error occurred while trying to copy the image in to extended memory. xx is the BIOS failur e code. ⢠PXE-E51: No DHCP or BOOTP offers received. The Boot Agent did not receive any DHCP or BOOTP re sponses to it s initial request. Be sure that your DHCP server (and/or proxyDHCP server, if one is in use) i s properly configured and has sufficient IP addresses available for lease. If you are using BOOTP, be sure that the BOOTP service is running and is properly configured. ⢠PXE-E53: No boot filename received. The Boot Agent received a DHCP or BOOTP offer, but has not received a valid filename to download. If you are using PXE, please check your PXE and BINL configu rat ion. If you are using BOOTP, be sure that the TFTP service is running an d that the specific path and filename are correct. ⢠PXE-E61: Media test failure. The adapter does not detect link. Be sure that the cable is good and is attached to a working hub or switch. The link light visible from th e back of the adapter should be lit. ⢠PXE-EC1: Base-code ROM ID structure was not found. No base code can be located. An incorrect fl ash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC3: BC ROM ID structure is invalid.
Adapter configurations 77 Base code cannot be installed. An incorrect fl ash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC4: UNDI ROM ID structure was not found. UNDI ROM ID structure signature is incorrect. An incorrect flash ima ge is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC5: UNDI ROM ID structure is invalid. The structure length is incorrect. An incorrect flash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC6: UNDI driver image is invalid. The UNDI driver image signature is invalid. An incorrect flash image is install ed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC8: !PXE structure was not found in UNDI driver code segment. The Boot Agent cannot locate the needed !PXE stru cture reso urce. An incorrect flash image is installed or the image has become corr upted. Try to update the flash image. ⢠PXE-EC9: PXENV structure was not found in UNDI driver code segment. The Boot Agent cannot locate the needed PXENV structure. An inco rrect flash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-M0F: Exiting Intel Boot Agent. Ending execution of the RO M image. ⢠This option has been locked and cannot be changed. You have attempted to change a configuration setting that has been locked by your system administrator with IBAUtil. This message can ap pear from the configuration setup menu when operating in a stand-alone environment. If you thin k you should be able to change the configuration setting, consult your system administrator. ⢠PXE-M0E: Retrying network boot; press ESC to cancel. The Boot Agent did not successfully complete a netw ork boot due to a network error (such as not receiving a DHCP offer). The Boot Agent continues to a ttempt to boot from the network until successful or until canceled by the user. This feat ure is disabled by default. For information on how to enable this feature, contact HP Customer Support. Troubleshooting procedures The following list of problems and associated sol ution s covers a representative set of problems that you might encounter while using the Boot Ag ent. If you are experiencing a problem that is not listed in this section, contact HP Customer Support. After booting, my computer experiences problems After the Boot Agent product has finished its sole task (remote booti ng), it no longer has any effect on the client computer operation. Thus, any issues that arise after the boot process is complete are most likely not related to the Boot Agent product. If you are having problems with the local (client) or network operating syst em, contact the operating system manufacturer for as sistance. If you are having problems with some application program, contact the application manufacturer for assistance. If you are having problems with any of your computer's hardware or with the BIOS, contact HP Customer Support for assist ance. Cannot change boot order
Adapter configurations 78 If you are accustomed to redefini ng your computer's boot order using the motherboard BIOS setup program, the default settings of the Boot Agent set u p program can override that setup. To change the boot sequence, you must first override the Boot Ag ent setup program defaults. To start the Boot Agent configuration setup progra m, press the Ctrl S keys during the time the initialization message appears. A configuration setup menu appears allow ing you to set configuration values for the Boot Agent. If you do not press the Ctrl S keys, the Boot Agent softwar e proceeds with th e boot process, eventually bringing up the operating system. To change your computer's boot order sett ing, refer to the "Configuring t he Boot Agent in a pre-boot PXE or RPL environm ent" section. My computer does not complete POST If your computer fails to boot with an adapter instal led, but does boot when you remove the adapter, try moving the adapter to another computer an d using IBAUtil to disable the flash chip. If this does not work, the proble m may be occurring before the Boot Agent softwa re even begins operating. In this case, there may be a BIOS proble m with your computer. Contact HP Customer Support for help in correcting your problem. Not getting a prompt to go to Boot Agent setup program This is actually a feature of the Boot Agent product. As part of the Boot Agent's Configurati on Setup Program, two of the options, Show Setup Prompt and Setup Wait Time, allow the boot process to proceed after POST without interruption by the Ctrl S prompt (Press the Ctrl S keys to enter the Setup Menu) If the Show Setup Prompt has previously been set to Disabled and/or if the Setup Menu Wait Time option is set to zero seconds, the prompt does not disp lay after POST. However, even though you are not prompted to press the Ctrl S keys, you can still enter the config uration setup menu to customiz e configuration settings as f ollows: Repeatedly press the Ctrl S keys immediately after POS T until th e Configuration Setup Menu appears. If the Configuration Setup Menu does not appear after repeatedly pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again. While in the Configuration Setup Men u, you can restore the prompt message as follows: ⢠If Show Setup Prompt has been set to Disabled, change it to Enabled. ⢠If Setup Menu Wait Time is set to zero seconds, change it to an y of the wait time values that exceed zero seconds. Refer to the list of possibl e Set Menu Wait Time values shown in the "Boot Agent configuration settings" section. There are configuration/operation problems with the boot process If your PXE client receives a DHCP address, but then fails to boot, you know the PXE client is working correctly. Check your network or PXE server config uration to trou bleshoot the problem. Contact HP Customer Support if you need further assistance. Diagnostics information for pre-boot PXE or RPL environments Anytime the Configuration Setup Men u is displayed (refer to the "Config uring the Boot Agent in a pre- boot PXE or RPL environment" section), you may press the D key to display diagnostics information on the screen. This information can be helpful during interact ion with HP Customer Support personnel or your IT team members. After you press the D key , the information remains displayed until you reboot your computer. Actual diagnostics information may vary, depending upon the adapter(s) installed in your computer. Diagnostics information may include the f ollowing items: ⢠PWA NumberâThe Printed Wire Assembly number of the device as stored in the EEPROM
Adapter configurations 79 ⢠MAC AddressâThe individual address of the device as stored in the EEPROM ⢠I/OâThe I/O address for PCI access determined by the software. On cards without an I/O address, all zeros are displayed ⢠MemoryâThe memory map PCI access address det ermined by the software ⢠SlotâThe slot number reported by the BIOS. The number displayed is the BIOS v ersion of the PCI slot number. Therefore, actual positions of adapte rs within slot s may not be displayed as expected. Slots are not always enumerated in an obvious manne r, and the diagnostic s will only report what is indicated by the BIOS. Boot agent glossary This glossary defines terms, abbrevia tions, and acronyms that apply directly to the Intel® Boot Agent. Term / Abbreviation / Acronym Definition ACPI Advanced Configuration and Power Interface AOL Alert on LAN API Application Program Interface. The spec ific method presc ribed by an OS or an application program by which a programmer writing an application program can make requests of the operating system or the application. BAID BIOS Aware IPL Device. The BIOS contains all code required to IPL from the device. BAR Base Address Register BBS BIOS Boot Specification BC BaseCode. The PXE BaseCode, included as a core prot ocol in EFI, is comprised of a simple network stack (UDP/IP) and a few common network protocols (DHCP, Bootserver Discovery, TFTP) that are useful for remote bootin g machines. BCV Boot Connectio n Vector. A field in the Plug and Play (PnP) header for a device with an associated option ROM. BEV Boot Entry Vector. A field in the Plug and Play (PnP) Header of a device with an associated option ROM. PXE is implemented as a BEV option ROM. BINL Binary Image Negotiation Layer BIOS Basic Input/Output System. The program a personal computer's microprocessor uses to get the computer system started after yo u turn it on. It also manages data flow between the computer's OS and attached devices. BIS Boot Integrity S ervices BOOTP Bootstrap Protocol. A legacy remote b ooting protocol developed originally for use with UNIX. Used as Linux's server side PXE host software. Runs a Daemon once installed. DDIM Device Driver Initialization Model DHCP Dynamic Host Configuration Protocol. An industry standard internet protocol defined by the IETF. DHCP was defined to dyna mically provide communications-related configuration values such as network address es to network client computers at boot time. DHCP is specified by IETF RFC s 1534, 2131, and 21 32. EEPROM Electrically Erasable Pr ogrammable Read-Only Memory ESD Electro-Static Discharge Flash A high-density, truly non-volatile, hig h-performance, read-write memory solution, also characterized by low power consumption, extreme ruggedness, and high reliability.
Adapter configurations 80 Term / Abbreviation / Acronym Definition Flb Flash Library IA32 Shorthand for computer platforms ba sed on Intel 32-bit architecture CPUs. IA64 Shorthand for computer platforms ba sed on Intel 64-bit architecture CPUs. IAL Intel Architecture Lab IBA Intel Boot Agent IETF Internet Engineering Task Force. The IETF is a large open int ernational community of network designers, operators, vendors, and researchers concerned with the evolution of the Internet architecture and the smooth operation of t he internet. It is open to any interested individual. For more details, refe r to the Internet Engineering Task Force website. IPL Initial Program Load. Also known as the bootstrap or boot process. LCM LANDesk Configuration Manager LOM LAN On Motherboard. This is a network de vice that is built onto the motherboard (or baseboard) of the machine. MTFTP Adds multicast support to a TFTP server. NBP Network Bootst rap Program. The first pr ogram that is downloaded into a machine that has selected a PXE capable devi ce fo r remote boot services. A typical NBP examines the machine it is running on to try to determine if t he machine is capable of running the next layer (OS or application). If the machine is not capable of running the next layer, control is returned to th e EFI boot manager and the next boot device is selected. If the machine is capable, t he next layer is downloaded and control can then be passed to the downlo aded program. Though most NBP s are OS loaders, NBPs can be written to be stan d-alone applications such as diagnostics, backup/restore, remote management agents, browsers, and so on. NIC Network Interface Controller. Also referred to as adapter. Technically, a network device that is inserted into a bus on the motherboard or into an expansion board. For the purposes of this document, the term adapter is used in a generic sense, meaning any device that enables a netwo rk connection (including LOMs and network devices on external buses, such as USB 1394, and so on). OS Operating System PMM POST Memory Mana ger. A mechanism used by option ROMs to allocate RAM memory for use during system startup. PnP Plug and Play POST Power On Self-Test proxyDHCP proxyDHCP is used to ease the transition of PXE clients and servers into an existing network infrastructure. proxyDHCP provid es additional DHCP information that is needed by PXE clients and boot servers wi thout making changes to existing DHCP servers. PXE Preboot Executi on Environment. PXE pr ovides a way for a system to initiate a network connection to various servers prior to loading an OS. This netw ork connection supports a number of standard IP protocols such as DHCP and TFTP, and can be used for purposes such as softwa re installation and system inventory maintenance. PXE 16 PXE Specification v2.1 or ea rlier. Also known as 16-bit PXE. RFC Request For Comments. Used to denote that a new standard or modification to standard is be introduced by IE TF or other standards bodies.
Adapter configurations 81 Term / Abbreviation / Acronym Definition RIS Remote Installation Services. A Micr osoft® service that uses PXE to deliver the Network Bootstrap Program (NBP) to start the boot process. Sh ips with Windows® 2000 Advanced Server. ROM Read-Only Memory. When used in this guide, ROM refers to a non-volatile memory storage device on an adapter RPL Remote Program Load. A bootable network device protocol, developed and used mostly by IBM. TCO Total Cost of Ownership. Business models often calculate the approximate total cost of ownership for technology initiative \s through formulas that incorporate both expenditure on hardware and software an d maintenance, replacement, training, and all aspects of business that are impacted by the core system. Remote management tools like PXE help lower the overall cost of t he ownership. TFTP Trivial File Tran sfer Protocol. An indust ry standard internet protocol defined by the IETF to enable the transmission of files across the internet. Trivial File Transfer Protocol (TFTP, Revision 2) t o support NBP downloa d is specified by IETF RFC 1350 . UNDI Universal Network Driver Interface. UNDI is an archit ectural interface to adapters. Traditionally adapters have had custom inte rfaces and custom drivers (each ada pter had a driver for each OS on each pl atform architecture). UNDI provides a hardware-dependent me chanism for the PXE base code to use a adapter for network access without controlling the adapter hardware directly. H/W UNDI is an architectural hardware interface to an adapter; S/W UNDI is a software implementation of the H/W UNDI. URL Uniform Resource Locator. The address of a file (resource) accessible on the internet. USB Universal Serial Bus. A P lug and Play (PnP) interface between a c omputer and add- on devices. WfM Wired for Management WOL Wake on LAN Keywords for N100 NDIS2 driver The following how to configure the parameter s for the NDIS 2 driver version 4.56 (N100 DOS) . General configuration parameters The following configuration parame ters are supported through the PR OTOCOL.INI file. The "Possible Errors" section describes the driver's behavior when erro r conditions are detected. If an error is fatal, the driver will not load. In all other cases the driver will load and use the default value. ⢠DRIVERNAME o Syntax: DRIVERNAME=[N100$ | N1002$ | etc] o Example: DRIVERNAME=N100$ o Default: None. This is a required parameter. o Normal Behavior: The driver finds its section in PR OTOCOL.INI by matching its instance ID to the value for this parameter. o Possible Errors: The device driv er uses a DOS and OS/2 function to display the name of the driver it is expecting. This function cannot disp lay a "$" character. For this reason, the user may
Adapter configurations 82 see a message referring to this value with the "$." The user must rem ember to enter the "$" character as part of the parameter's value. ⢠SPEED o Syntax: SPEED=[Auto | 10 | 100] o Example: SPEED=10 o Default: Auto-Speed-Detect as a function of PHY detection o Normal Behavior: If either 10 or 100 is specified, a message indicates that Auto-Speed-Detect has been disabled and the driver sets the speed us ing the value specified. If the parameter i s not provided at all, or if Auto is prov ided, Auto-Speed-Sensing is enabled. o Possibl e Errors: If the SPEED parameter is set to an invalid valu e, the parameter is ignored and the default value is used, and a message indicates a "Par ameter value out of range" error. If the FORCEDUPLEX parameter is assigned a valid non-zero value, this parameter is required. If the duplex is forced, Auto-Speed-Sensing is disabled and the speed and full-duplex modes are set as specified by the user. An N-Way PHY will not be able to detect speed changes in this configuration until the driver is reloaded. ⢠FORCEDUPLEX o Syntax: FORCEDUPLEX=[Auto | 1 | 2] o Example: FORCEDUPLEX=1 o Default: Duplex mode configuration detected during PHY initialization. o Normal Behavior: The parameter value of 1 wi ll force the duplex mode to half-dup lex. A parameter value of 2 will force th e duplex mode to full-duplex. The SPEED parameter must be specified and must be valid if either half or full-duplex is forced. Auto-detection by the PHY is disabled and the sp eed and full-duplex modes are set as specified. The N-Way PHY will not be able to auto-detec t the mode until the driver is reloaded. o Possibl e Errors: If the FORCEDUPLEX parameter is set to an invalid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. If the FORCEDUPLEX mode does not match the actu al mode of the port in use, the driver will function for certain levels of traffic and then begin to lose frames. If the port in use is FULL duplex but it is not capable of performing an N-way auto-negotiation, the user MUST use FOR CEDUPLEX set to FULL. ⢠PHYADDRESS o Syntax: PHYADDRESS=[0..32] o Example: PHYADDRESS=1 o Default: 1, implying logical PHY 1 at address 1. o An address of 0 implies logical PHY 0. PHY 0 can only be at address 0. PHY 1 can have addresses in the range 1 - 31. PHY address 32 can be used to indicate the presence of an 82503 interface.
Adapter configurations 83 o Normal Behavior: The value will be used by the PH Y detection rout ine to locate and initialize the PHY. If a PHY is not found at the address specifie d, the driver will load but not function properly, because PHY detection occurs afte r the initial load has occurred. o Possible Errors: If the PHYADDRESS parameter is set to an invalid value, the parameter is ignored and the default value is used, and a message indi cates a "Parameter value out of range" error. ⢠SLOT o Syntax: SLOT=[0x0..0x1FFF] Optional unless you wa nt to specify a particular a d apter or module that is not loaded by default. (By default, the driver loads on the first adapter it finds.) o Example: SLOT=0x1C o Default: The driver will Auto-Configur e o Normal Behavior: The driver uses the value of th e parameter to decide which adapter to control. o Possibl e Errors: If only one adapter is installed and the value does not correctly indicate the adapter slot, a message indicates that the value does not match th e actual configuration, and the driver find s the adapter and uses it. If more than one adapter is installed and the va lue does not indicate the adapter slot, a message displays indicating possible slots to use and the driver loads on the next available slot. ⢠NODE o Syntax: NODE="12 hexadecimal digits" The value must be exactly 12 hexadecima l digits, enclosed in double quot es The value cannot be all zeros. The value cannot have the Multicast bit set (LSB of 2nd digit = 1). o Example: NODE="00AA00123456" o Default: Value from EEPROM installed on adapter o Normal Behavior: The Current Station Address in th e NDIS MAC Service-Specific Characterist ics (MSSC) table is assigned the value of this parameter. The adapter hardware is programmed to receive frames with the destinat ion address equal to the Current Station Address in the MSSC table. The Permanent Station Address in the MSSC table will be set to reflect the node address read from the adapter's EEPROM. o Possible Errors: If any of the rules described above is violated , an error message indicates the rules for forming a proper address, and the driver treats this as a fatal error. ⢠CONGENB o Syntax: CONGENB=[0 | 1] o Example: CONGENB=1 o Default: Congestion Control is disabled. o Normal Behavior: The absence of th is parameter or the value 0 resu lts in no Congestion Control. The value of 1 enables the Congestion Contro l feature. o Possible Errors: If the CONGENB parameter is set to an invalid value, the parameter is ignor ed and the default value is used, and a message indi cates a "Parameter value out of range" error. ⢠RXFIFO o Syntax: RXFIFO=[0..15 | 0x0..0xF] o Example: RXFIFO=8
Adapter configurations 84 o Default: 8, corresponding to 32 byte s or 8 dwords in the receive FIFO. o Normal Behavior: No message wi ll be displayed. The driver will use the specified value to configure. o Possible Errors: If the RXFIFO pa rameter is set to an invalid valu e, the parameter is ignored and the default value is used, and a message indica tes a "Parameter value out of range" error. ⢠TXFIFO o Syntax: TXFIFO=[0..15 | 0x0..0xF] o Example: TXFIFO=8 o Default: 8, corresponding to 32 bytes or 8 dwords in the transmit FIFO. o Normal Behavior: No message wi ll be displayed. The driver will use the specified value to configure. o Possible Errors: If the TXFIFO pa rameter is set to an invalid valu e, the parameter is ignored and the default value is used, and a message indica tes a "Parameter value out of range" error. ⢠TXTHRESHOLD o Syntax: TXTHRESHOLD=[0..254 | 0x0..0xFE] o Example: TXTHRESHOLD=32 o Default: 32, corresponding to 32 * 8=256 bytes o Normal Behavior: The value will be set in every transmit command. The initial value can only be changed at load time, although the driver will increase the value if underruns are d etected. Setting a low value (less than 16) in high bus latency sy stems could cause excessive TX DMA underruns, reducing throughput until th e driv er reacts. Setting a value of more than 200 completely eliminates the possibility o f TX DMA underruns. o Possibl e Errors: If the TXTHRESHOLD parameter is set to an in valid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. ⢠IFS o Syntax: IFS=[0..255 | 0x0..0xFF] o Example: IFS=16 o Default: 0 (minimum spacing between frames) o Normal Behavior: The value will be used to config ure the 82557 at initiali zation time. A value of 0 implies minimum IFS between 2 transmit frames. Increasing the value reduces the probability of collisions under heavy load due to multiple client s, but will have a negative impact on transmit throughput in a point-to-point single-client packet burst. o Possible Errors: If the IFS paramet er is set to an invalid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. ⢠RXDMACOUNT o Syntax: RXDMACOUNT=[0..127 | 0x0..0x7F] o Example: RXDMACOUNT=16 o Default: 0 implying that the Rx DMA will contin ue until it is done. No other DMA request can preempt it.
Adapter configurations 85 o Normal Behavior: The value will be used to config ure the 82557 at initialization ti me. Setting it to a low value will force frequent internal arbitr ation, which will affect overall performance. The value can be tuned to obtain optimum balance between the TX and Rx DMA machines. o Possibl e Errors: If the RXDMACOUNT parameter is set to an in valid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. ⢠TXDMACOUNT o Syntax: TXDMACOUNT=[0..127 | 0x0..0x7F] o Example: TXDMACOUNT=16 o Default: 0 implying that the Tx DMA will contin ue until it is done. No other DMA request can preempt it. o Normal Behavior: The value will be used to config ure the 82557 at initialization ti me. Setting it to a low value will force frequent internal arbitr ation, which will affect overall performance. The value can be tuned to obtain optimum balance between the Tx and Rx DMA machines. o Possibl e Errors: If the TXDMACOUNT parameter is set to an in valid value, the parameter is ignored and the default value is used, and a message is indica tes a "Parameter value out of range" error. ⢠TRANSMITBUFFERS o Syntax: TRANSMITBUFFERS=[4 .. 30] o Example: TRANSMITBUFFERS=14 o Default: 4 for DOS driver, 10 for OS/2 driver o Normal Behavior: The driver allocates the number of buffer s indicated by the parameter value. o Possibl e Errors: If the value provided is out of range, the paramet er is ignored and the default value is used, and a message indicates a "Parameter value out of range" error. CAUTION: If the number of buffers allocated caus es the driverâs size to exceed MAXIMUM, the driver adjusts the number of buffers d own ward, by reducing each of the counts of transmit/receive buffers by one unt il the total allocation will fit in the available memory. A message indicates the actual number of buff ers allocated. Under DOS without Window for Workgroups (WFW), MAXI MUM is define d as 64Kb for all code and data. Under DOS and WFW, the driver must not claim more space than its load size, but since the driver cannot detect that it is running under WFW, a message warns the user when the size allocated exceeds the size of the driver when loaded. Under OS/2, MAXIMUM is defin ed as 64Kb for all data. ⢠RECEIVEBUFFERS o Syntax: RECEIVEBUFFERS=[4 .. 30] o Example: RECEIVEBUFFERS=14 o Default: 8 for DOS driver, 10 for OS/2 driver o Normal Behavior: The driver allocates the number of buffers indicated by the parameter value. o Possibl e Errors: If the value provided is out of range, the paramet er is ignored, the d efault value is used, and a message indicates a "Parameter value out of range" error.
Adapter configurations 86 CAUTION: If the number of buffers allocated caus es the driverâs size to exceed MAXIMUM, the driver adjusts the number of buffers d own ward, by reducing each of the counts of transmit/receive buffers by one unt il the total allocation will fit in the available memory. A message indicates the actual number of buff ers allocated. Under DOS without Window for Workgroups (WFW), MAXI MUM is define d as 64Kb for all code and data. Under DOS and WFW, the driver must not claim more space than its load size, but since the driver cannot detect that it is running under WFW, a message warns the user when the size allocated exceeds the size of the driver when loaded. Under OS/2, MAXIMUM is defined as 64Kb for all data. ⢠MCWA o Syntax: MCWA=[0..255 | 0x0..0xFF] o Example: MCWA=8 o Default: 6 o Normal Behavior: This function is based on the system timer. When the counter reaches 2 seconds (in both OS/2 and DOS), the driver checks to see if the driver has received no frames since the last time the counter ran, and the Re ceive Ha ng workaround will be executed. When the workaround condition exists, the serv ice routin e that runs the workaround will attempt to run the workaround every timer tick (18.3/sec in DOS, ~32/sec in OS/2). The EEPROM contains information that will control the running of the workaround. If in the event that the EEPROM is configured so that the workaround is not need ed, the counter will not be added to the timer chain, saving system resourc es. o Possible Errors: The value is not checked in any wa y, only 8 bits of value are used. The counter's effectiveness could be compromised if anot her execut able disrupts the timer chain. ⢠FRAMEGAP o Syntax: FRAMEGAP=[6..64 | 0x06..0x40] o Example: FRAMEGAP=12 o Default: 6 o Normal Behavior: In normal circumstances, this will not be needed. If packets are being dropped while running a T4 card, the FRAMEGAP must be increased. A value of 12 or 15 will fix the problem in most cases. Do not use this p arame ter unless you are exp eri encing dropped packets. o Possible Errors: If an invalid ra nge or syntax for this parameter occurs, the default value is used. The driver does not give any outward indica tion as to the state of this parameter. ⢠CACHEFLUSH o Syntax: CACHEFLUSH=[0|1] o Example: CACHEFLUSH=1 o Default: 0 o Normal Behavior: Use this parame ter during a remote installation or "push-install" of Windows NT®. o Possible Errors: Any non-zero valu e sets this parameter to 1. The driver does not give any outward indication as to th e state of this parameter.
Adapter configurations 87 Keywords for N1000 NDIS2 driver The following describes how to config ure the parameters for the NDIS 2 driver vers ion 3.62 (N1000.DOS). Configuration The following configuration parame ters are supported through the PR OTOCOL.INI file. The "Possible Errors" section describes the driver's behavior when erro r conditions are detected. If an error is fatal, the driver will not load. In all other cases the driver will load and use the default value. ⢠DRIVERNAME o Syntax: DRIVERNAME=[N1000$ | N10002$ | etc] o Example: DRIVERNAME=N1000$ o Default: None. This is a required parameter. o Normal Behavior: The driver finds its section in PR OTOCOL.INI by matching its instance ID to the value for this parameter. o Possible Errors: The device driver uses a DOS fu nction to display the name of the driver it is expecting. This function cannot display a "$" ch aracter. For this reason, the user may see a message re ferring to this valu e without the "$." The user must remember to enter the "$" character as part of the parameter's value. ⢠SPEEDDUPLEX o Syntax: SPEEDDUPLEX = [0 | 1 | 2 | 3] o Example: SPEEDDUPLEX = 1 o Default: Speed and Duplex mo de configuration detected during PHY initialization. o Normal Behavior: The parameter value of: 0 will force the speed to 10 Mbps and the duplex mode to half duplex. 1 will force the speed to 10 Mbps and the duplex mode to full duplex. 2 will force the speed to 100 Mbps an d the duplex mode to half duplex. 3 will force the speed to 100 Mbps an d the duplex mode to full duplex. o Auto-detection by the PHY is disabled and the sp eed and full duplex modes are set as specified. The PHY will not be able to auto-detect the mode until the driver is reloaded. o Possibl e Errors: If the SPEEDDUPLEX parameter is set to an inva lid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. If the SPEEDDUPLEX mode does not match the actual mode of the port in use, the driver may or may not function for certain levels of traffic and then begin to lose frames. The SPEEDDUPLEX parameter is valid for HP NC71 xx Server Adapters or Modules only and is ignored if used with HP NC61 xx Server Adapters or Modules. The HP NC71xx Server Adapters and Modules cannot be forced to 1000 Mbps. ⢠SLOT o Syntax: SLOT=[0x0..0x1FFF] Optional unless you wa nt to specify a particular a d apter or module that is not loaded by default. o Example: SLOT=0x1C o Default: The driver will Auto-Configur e if possible.
Adapter configurations 88 o Normal Behavior: The driver uses the value of th e parameter to decide which adapter to control. o Possibl e Errors: If only one adapter is installed and the value do es not correctly indicate adapter slot, a message indicates that the value does not match the actual configuration, and the driver finds the adapter and uses it. If more than one adapter is installed and the va lue does not indicate the adapter slot, a message indicates that the specified slot is invalid for the driver instance. The parameter is ignored and the default value is used. ⢠NODE o Syntax: NODE = "12 hexadecimal digits" The value must be exactly 12 hexadecima l digits, enclosed in double quotes. The value cannot be all zeros. The value cannot have the Multicast bit set (LSB of 2nd digit == 1). o Example: NODE = "02AA00123456" o Default: Value from EEPROM installed on adapter o Normal Behavior: This value sets the adapter's Individual Address, overriding the value read from the adapter EEPROM. The address must cons ist of 12 hexadecimal digits, enclosed in double quotes. The value cannot be all zeros and cannot have the Multicast bit set. Use only the hexadecimal digits 2, 6, A, or E as the seco nd digit of the twelve required in the NODE ADDRESS parameter. Use any hexadecimal digits in all other positions. o Possible Errors: If any of the rules described above is violated , an error message indicates the rules for forming a proper address, and the driver treats this as a fatal error. ⢠CACHEFLUSH o Syntax: CACHEFLUSH= [0|1] o Example: CACHEFLUSH=1 o Default: 0 o Normal Behavior: Use this parame ter during a remote installation or "push-install" of Windows NT®. o Possible Errors: Any non-zero valu e sets this parameter to 1. The driver does not give any outwards indication as to th e state of this parameter. Example PROTOCOL.INI file PROTOCOL.I NI sectio n for the HP Gi ga bit Server Adapters and Modules: ⢠DRIVERNAME parameter o Use this parameter to set 1st adapter = N1000$, 2nd adapter = N10002$, and so on o DRIVERNAME = N1000$ ⢠NODE parameter o Use this to change the MAC address us ed by the adapter o NODE="02AA00123456" ⢠SLOT parameter o Use this to configure a particular instance of th e driver designated by DRIVERN AME, to use the adapter in a particular PCI slot. Slots should be specified in hex notation.
Adapter configurations 89 o SLOT=0x7 ⢠CACHEFLUSH parameter o Set this to 1 to hook the system Disk Buffer Flush interrupt. This causes the driver to properly tur n off the adapter when the system is shut down by Windows NT® during remo te installation (push- install). o CACHEFLUSH=1 o Known Limitations: This driver doe s not suppor t any application that uses protected mode, for example, EMM386. In addition, testing has demons trated that the GHOST software application fails when used with this driver under Netx (VLMs are OK). PXE configuration for HP NC-Series Intel adapters The Pre-boot Execution Environment (PXE) of HP adap ters has been upgraded to allow PXE functions to work with the HP ProLiant servers. The Boot Agent software provides the ability to custom ize it s behavior through its configurati on options. Changing the Boot Agent configuration is required only the first time you add a new network adapter to a client computer within your network. You can configure the Boot Agent within any of the followin g environments: ⢠A pre-boot environment (before an operating syst em is loaded) ⢠A Windows® environment ⢠A DOS environment Configuring the Boot Agent in a pre-boot environment You can customize the behavior of the B oot Agent software through a pre-boot (operating system- independent) configuration program contained within the adapter's fl ash ROM. You can access this pre- boot configuration setup program each time the cl ient computer cycles through the boot process. When you power up or perform a hard or soft system reset, the following message is displayed during the initialization process of the Boot Agent firmware . It indicates that the boot agent is active. Initializing Intel(R) Boot Agent Version 3.0.05 PXE 2.0 Build 078 (WFM 2.0), RPL V2.73 The following message is then displayed duri ng the boot process to Boot A gent Services: Intel(R) Boot Agent 3.0.05 Copyright(c) 1997-2000, Intel Corporation Client MAC Address: 00 50 8B xx xx xx GUID:00000000-0000-0000-000000000000 Client IP: 123.xxx.xxx.xxx MASK: 123.xxx.xxx.xxx DHCP IP: 123.xxx.xxx.xxx Gateway IP: 123.xxx.xxx.xxx Changing the Boot Agent configuration settings To access the Boot Agent Setup Menu, immediately after the initialization messages, and befor e any subsequent message is displayed, quickly press Ctrl S . This menu provides the ability to customize the "Boot Agent Configuration Settings." The Boot Agent setup menu displays a list of configur ation setting s on the left, and their corresponding values on the right. Key functions are described near the bottom of th e menu. The menu indicates how to change the values of settings. To access a brief Help descript ion of a setting, select it.
Adapter configurations 90 To customize the Boot Agent configuration settings: 1. Highlight the setting you want to chan ge using the arrow keys. Then press Enter. 2. Press the spacebar until the desired value appears. 3. If you want to change additional settings, repeat steps 1 and 2. 4. After you have completed your changes, press F4 to update the adapter with the new values. Boot Agent configuration settings The following table describes the Boot Agent config uration settings. Configuration settings Possible values Description Network Boot Protocol ⢠PXE (Preboot eXecution Environment) ⢠RPL (Remote Program Load) Controls whether the RPL or PXE boot protocol will be used. Select PXE for use with WfM-compatible network management programs, such as Intel® LANDesk® Manage ment Suite, Windows 2000 RIS, and Linux®. Select RPL for legacy-style remote booting, as well as for Novell® Netware® remote boot solutions. Depending on the configuration of the Boot Agent, this parameter may not be changeable. Boot Order ⢠Use BIOS Setup Boot Order ⢠Try network first, then local drives ⢠Try local drives first, then network ⢠Try network only ⢠Try local drives only Sets the boot order in which devic es are selected during boot up if the computer does not have its own control method. If your client computer's BIOS supports the BIOS Boot Specification (BBS), or allows PnP-compliant selection of the boot order in the BIOS setup program, then this setting will always be Us e BIOS Set up Boot Order and cannot be changed. In this ca se, refer to the BIOS setup manual specific to your client computer to set up boot options. If your client computer does not have a BBS- or PnP-compliant BIOS, you can select any one of the othe r possible values listed for this setting except for Use BIOS Setup Boot Order. Show Setup Prompt ⢠Enabled ⢠Disabled Controls whether or not the Boot Agent displays the Ctrl S prompt after POST. If you select Enabled, the Ctrl S prompt appears after POST so that you can press Ctrl S within the number of seconds allowed by the Setup Menu Wait Time setting to display the configuration setup menu. If you select Disabled and/or the Setup Menu Wait Time setting is zero, the Ctrl S setup prompt will not appear during the boot process. However, you can display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedly pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again.
Adapter configurations 91 Configuration settings Possible values Description Setup Menu Wait Time ⢠0 seconds ⢠2 seconds ⢠3 seconds ⢠5 seconds Controls the number of seconds the Boot Agent waits for you to press the Ctrl S keys, so as to suspend the boot process and then configure the way the Boot Agent operates. If you select a Setup Menu Wait Time setting of zero and/or a Show Setup Prompt setting of Disabled, the Ctrl S setup prompt will not appear during the boot process. Ho wever, you can still display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedl y pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again. Windows Modifying adapter properties using the NCU This following describes the adapter properties available in the HP Ne twork Configuration Utility (NCU). IMPORTANT: NC31xx, NC6132, NC6133, and NC7132 adapters are not supported in Windows Server 2003 x64 . HP 10/100 server adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default = Auto/Auto o Range = Auto/Aut o; 10/Half; 10/Full; 100/Half; 100/Full NOTE: The NC3133 is a 100BASE-FX fiber optic upg rade module that supports 100 Mbps only. ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Adaptive Performance Tuning: NIC Bandwidth/CPU Utilization (NC31xx adapters only). Specifies the number of frames the adapter receives before trig gering an interrupt. Under normal operation, the adapter generates an interrupt each time a frame is received. Reducing the number of interrupts improves CPU use. Setting a low value for this param eter causes a high rate of receive interrupts (that is, a setting of the lowest value will result in an interrupt for every frame recei ved). A low value increases adapter bandwidth, but may reduce CPU effi ciency, slowing your computer. Setting a high value for this parameter reduces the interrupt rate (that is, increases the number of frames the
Adapter configurations 92 adapter receives before generating an interrupt). A high valu e improves CPU efficiency, but may reduce adapter bandwidth. o Default = 1536 o Range = 0â4096 (increments of 32) ⢠Number of Coalesce Buffers (NC31xx adapters only). Specifies the number of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map registers are available, the driver is forced to queue the packe t for later transmission. The preferred method of transmitting data is to use map registers, because it is the most efficient method. o Default = 8 o Range = 1â 32 ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 48 o Range = 8â1024 (in increments of 8) ⢠Transmit Control Blocks. Specifies how many Transmit Control Blocks (TCBs) the driver allocates for adapter use. This directly corresponds to the number of outstanding packets the driver can have in its send queue. If too few TCBs are used, performance suffers. If too ma ny TCBs are used, the driver unnecessarily consumes memory resources. o Default = 32 o Range = 1â64 ⢠Adaptive Transmit Threshold. Determines how many bytes are in the FIFO before transmission begins. o Default = 200 o Range = 0â200 ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The recommended selection is Off and it is t he default se tting. To disable Flow Control, select Off from the Value list on the Advanced tab. o Off. (Default) PAUSE frame receipt and transmission is disabled o Generate. PAUSE frame transmission is enabled o Respond. PAUSE frame receipt is enabled o Both on. PAUSE frame receipt and transmission is enabled HP NC110x, NC310F, NC340x, NC360x, NC36 4x, NC61xx, and NC71xx server adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default 1000BASE-SX or -LX Adapters: Auto/Auto 10/100/1000 Adapters: Aut o/Auto
Adapter configurations 93 o Range 1000BASE-SX or -LX Adapters: Auto/Auto 10/100/1000 Adapters: Aut o/Auto; 10/Hal f; 10/Full; 100/Half; 100/Full; 1000/Full NOTE: This property is disabled fo r the NC310F and NC61xx adapters. ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. o Default = 1514 Bytes o Range NC310F and NC61xx: 1514 Bytes; 40 88 Bytes; 9014 Bytes; 16128 Byt es NC110x, NC340x, NC360x, NC364x, and NC71xx: 1514 Bytes; 4088 Byt es; 9014 Bytes ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default Windows® 2000: 256 (NC110x and NC36 4x not supported) Windows Server 2003 and Windows Server 2003 x6 4: 128 o Range = 16â768 (increments of 8) ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default Windows® 2000: 256 (NC110x and NC36 4x not supported) Windows Server 2003 Wi ndows Server 2003 x64 NC110x, NC310F, NC340x, NC360x, NC364x, N C6170, and NC7170: 256 All other adapters: 160 o Range Windows® 2000
Adapter configurations 94 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) Windows Server 2003 Wi ndows Server 2003 x64 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) NC110x, NC360x, and NC364x: 80-2048 (increments of 8) ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 256 o Range Windows® 2000 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) Windows Server 2003 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) NC110x, NC360x, and NC364x: 80-2048 (increments of 8) ⢠Rx TCP Checksum Offload. Offloads the task of computing the checksum for incoming TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx TCP Checksum Offload. Offloads the task of computing the checksum for outgoing TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx IP Checksum Offload. Offloads the task of computing th e checksum for outgoing IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Rx IP Checksum Offload. (NC110x, NC310F, NC340 x, NC360x, NC364x, NC6170, and NC7170 only). Offloads the task of computing the checksum for incoming IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled
Adapter configurations 95 ⢠Enable PME. (NC110T, NC310F, NC340x, NC360x, NC364T, NC61 70, and NC7170 only) Allows the use of power management and wake-up functions. o Default = No Action o Range = Disable; Enable; No Action; H/W Default ⢠Wake On. (NC110T, NC310F, NC340T, NC360x, NC364T, NC6170, and NC71 70 only) Available only when Enable PME is enabled. Allows you to choose what types of packets will cause the adapter to wake. o Default = Disable (NC110, NC360x, NC364x Defa ult=OS Controlled) o Range = Disable; Magic Packet; Direct Packe t Magic and Direct Packets; OS Controlled ⢠Large Send Offload. Enables offloading of large TCP packets (NC11 0T, NC310F, NC340x, NC360x, NC364T, NC6170, and NC7170 only). Windows 2003 and Windows 2003 x64 only. o Default = Enabled o Range = Disabled; Enabled ⢠Gigabit Master Slave Mode. (NC110x, NC360x, NC3 64x only) Determines IEEE 802.3ab Master- Slave resolution during gi gabit auto negotiation. o Default = Auto Detect o Range = Auto Detect, Force Master Mode, Force Slave Mode ⢠Smart Power Down. (NC110T, NC360x, and NC364T only) Minimizes power consumpti on by enabling the adapter to enter a deep sleep mode under certain conditions. o Default = H/W Default o Range = Disable, H/W Default, Enable ⢠Wake on Link. (NC110T, NC360x, and NC364T only) Allows wake-up from Advanced Configuration and Power Interface (ACPI) when link is reconnected. Windows 2003 and Windows 2003x64 o Default = Disable o Range = Disable, Force ⢠Flow Control. (Windows Server 2003 only) Determines the type of fl ow control that is needed. (NC110x, NC360x, NC364x only): o Default = Both On o Range = Off, Respond, Generate, Both On HP NC150T, NC32x, NC37xx, NC380T, NC1 020, NC67xx, and NC77xx Server Adapters ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 200 o Range = 100â600 NOTE: This property is not supported on HP Multifunction Server Adapters. ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of
Adapter configurations 96 system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 200 o Range = 100â512 NOTE: This property is not supported on HP Multifunction Server Adapters. ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o NC150T and NC67xx Range = Auto/Aut o Default = Auto/Auto o NC32x, NC37xx, NC380T, NC1020, and NC77xx Range = Auto/Aut o; 10/Half; 10/Fu ll; 100/Half; 100/Full; 1000/Full Default = Auto/Auto ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. When TCP/IP Offload Engine (TOE) is enabled, th is property cannot be set to greater than 1514 bytes. If this property is set to grea ter than 1514 bytes, TOE cannot be enabled. o Default = 1514 Bytes o Range = 1514 Bytes; 4088 Bytes; 9014 B ytes NOTE: This property is not supporte d by NC150T, NC32x, NC1020, and NC7761 adapters. ⢠Wake Up Capabilities. Determines the type of packets that will cause the adapter to wake up the system. o Default = Both o Range = None; Magic Packet; Wake Up Frame; Both NOTE: This property is not supported on HP Multifunction Server Adapters. ⢠WOL Speed. (Not supported on NC150T, NC37x, NC380T, NC67xx, and NC7782 adapters) Specifies the speed at which the adapter connects to the network during Wake-on- LAN mode.
Adapter configurations 97 o Default = Auto o Range = Auto, 10Mb, 100Mb ⢠Flow Control. Specifies the type of flow control that is needed. o Default = Auto o Range = Disable; Rx Pause; Tx Pause; Rx/Tx Pause; Auto ⢠Rx TCP Checksum Offload. Offloads the task of computing the checksum for incoming TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx TCP Checksum Offload. Offloads the task of computing the checksum for outgoing TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx IP Checksum Offload. Offloads the task of computing th e checksum for outgoing IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Rx IP Checksum Offload. Offloads the task of computing th e checksum for incoming IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Ethernet@WireSpeed (NC37xx, NC380x and NC77xx only). Enables linking at 100 Mbps when using a cable that does not support Gigabit speed. o Default = Enabled o Range = Disabled; Enabled NOTE: This property is not su pported by NC7761 adapters. ⢠Large Send Offload. Normally, the TCP segmentat ion is performed by the protocol stack. By enabling the Large Send Offload parameter, the TCP segmentation can be performed by the Gigabit Ethernet Adapter. To enable the Large Send Offload parameter, select Enable fr om the Value list on the Advanced tab. o Default = Enabled o Range = Disabled; Enabled ⢠IPMI . Enables Intelligent Platform Management Inte rface, which provides a means for monitoring, control, and automatic recovery of servers. IPMI can only enabled on one port at a time. IPMI cannot be enabled on a port that has Large Send Offload enabled. (This property is only supported on the NC320i, NC7781 and NC7782 adapters.) Multifunction Gigabit Server Adapters only ⢠iSCSI properties
Adapter configurations 98 o SRB Timeout Delta. Increases or decreases the tim eout value for I/O operations through the iSCSI HBA. Increasing this value may help prevent prem ature I/O timeouts if yo ur network is highly congested and packets are being dropped frequently. The default setting is 15 seconds. This property only displays if iSCS I is enabled on the adapter. o iSCSI VLAN. Allows you to create a VLAN for an iSCSI-enabled device. ⢠TCP/IP Offload Engine (TOE). Offloads TCP connections to th e Multifunction Server Adapter. (Microsoft Scalable Networking Pack (SNP) fo r Windows 2 003 SP1 or later and Windows 2003 x64 SP1 or later environments only.) See the Microsoft ( http://www.microsoft.com ) website to download the latest version of the Microsoft Wind ows Server 2003 Scalable Networking Pack (SNP) and the required Microsoft hotfixes in arti cle numbers 921136, 919948, and 923187 o f the Microsoft Knowledge Base (KB). o Default = Enabled o Range = Disabled; Enabled ⢠Receive-Side Scaling (RSS). Allows the network load from a ne twork adapter to be balanced across multiple CPUs. (Microsoft Scalable Networking Pack (SNP) for Windows 2003 SP1 or later and Windows 2003 x64 SP1 or later) o Default = Enabled o Range = Disabled; Enabled ⢠Optimized Interrupt Moderation . When this configuration is set to optimize, the driver automatically and continuously determines the best Target DPC (d eferred procedure call) Rate to use based on the current network conditions. When th is configuration is set to manual , the user can select a Target DPC Rate and the driver adjusts the adapter's inte rrupt setting to achieve that DPC rate. Changing this setting may improve performance for some conf igurations. ⢠Target DPC Rate . Determines the desired DPC rate. The driver adjusts the adapter's interrupt moderation setting dynamically bas ed on network conditions to achieve the desired DPC rate. This configuration is valid only when "Optimized Interrupt Moderation" is set to manual. Modifying HP NC-Series Broadcom and Multifunction adapter properties Although the default values should be appropriate in most cases, you may change any of the available options to meet the requirements of your specific system. Use the followin g procedure to verify or change the properties after the adapter driver is installed. To change the adapter properties, display the Driver Properties Advanced tab as follows: 1. On the desktop, click Start then select Control Panel > System to display the System Properties window. 2. Select the Hardware tab and click Device Manager . The Device Manager window appears. 3. Scroll down the list of hardware devices to Network Adapters and click the plus ( ) . This displays the list of adapters currently configured. 4. Right-click the adapter to configure and select Properties. The Properti es window appears. 5. Select the Advanced tab. The list of properties and cu rrent values for the adapter appears. 6. Select the property to modify and adjust the Value as needed. 7. Save the settings. a. When the adapter property conf iguration is complete, click OK to accept the settings.
Adapter configurations 99 b. If prompted to restart your computer, click Yes . Rebooting the system for new adapter properties to take effect is not required; however, rebooting is recommended to re-initialize all regi sters. Windows Server 2003 and Windows Server 2003 x64 adapter properties ⢠802.1p QoS Packet Tagging . The 802.1p QoS parameter is a standard that enables Quality of Service. This property is disabled by de fault. To change this property, select Enable or Disable. ⢠Checksum Offload. Normally the Checksum Offload function is computed by the protocol stack. By selecting one of the Checksum Offload properti es , the checksum can be computed by the Gigabit Ethernet Adapter. To enable one of the Checksum Offload properties s elect the property from the Value list on the Advanced tab. o None. Disables checksum offloading o Rx TCP/IP Checksum. Enables receive TC P, IP, and UD P checksum offloading o Tx TCP/IP Checksum. Enables transmit TCP, IP, and UDP checksu m offloading o Tx/Rx TCP/IP Checksum. (Default) Enables transmit and receive TCP, IP, and UDP checksum offloading ⢠Ethernet@WireSpeed. This property is only available throug h the HP Network Configuration Utility. The Ethernet@WireSpeed property enables a 1000Bas e-T Ethernet adapter to establish a link at a lower speed when only two pairs of wires are available in the cable plant. By default, this property is enabled. To disable the Ethernet @ WireSpeed prop erty, select Disable from the Value list. o Disable. Disables Ethernet@Wir eSpeed o Enable. (Default) Enables Ethernet@WireSpeed ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The recommended selection is Auto and it is the default setting. To dis able Flow Control, select Disable from the Value list on the Advanced tab. o Auto. (Default) Auto detect flow control o Disable. PAUSE frame receipt an d transmission is disabled o Tx PAUSE. PAUSE frame transmission is enabled o Rx PAUSE. PAUSE frame receipt is enabled o Rx/Tx PAUSE. PAUSE frame receip t and transmission is enabled ⢠Large Send Offload. Normally, the TCP segmentat ion is performed by the protocol stack. By enabling the Large Send Offload property, the TC P segmentation can be performed by the Gigabit Ethernet Adapter. To enable the Large Send Offl oad property, sele ct Enable from the Value drop- down menu on the Advanced tab. o Disable. Disables large send offloading o Enable. Enables large send offloading ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = Not Present o Range = Value; Not Present
Adapter configurations 100 ⢠Speed & Duplex. The Speed & Duplex property allows the user to s et the connection speed to the network and mode. The Duplex Mode allows the adapter to transmit and receive network data simultaneously. The adapter is set to Auto (optimum connection) by default. Set the speed and mode as described below: o 1 Gb Full Auto. Sets the speed to 1 Gbps and mode to Full-Duplex o 10 Mb Full. Sets the speed to 10 Mbps and mode to Full-Duplex o 10 Mb Half. Sets the speed to 10 Mbps and mode to Half-Duplex o 100 Mb Full. Sets the speed to 100 Mbps and mode to Full -Duplex o 100 Mb Half. Sets the speed to 100 Mbps and mode to Half-Duplex o Auto. (Default) Sets the speed and mode fo r optimum network connection (recommended) NOTE: Auto is the recommended selection. It allo ws the adapter to dynamically detect the line speed and duplex mode of the network. Wh enever the network capability changes, the adapter will automatically detect and adjust to the new line sp eed and duplex mode. The Half-Duplex selection forces the adapter to connect to the network in Half-Duplex mode. The adapter may not function if the network is not configured to operate at the same mode. The Full-Duplex selection forc es the adapter to connect to the network in Full-Duplex mode. The adapter may not function if the network is no t configured to operate at the same mode. ⢠Wake Up Capabilities. This property is only available throug h the HP Network Configuration Utility. The Wake Up Capabilities property allows the user to set the adapte r to wake up from a low power mode when it receives a networ k wake up frame. Two wake up frames are possible: Magic Packet and Wake Up Frame. By default, the adapter is set to Both. To choose the type of frame that will cause the adapter to wake up, select the appropriate option from the Value list on the Advanced tab. o Both. (Default) Selects both Magic Packet and Wake up Frame as wake up frames o Magic Packet. Selects Magic Pa cket as the wake up frame o None. Selects no wake up frames o Wake Up Frame. Selects Wake up Frame as the wake up frame ⢠WOL Speed. This property is only available through the HP Network Confi guration Utility. The WOL Speed property allows the user to select t he sp eed at which the adapter connects to the network while in Wake-on-LAN mode. By default, the adapter is set to Auto. o 10 Mb. Sets the speed to 10 Mb o 100 Mb. Sets the speed to 100 Mb o Auto. (Default) Sets the speed for optimum network connection ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers are alloca ted).This property is only available through the HP Network Configuration Utility. o Default = 200 o Range = 100â600 ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too
Adapter configurations 101 many receive descriptors are used, the driver un necessarily consumes me mory resources. This property is only available through the HP Network Configuration Utility. o Default = 200 o Range = 100â512 Multifunction Gigabit Server adapter properties ⢠SRB Timeout Delta. Increases or decreases the tim eout value for I/O operations through the iSCSI HBA. Increasing this value may help prevent prem ature I/O timeouts if yo ur network is highly congested and packets are being dropped frequently. The default setting is 15 seconds. This property only displays if iSCSI is enabled on the adapter. iSCSI capabilities are only available through HP Multifunction Gigabi t Server Adapters and Multifunction Server Adapters for HP BladeSystem running Windows 2000, Window s Se rver 2003, and Windows Server 2003 x 64. ⢠Receive-Side Scaling (RSS). Allows the network load from a ne twork adapter to be balanced across multiple CPUs. (Microsoft Scalable Networking Pack (SNP) for Windows 2003 SP1 or later and Windows 2003 x64 SP1 or later) ⢠TCP Offload Engine (TOE). Offloads TCP connections to the Mult ifunction Server Adapter. (Microsoft Scalable Networking Pack (SNP) for Windows 2003 SP1 or later an d Windows 2003 x64 SP1 or later environments only.) See the Microsoft ( http://www.microsoft.com ) website to download the latest version of the Microsoft Windows Serv er 2003 Scalable Networking Pack (SNP) and the required Microsoft hotfixes in article num bers 921136, 919948, an d 923187 of the Microsoft Knowledge Base (KB). Windows 2000 adapter properties ⢠802.1p QoS. The 802.1p QoS parameter is a standard th at enables Quality of Service. This parameter is disabled by default. To change this parameter, select Enable or Disable . ⢠Checksum Offload. Normally the Checksum Offload function is computed by the protocol stack. By selecting one of the Checksum Offload parameter s , the checksum can be computed by the Gigabit Ethernet Adapter. To enable one of the Checksum Offload parameters, select the parameter. o None. Disables checksum offloading o Rx TCP/IP Checksum. Enables receive TC P, IP, and UD P checksum offloading o Tx TCP/IP Checksum. Enables transmit TCP, IP, and UDP checksu m offloading o Tx/Rx TCP/IP Checksum. (Default) Enables transmit and receive TCP, IP, and UDP checksu m offloading ⢠Ethernet@WireSpeed. The Ethernet@WireSpeed parameter enables a 1000Base-T Ethernet adapter to establish a link at a lower speed when only two pairs of wires are available in the cable plant. By default, this parameter is en abled. To disable the Ethernet@WireSpeed parameter, select Disable from the Value list on the Advanced tab. o Disable. Disables Ethernet@Wir eSpeed o Enable. (Default) Enables Ethernet@WireSpeed NOTE: This property is not supported by NC1020 and NC7761 adapters. ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The
Adapter configurations 102 recommended selection is Auto and it is the default setting. To dis able Flow Control, select Disa ble from the Value list on the Advanced tab. o Auto. (Default) Auto detect flow control o Disable. PAUSE frame receipt an d transmission is disabled o Tx PAUSE. PAUSE frame transmission is enabled o Rx PAUSE. PAUSE frame receipt is enabled o Rx/Tx PAUSE. PAUSE frame receip t and transmission is enabled ⢠Speed & Duplex. The Speed & Duplex parameter allows the user to set the connection speed to the network and mode. The Duplex Mode allows the adapter to transmit and receive network data simultaneously. The adapter is set to Auto (optimum connection) by default. Set the speed and mode as described below: o 10 Mb Full. Sets the speed to 10 Mbps and mode to Full-Duplex o 10 Mb Half. Sets the speed to 10 Mbps and mode to Half-Duplex o 100 Mb Full. Sets the speed to 100 Mbps and mode to Full -Duplex o 100 Mb Half. Sets the speed to 100 Mbps and mode to Half-Duplex o Auto. (Default) Sets the speed and mode fo r optimum network connection (recommended) NOTE: 1 Gbps speed is enabled by selecting Auto. NOTE: Auto is the recommended selection. It allo ws the adapter to dynamically detect the line speed and duplex mode of the network. Wh enever the network capability changes, the adapter will automatically detect and adjust to the new line sp eed and duplex mode. The Half-Duplex selection forces the adapter to connect to the network in Half-Duplex mode. The adapter may not function if the network is not configured to operate at the same mode. The Full-Duplex selection forc es the adapter to connect to the network in Full-Duplex mode. The adapter may not function if the network is no t configured to operate at the same mode. ⢠Wake Up Capabilities. The Wake Up Capabilities parameter a llows the user to set the adapter to wake up from a low power mode when it receives a network wake up frame. Two wake up frames are possible: Magic Packet and Wake Up Frame. By default, the adapter is set to Both. To choose the type of frame that will cause the adapter to wake up, select the appropriate option from the Value drop-down list on the Advanced tab. o Both. (Default) Selects both Magic Packet and Wake up Frame as wake up frames o Magic Packet. Selects Magic Pa cket as the wake up frame o None. Selects no wake up frames o Wake Up Frame. Selects Wake up Frame as the wake up frame NOTE: This property is not su pported by NC150T adapters. ⢠WOL Speed. The WOL Speed parameter allows the user to select the spe ed at which the adapter connects to the network while in Wake-on-LAN mode . By default, the adapter is set to Auto. o 10 Mb. Sets the speed to 10 Mb o 100 Mb. Sets the speed to 100 Mb
Adapter configurations 103 o Auto. (Default) Sets the speed fo r optimum network connection NOTE: This property is not supported on the NC150T and NC7782 adapters. Multifunction Gigabit Server adapter iSCSI properties ⢠SRB Timeout Delta. Increases or decreases the tim eout value for I/O operations through the iSCSI HBA. Increasing this value may help prevent prem ature I/O timeouts if yo ur network is highly congested and packets are being dropped frequently. The default setting is 15 seconds. This property only displays if iSCSI is enabled. Modifying HP NC-Series Intel adapter properties Although the default values should be appropriate in most case s, you may change any of the available options to meet the requirements of your specific system. Use the followin g procedure to verify or change the properties after the adapter driver is installed. To change the adapter properties, display the Driver Properties Advanced tab as follows: 1. On the desktop, click Start then select Control Panel > System to display the System Properties window. 2. Select the Hardware tab and click Device Manager . The Device Manager window appears. 3. Scroll down the list of hardware device s to Network Adapters and click the plus ( ). This displays the list of adapters currently configured. 4. Right-click the adapter to configure and select Properties. The Properti es window appears. 5. Select the Advanced tab. The list of properties and cu rrent values for the adapter appears. 6. Select the property to modify and adjust the Value as needed. 7. Save the settings. a. When the adapter property conf iguration is complete, click OK to accept the settings. b. If prompted to restart your computer, click Yes . Rebooting the system for new adapter properties to take effect is not required; however, rebooting is recommended to re-initialize all regi sters. Adapter properties ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 128 (for NC110x, NC360x, NC364x = 25 6) o Range = 16â768 (increments of 8) ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The
Adapter configurations 104 recommended selection is Auto and it is the default setting. To dis able Flow Control, select Disa ble from the Value list on the Advanced tab. o Auto. (Default) Auto detect flow control o Disable. PAUSE frame receipt an d transmission is disabled o Tx PAUSE. PAUSE frame transmission is enabled o Rx PAUSE. PAUSE frame receipt is enabled Rx/Tx PAUSE. PAUSE frame receip t and transmission is enabled NOTE: User control of Flow Control parameter is not available for NC71xx adapters. ⢠Link Speed & Duplex. Allows the user to set the connectio n speed to the network and mode. Full- Duplex Mode allows the adapter to transmit and re ceive network data simultaneously. This property is not available for NC310F and NC61xx adapters through the Microsoft® User Interface. o Default = Auto Detect o Range = NC340x and NC71xx Adapters: Auto Detect; 10Mbps/Half-Duplex; 10Mbp s/Full- Duplex; 100Mbps/Half-Duplex; 100Mbps/ Full-D uplex; 1000 Mbps/Full-Duplex ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = Not Present o Range = Value; Not Present ⢠Number of Receive Descriptors. Specifies the number of buffers used by the driver when copying data to the protocol memory. In high network lo ad situations, increasing receive desc riptors can increase performance. The tradeoff is that this also increases the amount of syste m memory used by the driver. If too few receive buff ers are used, perf ormance suf fers. If too many receive buffers are used, the driver unnecessarily consum es memory resources. o Default NC110x, NC310F, NC340x, NC360x, NC364x, N C6170, and NC7170 = 256 All others = 160 o Range NC6132, NC6133, and NC6134 = 80â768 (incr ements of 8) NC6136, NC7131, and NC7132 = 80â256 (incr ements of 8) NC310F, NC340x, NC6170, and NC7170 = 80â4 096 (increments of 8) NC110x, NC360x, and NC364x = 80-2048 (increments of 8 ) ⢠Number of Transmit Descriptors. Specifies the number of resources allocated to transmit packets. o Default = 256 o Range NC6132, NC6133, and NC6134 = 80â768 (incr ements of 8) NC6136, NC7131, and NC7132 = 80â256 (incr ements of 8) NC310F, NC340x, NC6170, and NC7170 = 80â4 096 (increments of 8) NC110x, NC360x, and NC364x = 80-2048 (increments of 8 )
Adapter configurations 105 ⢠Offload Receive TCP Checksum. Offloads the task of computing the checksum for incoming TCP or UDP packets, thereby improving performance. o Default = On o Range = Off; On ⢠Offload Transmit TCP Checksum. Offloads the task of computing the checks um for outgoi ng TCP or UDP packets, thereby improving performance. o Default = On o Range = Off; On ⢠Offload Transmit IP Checksum. This property does not apply to the NC6132, NC6133 or NC6134 adapters. Offloads the task of computing the checksum for outgoing IP packets, thereby improving performance. o Default = On o Range = Off; On ⢠Offload Receive IP Checksum. (NC110x, NC310 F, NC340x, NC360x, NC364x, NC6170, and NC7170 only). Offloads the task of computing the checksum for incoming IP packets, thereby improving performance. o Default = On (NC110, NC360x, NC364x = Off) o Range = Off; On ⢠Enable PME (NC110x, NC310F, NC340x, NC360x, NC364x, N C6170, and NC7170 only). Allows the use of power management and wake-up functions. o Default = No Action (NC110x, NC360x, NC364x = OS Controlled) o Range = Disabled; Enabled; Hardware Default; No Action (NC110x, NC360x, NC36 4x = OS Controlled, Disabled, Enabled) ⢠Wake On Settings (NC110x, NC310F, NC34 0x, NC360x, NC364x, NC6170, and NC7170 only). Available only when Enable PME is enabled. Al lows you to choose what types of packets will cause the adapter to wake. o Default = Disabled (NC110x, NC360x, NC364x = OS Controlled) o Range = Disabled; Wake on Directed Packet; Wake on Magic & Directed; Wake on Magic Packet; OS Controlled ⢠Large Send Offload (Windows® Server 2003 only; NC 110x, NC310F, NC340T, NC360x, NC364x, NC6170, and NC7170 only). Enables offloading of larg e TCP packets. o Default = On o Range = Off; On ⢠Smart Power Down. (NC110T, NC360x, and NC364T only) Minimizes power consumpti on by enabling the adapter to enter a deep sleep mode under certain conditions. o Default = Hardware Default o Range = Disable, Enable, Hardware Default ⢠Wake on Link. (NC110T, NC360x, and NC364T only) Allows wake-up from Advanced Configuration and Power Interface (ACPI) when link is reconnected. o Default = Disable o Range = Disable, Force
Adapter configurations 106 Modifying NC31xx Fast Ethernet adapter properties Although the default values should be appropriate in most cases, you may change any of the available options to meet the requirements of your specific system. Use the followin g procedure to verify or change the properties after the adapter driver is installed. To change the adapter properties, display the Driver Properties Advanced tab as follows: 1. On the desktop, click Start then select Control Panel > System to display the System Properties window. 2. Select the Hardware tab and click Device Manager . The Device Manager window appears. 3. Scroll down the list of hardware devices to Network Adapters and click the plus ( ) . This displays the list of adapters currently configured. 4. Right-click the adapter to configure and select Properties. The Properti es window appears. 5. Select the Advanced tab. The list of properties and cu rrent values for the adapter appears. 6. Select the property to modify and adjust the Value as needed. 7. Save the settings. a. When the adapter property conf iguration is complete, click OK to accept the settings. b. If prompted to restart your computer, click Yes . Rebooting the system for new adapter properties to take effect is not required; however, rebooting is recommended to re-initialize all regi sters. Adapter Properties ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p tagging to send network traffic with different priority levels. o Default = Off o Range = Off; On ⢠Coalesce Buffers. Specifies the number of memory buffers avai lable to the driver in case the driver runs out of available map registers. This buffer ar ea is also used when a packet consists of many fragments. If no coalesce buffer s or map registers are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 8 o Range = 1â 32 ⢠Flow Control Settings. The Flow Control property allows the user to enable or disable the recei pt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The recommended selection is Off and it is t he default se tting. To disable Flow Control, select Off from the Value list on the Advanced tab. o Default = Off o Range = Off, Generate, Re spond, Respond and Generate Off. PAUSE frame receipt and transmission is disabled Generate. PAUSE frame transmission is enabled Respond. PAUSE frame receipt is enabled Respond and Generate. PAUSE frame re ceipt and transmission is enabled
Adapter configurations 107 ⢠Link Speed & Duplex. Allows the user to set the connectio n speed to the network and mode. Full- Duplex Mode allows the adapter to transmit and receive network data simultaneously. o Default = Auto Detect o Range 100 Mbps/Full-Duplex: Sets the speed at 100 Mbps and mode to Full-D uplex 100 Mbps/Half-Duplex: Sets the speed at 100 Mbps and mode to Half-Dupl ex 10 Mbps/Full-Duplex: Sets the speed at 10 Mbps and mode to Full-Duplex 10 Mbps/Half-Duplex: Sets the speed at 10 Mbps and mode to Half-Duplex Auto Detect: Sets the speed and mo de for optimum network connection ⢠Locally Administered Address. Specifies the user-define d MAC Address of the adapter; overrides the burned-in MAC Address. This is a read-o nly field if the adapter is teamed. o Default = Not Present o Range = Value; Not Present ⢠Receive Buffers. Specifies the number o f buffers used by the driver when copying data to the protocol memory. In high network load situ ations, increasing receive buffers can increase p erformance. The tradeoff is that this also increases the amount of sy stem memory used by the driver. If too few receive buffers are used, performance suf fers. If too many receive buffers are used, the driver unnecess arily consumes memory resources. o Default = 48 o Range = 8â1024 (increments of 1) ⢠Transmit Control Blocks. Specifies the number of resources that are allocated to transmit packets. o Default = 16 o Range = 1â64 Modifying adapter properties in Windows NT This document describes the Windows NT® 4.0 adapter properties available for fiber and copper server adapters. Properties for 10/100 Co pper Server Adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default = Auto/Auto o Range = 10/Half; 10/Full; 100/Half; 100/Full; Auto /Auto NOTE: The NC3133 is a 100BASE-FX fiber optic upg rade module that supports 100 Mbps only. ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address
Adapter configurations 108 ⢠Adaptive Interframe Spacing. Compensates for excessive Ethernet p acket collisions on the network. The default setting works best for most c omputer s and networks by dynamically adapting to the network traffic conditions. However, in some rare cases, you ma y obtain better performance by manually setting the spacing value. Setting a valu e forces a static gap between packets. Increasing the value increases the delay between frames bein g transmitted. o Default = 1 o Range = 1â255 ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 8 o Range = 1â32 ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 32 o Range = 8â1024 (in increments of 8) ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 12 o Range = 1â16 ⢠Transmit Control Blocks. Specifies how many Transmit Control Blocks (TCBs) the driver allocates for adapter use. This directly corresponds to the number of outstanding packets the driver can have in its send queue. If too few TCBs are used, performanc e suffers. If too ma ny TCBs are used, the driver unnecessarily consumes memory resources. o Default = 16 o Range = 1â80 ⢠Transmit Threshold. Specifies the number of bytes before the PCI adapter emptie s its internal transmit FIFO onto the wire. The value is multiplied by 8 to produce the number of bytes. For example, if Transmit Threshold = 200, the number of bytes is 1600. This is greater than the maximum packet size for Ethernet. Consequently, the adapter will no t attempt early transmits. Although this is the safest setting, the best performance is ac hieved wh en the Threshold parameter is as low as possible without producing underruns. o Default = 200 o Range = 0â200 Properties for NC61 xx and NC71xx Server Adapters Exception: HP NC6170 and NC7170 Server Adapte rs are not supported under Windows NT® 4.
Adapter configurations 109 ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default 1000BASE-SX or -LX Adapters: 1000/Fu ll 10/100/100 Adapters: Auto/Auto o Range 1000BASE-SX or -LX Adapters: 1000/Fu ll 10/100/100 Adapters: Auto/Auto; 10/ Half ; 10/Full; 100/Half; 100/Full; 100/Full ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. o Default = 1514 Bytes o Range = 1514 Bytes; 4088 Bytes; 9014 B ytes; 16128 Bytes NOTE: All Gigabit team members must support maximum frame size for this parameter to be enabled. If any one team member is not supported, then it is disabled for the team. Size of the frame is determined by the lowest value for the team members. The 16128 sett ing applies for Fiber Gigabit Ethernet server adapters only. ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also us ed when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 128 o Range = 16â768 (increments of 8) ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 160 o Range NC6132, NC6133, and NC6134: 64â76 8 (increments of 8) NC6136, NC7131, and NC7132: 64â25 6 (increments of 8) ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 256 o Range NC6132, NC6133, and NC6134: 64â76 8 (increments of 8) NC6136, NC7131, and NC7132: 64â25 6 (increments of 8)
Adapter configurations 110 Properties for NC67 xx and NC77xx Server Adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default = Auto/Auto o Range = Auto/Aut o; 10/Half; 10/Fu ll; 100/Half; 100/Full; 1000/Full ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. o Default = 1514 Bytes o Range = 1514 Bytes; 4088 Bytes; 9014 B ytes NOTE: All Gigabit team members must support maximum frame size for this parameter to be enabled. If any one team member is not supported, then it is disabled for the team. Size of the frame is determined by the lowest value for the team members. The 16128 sett ing applies for Fiber Gigabit Ethernet server adapters only. ⢠Flow Control. Specifies the type of flow control that is needed. o Default = Disable o Range = Disable; Rx Pause; Tx Pause; Rx/Tx Pause; Auto Novell Netware Advanced Server Program for NetWare The following describes how to install and co nfigure th e Advanced Server Program (ASP). Installing ASP IMPORTANT: Teaming NC150, NC32x, NC37x, NC1020, NC6770, or NC77xx adapters with NC110x, NC310, NC31xx, NC340x , NC360x, NC364x, NC61xx, or NC71xx adapters is not supported for NetWare. Load QASP.LAN just like a standard LAN driver with all necessary frame types for the team. QASP.LAN requires a special VSLOT parameter to specify the virt ual slot. The virtual slot can be viewed as a team number. NOTE: Before loading the network drivers, the recommended sequence is to loa d QASP.LAN. This allows QASP.LAN to determine the initial li nk state of the bound adapters without delay. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 If you are using the NetWare NIC agent ( for example, Insight Manager) and teaming, you must load BMAPI.NLM. Loading network drivers To load the network drivers:
Adapter configurations 111 1. Load the network drivers for the network adapters th at will be part of the team. The frame types loaded should be the same for all adapters in the team and the same as those loaded for QASP.LAN. Do not bind protocols directly to these adapters. 2. Load ODINEB.NLM (a Novell supplied NLM) before all network drivers. Example: LOAD ODINEB.NLM LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 3. Load QASP.LAN. a new screen is displayed with all virtual adapter settings and statistics. Press the Alt Esc keys to switch back to the console and continue with step 3. 4. Bind QASP.LAN to the network adapters in the team by using a custom QASP BIND command at the console. Example: BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII If there are multiple frame types loaded on the virt ual and the physical adapters, it is only necessary to bind one frame type on the virtual adapter to the same frame type on the physic al adapter. The other frame types will be automatically bound. 5. Bind protocols to QASP.LAN. Example: BIND IP QASP_1_EII ADDR=x.x.x.x MASK=x.x.x.x NOTE: Configuration of QASP.L AN should be performed manually by editing the AUTOEXEC.NCF file. NWCONFIG.NLM (or INSTALL.NLM) cannot completely configure QASP.LAN. Uninstalling ASP To uninstall the adapter and the QASP drivers: 1. For the adapter driver, at the Command Line Interface (CLI) enter the following command: UNLOAD Q57 The response is: Q57 "Driver Name" unload Module Q57.LAN unloaded 2. For the QASP driver, at the Command Line Interface (CLI) ent er the following command: UNLOAD QASP The response is: Module QASP.LAN unloaded NOTE: QASP cannot be unloaded if one or more adapters are bound to QASP. SLB and TRUNK MODE selection Use "MODE=SLB" for Smart Load-B alance mode or "MODE=TRUNK" for Generic Trunking mode. The default is SLB. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 MODE=TRUNK
Adapter configurations 112 NOTE: In SLB mode, IPX traffic is load-balanced on the send side but not on the receive side. Loading frame types After one or more network adapters are bound to a vi rtual adapter, additional frame types can only be loaded in the virtual adapter if the corresponding fram e types are also loaded in the bound adapters. For example, ETHERNET_802.2 can be loaded in QASP VS LOT 1 only if ETHERNET_802.2 is loaded for the Q57 driver in SLOT 1 and 2. Similarly, a virtual adapter can only be bound to a physical adapter if the ph ysical adapter has all the frame types loaded in the virtual adapter. Example: LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_802.2 NAME=QASP_E82 VSLOT=1 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_1_E82 SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_2_E82 SLOT=2 Hot STANDBY In Smart Load-Balance (SLB) mode, one or more adapters can be designated as hot standbys. Use the keyword STANDBY in the QASP BIND command to indicate binding an adapter as a hot standby. Standby is only valid for Smart Load-Balance mode. Example: BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII STANDBY In the above example, Q57_1_EII and Q57_2_EII are bound as primary and hot standby adapters respectively. Configuring VLANs NOTE: VLANs are not supported on non-HP adapters. To add VLANs to a team: 1. Load QASP.LAN with the all necessary frame types and specify the VLAN ID for ea ch frame type. You can specify a maximum of 64 VLAN IDs and each VLAN ID can be loaded up to four times with four different frame types. Example: (VLAN ID 2 for Ethernet II) LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 2. Load the network drivers for the adapters in the team with all the frame type s specified in step 1. One or more VLAN IDs specified in step 1 do no t have to be specified when loading the network drivers. Each frame type loaded in step 1 only needs to be loaded once for each network driver even if in step 1 it is loaded multiple tim es with different VLAN IDs. Example: LOAD ODINEB.NLM LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 3. Bind QASP.LAN to the adapters in the team for each protocol. Example: BIND QASP1 Q57_1_EII
Adapter configurations 113 BIND QASP1 Q57_2_EII If there are multiple VLANs (each with one or more frame types) loaded on the virtual adapter, it is only necessary to bind one frame type on one VLAN on the virtual adapter to the same frame type on the physical adapter. The other VLANs will be automatically bound. 4. Bind protocols to QASP.LAN. Example: BIND IP QASP_1_V2_EII ADDR=x.x.x.x MASK=x.x.x.x This example creates a team with two adapters us ing VLAN ID 2. Outbound packets will be tagged with VLAN ID 2 and only similarly tagged packets will be received by the adapters in the team. Additional VLANs with different VLAN IDs can be created in the same team. The maximum number of VLANs per virtual slot is 64. The valid range of VLAN IDs is from 1 to 4094. VLAN=0 indicates the VLAN is untagged and is the default. Us e decimal numbers to specify the VLAN ID. The following are examples of multiple VLAN configura tions: LOAD QASP FRAME=ETHERNET_II NAME=QASP_1_V100_EII VSLOT=1 VLAN=100 LOAD QASP FRAME=ETHERNET_II NAME=QASP_1_V200_EII VSLOT=1 VLAN=200 LOAD QASP FRAME=ETHERNET_II NAME=QASP_1_V300_EII VSLOT=1 VLAN=300 LOAD ODINEB.NLM LOAD Q57 FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57 FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=4 VLAN=1 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_V100_EII ADDR=172.16.210.1 MASK=255.255.0.0 BIND IP QASP_1_V200_EII ADDR=172.17.210.1 MASK=255.255.0.0 BIND IP QASP_1_V300_EII ADDR=172.18.210.1 MASK=255.255.0.0 NOTE: When BIND QASP1 Q57_1_EII is ex ecuted, the adapter Q57_1_EII is b ound to all three VLANs. If you are unable to log in to th e server after configurin g ASP, add the following command lines before loading ASP: UNLOAD SLPTCP LOAD SLPTCP Jumbo Frames Jumbo Frames are supported in all balance modes. Th e maximum frame size will be automatically set to the smallest maximum frame size of all adapters in the team. Use appropriate keywords to ena ble jumbo frames when loading the adapter drivers. NOTE: This property is not supported on the NC150T and NC1020. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 JUMBO=9000 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 JUMBO=9000 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII Additional command line keywords Additional command line keywords include: ⢠CHECKSUM=ON. This keyword enables QASP .LAN to offload TCP/UDP and IP checksums to the bound adapters if supported by the OS. This improv es performance if some or all adapters in the
Adapter configurations 114 team support hardware checksums. Be sure to load the adapter drivers with hardware checksums enabled. Default is ON. Example: A team of two adapters with hardware checksums enabled. LOAD QASP.LAN FRAME=ETHERNET_II NAME=Q57_1_EII CHECKSUM=ON VSLOT=1 LOAD ODINEB.NLM LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII CHECKSUM=ON SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII CHECKSUM=ON SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII ⢠NOSCREEN. This keyword disables the menu-driven screen when QASP.LAN is loaded for the first time. By default, this screen/menu int erface will be loaded when QASP.LAN is loaded and cannot be turned off without unloading QASP.LAN. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 NOSCREEN ⢠GVRP. This keyword enables GVRP (GARP VLAN Registra tion Protocol) for the VLAN that is loaded. An untagged 802.2 frame type must be loaded in the virtual adapter and all bound physical adapters for GVRP to take effect. This is necessary because GVRP uses untagged 802.2 frames to advertise VLAN memberships. Use VLAN =0 FRAM E=ETHERNET_802.2 in the LOAD command to specify untagged 802.2 frame type. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 GVRP LOAD QASP.LAN FRAME=ETHERNET_802.2 NAME=QASP_1_E82 VSLOT=1 VLAN=0 Editing the AUTOEXEC.NCF file When an adapter configuration is saved, the NetWar e install program adds load and bind statements to the AUTOEXEC.NCF file. By accessing this file, you can verify the parameters configured for e ach adapter, add or delete parame ters, or modify parameters. A valid AUTOEXEC.NCF file is shown below wi th various VLAN and teaming examples. ⢠Example 1: Team of 2 network adapters with frame type Ethernet_II and one VLAN, number 2. Load QASP.LAN with the frame types and VLAN ID(s) specified. LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 Load the network drivers for the adapters in the team with the same frames types. LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 Bind QASP.LAN to the adapters in the team for each protocol. BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII Bind protocols to QASP.LAN. BIND IP QASP_1_V2_EII ADDR=192.168.2.200 MASK=255.255.255.0 ⢠Example 2: Team of two network adapters with fram e type Ethernet_II and three VLANs, number 2, 3, and 4.
Adapter configurations 115 Load QASP.LAN with the frame types and VLAN ID(s) specified. LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V3_EII VSLOT=1 VLAN=3 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V4_EII VSLOT=1 VLAN=4 Load the network drivers for the network adapters in the team with the same frame types and VLANs specified. LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 Bind QASP.LAN to the network adapters in the team for each protocol. QASP BIND is only used for the first VLAN all other VLANs are automatica lly # bound to the virtual adapter (VSLOT=1). BIND QASP1 Q57_1_V_EII BIND QASP1 Q57_2_V_EII Bind protocols to QASP.LAN. BIND IP QASP_1_V2_EII ADDR=192.168.2.200 MASK=255.255.255.0 BIND IP QASP_1_V3_EII ADDR=192.168.3.200 MASK=255.255.255.0 BIND IP QASP_1_V4_EII ADDR=192.168.4.200 MASK=255.255.255.0 mount all NOTE: If you modify any adapter parameters, you must reboot the system for the changes to take effect. If you make changes and do no t reboot, you may experience configuration problems. ⢠Example 3: Multiple SLB Teams LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_II NAME=QASP_1_EII LOAD Q57.LAN SLOT=2 FRAME=Ethernet_II speed=100fd NAME=Q57_3_EII LOAD Q57.LAN SLOT=3 FRAME=Ethernet_II speed=100fd NAME=Q57_4_EII BIND QASP1 Q57_3_EII BIND QASP1 Q57_4_EII BIND IP QASP_1_EII ADDR=192.222.20.1 MASK=255.255.255.0 LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_II NAME=QASP_2_EII LOAD Q57.LAN SLOT=10016 FRAME=ETHERNET_ii 100hd NAME=Q57_1_Eii LOAD Q57.LAN SLOT=10017 FRAME=Ethernet_ii 100hd NAME=Q57_2_Eii BIND QASP2 Q57_1_EII BIND QASP2 Q57_2_EII BIND IP QASP_2_EII ADDR=192.222.21.1 MASK=255.255.255.O NOTE: When binding Q57_ 1_EII to QASP_ 1_EII, Q57_1_ 82 is also bo und to QASP7_1_ E82. NOTE: The VLSOT range is from 1 to 4 (only four teams can be configured). When binding, add the VSLOT number to the QASP driver. ⢠Example 4: Multiple SLB Teams with Multiple Frame Types
Adapter configurations 116 LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_II NAME=QASP_1_EII LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_802.2 NAME=QASP_1_E82 LOAD Q57.LAN SLOT=1 speed=100fd FRAME=Ethernet_II NAME=Q57_1_EII LOAD Q57.LAN SLOT=2 speed=100fd FRAME=Ethernet_II NAME=Q57_2_EII LOAD Q57.LAN SLOT=1 speed=100fd FRAME=Ethernet_802.2 NAME=Q57_1_E82 LOAD Q57.LAN SLOT=2 speed=100fd FRAME=Ethernet_802.2 NAME=Q57_2_E82 bind qasp1 q57_1_EII bind qasp1 q57_2_EII BIND IP QASP_1_EII ADDR=192.222.20.1 MASK=255.255.255.0 BIND IPX QASP_1_E82 NET=100 LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_II NAME=QASP_2_EII LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_802.2 NAME=QASP_2_E82 LOAD Q57.LAN SLOT=3 FRAME=Ethernet_II NAME=Q57_3_EII LOAD Q57.LAN SLOT=4 FRAME=Ethernet_II NAME=Q57_4_EII LOAD Q57.LAN SLOT=3 FRAME=Ethernet_802.2 NAME=Q57_3_E82 LOAD Q57.LAN SLOT=4 FRAME=Ethernet_802.2 NAME=Q57_4_E82 bind qasp2 q57_3_EII bind qasp2 q57_4_EII BIND IP QASP_2_EII ADDR=192.222.21.1 MASK=255.255.255.0 BIND IPX QASP_2_E82 NET=100 ⢠Example 5: Multiple SLB Teams with Maximum VSLOTS LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_II NAME=QASP_1_EII LOAD Q57.LAN SLOT=1 speed=100fd FRAME=Ethernet_II NAME=Q57_1_EII LOAD Q57.LAN SLOT=2 speed=100fd FRAME=Ethernet_II NAME=Q57_2_EII bind qasp1 q57_1_eii bind qasp1 q57_2_eii BIND IP QASP_1_EII ADDR=192.222.20.1 MASK=255.255.255.0 LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_II NAME=QASP_2_EII LOAD Q57.LAN SLOT=3 speed=100fd FRAME=Ethernet_II NAME=Q57_3_EII LOAD Q57.LAN SLOT=4 speed=100fd FRAME=Ethernet_II NAME=Q57_4_EII bind qasp2 q57_3_eii bind qasp2 q57_4_eii BIND IP QASP_2_EII ADDR=192.222.21.1 MASK=255.255.255.0 LOAD QASP.LAN VSLOT=3 FRAME=Ethernet_II NAME=QASP_3_EII LOAD Q57.LAN SLOT=5 speed=100fd FRAME=Ethernet_II NAME=Q57_5_EII LOAD Q57.LAN SLOT=6 speed=100fd FRAME=Ethernet_II NAME=Q57_6_EII bind qasp3 q57_5_eii bind qasp3 q57_6_eii BIND IP QASP_3_EII ADDR=192.222.22.1 MASK=255.255.255.0
Adapter configurations 117 LOAD QASP.LAN VSLOT=4 FRAME=Ethernet_II NAME=QASP_4_EII LOAD Q57.LAN SLOT=7 speed=100fd FRAME=Ethernet_II NAME=Q57_7_EII LOAD Q57.LAN SLOT=8 speed=100fd FRAME=Ethernet_II NAME=Q57_8_EII bind qasp4 q57_7_eii bind qasp4 q57_8_eii BIND IP QASP_4_EII ADDR=192.222.23.1 MASK=255.255.255.0 ⢠Example 6: VLAN with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V100_EII VSLOT=1 VLAN=100 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V200_EII VSLOT=1 VLAN=200 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V300_EII VSLOT=1 VLAN=300 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_V100_EII ADDR=172.16.210.1 MASK=255.255.0.0 BIND IP QASP_1_V200_EII ADDR=172.17.220.1 MASK=255.255.0.0 BIND IP QASP_1_V300_EII ADDR=172.18.230.1 MASK=255.255.0.0 NOTE: When binding Q57_ 1_V100_EII to Q57_1_EII , Q57_1_EII is al so bound to the other VLANs on the same VSLOT. VLAN refers to VLAN ID, and a vali d VLAN ID ranges from 1 to 4094. ⢠Example 7: Trunking with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 MODE=TRUNK LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_EII ADDR=172.16.210.1 MASK=255.255.0.0 NOTE: Switch ports must be configured for Trun king (for example, FEC or GEC for Cisco switches). ⢠Example 8: GVRP with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 VLAN=100 GVRP LOAD QASP.LAN FRAME=ETHERNET_802.2 NAME=QASP_1_E82 VSLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_1_E82 SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_2_E82 SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_EII ADDR=172.16.210.1 MASK=255.255.0.0 NOTE: In this example, VLAN 100 will be advertis ed to the switch using GVRP. Notice that untagged 802.2 frames must be loaded to allow send ing and receiving GVRP frames. Switch ports must be configured with GVRP enabled. ⢠Example 9: 802.3ad with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 MODE=802.3ad LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1
Adapter configurations 118 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_EII ADDR=172.16.210.1 MASK=255.255.0.0 NOTE: 802.3ad mode requires the Ethernet sw itch to support 802.3ad with LACP (Link Aggregation Control Protocol). Configuration parameters for Q57.LAN driver Configuration parameters for Q57.L AN driver include: ⢠TxDescriptors= o This keyword is to initialize Descripto r resources on the adapter for transmits. o Recommended Min = 100 o Max = 512 o Default = 200 NOTE: Min = 1 when used with DOS Client32 and when Keyw ord P3=1 or 2. ⢠RxBuffers= o This keyword is to pre-allocate receive ECBs & Receive adapter re sources. This setting may be affected by the NetWare server maximum/ minimum packet receive buffer settings. o Recommended Min = 32 o Max = 512 o Default = 200 NOTE: Min = 1 when used with DOS Client32 and when Keyw ord P3=1 or 2. ⢠Speed= o This keyword is to force the line speed of the adapter. When this keyword is used, the switch to which the adapter cable is connected must be forc ed. If not, the adapter and switch may run at different duplex modes. It is important to have the switch and adapter set properly. The best way to ensure good connectivity is to not set the speed keyword and allow fo r automatic link setup (auto-negotiation). o Choices are: AUTO, 10FD,10HD, 100FD, 100HD o Default is Auto. (Note that 1000 FD is only auto-detec ted at this time.) ⢠Link= o This keyword is only used to allow the adapter to negotiate a specific or forced line speed with a switch that is not forced, but instead setup for auto-negotiation. It is best to allow for auto- negotiation of the card and switch by not setting this keyword or the speed keyword. Only use this keyword if the speed keyword is set to something other than AUTO. o Choices are: AUTO, FORCE o Default is FORCE because the speed keyword is usually used when a switch and the adapter speeds are both forced to a specific value. ⢠RxTicks=
Adapter configurations 119 o This keyword is to enable the use of batching receives within a speci fic time period. o Min = 0, disabled o Max = 5000000, 5 seconds o Units are in micro seconds o Default = 360 ⢠TxPacketsPer= o This keyword is to enable the use of batching transmits to a specific amount of packets. o Min = 0, disabled o Max = 100 o Default = 64 ⢠CheckSum= o This keyword is to enable or disable the transmit & receive checksum off loading feature. The checksum off loading support is only for TCP/IP packets. For that reason, the default OFF. o Choices are: ON, OFF, TX, RX o Default is ON ⢠TxFlow= o This keyword allows enabling/di sabling of TxFlow control. o Choices are: ON, OFF o Default is O FF ⢠RxFlow= o This keyword allows enabling/di sabling of RxFlow control. o Choices are: ON, OFF o Default value is OFF ⢠VLAN= o This keyword allows enabling/dis abling of VLAN tagging support. This keyword is only used when the QASP program is loaded for VLAN support on the Q57 driver. Do not set the VLAN keyword to 1 if the QASP program is not being used. o Choices are: 0 (Disabled), 1 (Enabled) o Default value is 0 ⢠Jumbo= o This keyword enables Jumbo frame support. When enabled, jumbo packets of up to 9000 bytes are supported. o Jumbo=1514-9000 o Default is jumbo packets not supported. ⢠PDriver= o This keyword allows for the driver to operate in persistent driver mode. Only use it if the adapter is placed in a Hot Plug PCI slot and only if required to swap with an exact same board. o Choices are: ON, OFF o Default value is OFF
Adapter configurations 120 NOTE: If you modify any adapter parameters, you must reboot the system for the changes to take effect. If you make changes and do no t reboot, you may experience configuration problems. NetWare Client 32 driver installation for HP adapters The NC10032.LAN driver is a high-p erformance server driver that meets the Op en Data Link Interface (ODI) v1.11 C specification. It can be used on Cl ient 32 workstat ions with NetWare v3.X servers and greater. Installing Client 32 on DOS 1. Install the driver and its support files for DOS. 2. From the root directory, run MAKEDISK .BAT to create driver diskettes. 3. Locate the Client 32 Software. NOTE: For NetWare 5, the software is located on the Novell Client Software CD. 4. For DOS installation, run INSTALL.EXE. NOTE: INSTALL.EXE is located in the follow ing CD directory: \PRODUCTS\DOSWIN32. 5. Follow the Novell prompts. 6. For TCP/IP, select TCP/IP protocol stack at the Options/Product screen. 7. At the 32-bit LAN Driver screen, scroll to the end of li sted drivers and select User Specified 32-bit driver. 8. Insert the HP driver diskette. 9. Select HP Ethernet or Fast Ethernet NIC. 10. Select needed setting and press the F10 key to continue. 11. Follow the remaining prompts. NET.CFG parameters This section describes NET.CFG parameters for the HP driver. Each statement must appear on a separate line under LINK DRIVER N100C32 and each must be indented with at least a single space or tab character. Example NET.CFG Statements DRIVER N100C32 SLOT 16 FRAME ETHERNET_802.2 NODE ADDRESS 220023456789 ⢠FORCEDUPLEX. If no duplex setting is specified, the a dapter auto-negotiates the duplex with the hub/switch. If the adapter and hub/switch are unable to auto-negotiate, HP adap ters default to half- duplex. Full-duplex allows for simultaneous transm it and receive data transfer. Full-dupl ex can be used only when the adapter is connected to a hub th at supports and is enabled both for full-duplex mode and for the specified speed.
Adapter configurations 121 o Syntax: FORCEDUPLEX=n Where n= 0âauto-negotiate 1âhalf-duplex 2âfull-duplex NOTE: To set the adapter to half- or full-duplex, yo u must set the SPEED parameter to either 10 or 100. o Default: auto-negotiation o Examples: 100 Mbps full-duplex: FORCEDUPLEX=2 SPEED=100 10 Mbps full-duplex: FORC EDUPLEX=2 SPEED=10 ⢠SPEED. If no speed setting is specified, the adapter auto-negotiates the speed with the hub/swit ch, either 10 or 100 Mbps. If the adapter and hub/sw itch are unable to auto-negotiate, the adapter defaults to 10 Mbps. o Syntax: SPEED=n (n = 10 or 100) o Default: The adapter automatically senses speed. Examples of system file settings ⢠Example NET.CFG Link Driver N100C32 NetWare DOS Requester FIRST NETWORK DRIVE F NETWARE PROTOCOL NDS BIND Protocol IPX IPX SOCKETS 40 Protocol TCPIP IP_ADDRESS 201.201.202.25 IP_ROUTER 201.201.202.1 IP_NETMASK 255.255.255.0 PATH TCP_CFG C:\NW5C32\TCP BIND N100C32 ⢠Example STARTNET.BAT for Client 32 SET NWLANGUAGE=ENGLISH C:\NW5C32\NIOS.EXE LOAD C:\NW5C32\NBI32.NLM LOAD C:\NW532\LSLC32.NLM LOAD C:\NW5C32\CMSM.NLM LOAD C:\NW5C32\ETHERTSM.NLM LOAD C:\NW5C32\N100C32.LAN FRAME=ETHERNET_II LOAD C:\NW5C322\TCPIP.NLM LOAD C:\NW5C322\TRANNTA.NLM LOAD C:\NW5C322\IPX.NLM LOAD C:\NW5C322\SPX_SKTS.NLM LOAD C:\NW5C322\CLIENT32.NLM ⢠Example CONFIG.SYS DEVICE=C:\DOS\HIMEM.SYS device=c:\dos\emm386.exe noems DOS=HIGH,umb
Adapter configurations 122 FILES=100 buffers=25 LASTDRIVE=m fcbs=4,4 device=c:\6150\amsida.sys device=c:\6150\amsicd.sys /d:cpqcd001 ⢠Example AUTOEXEC.BAT PROMPT $p$g @ECHO OFF C:\dos\mscdex.exe /d:cpqcd001 @CALL C:\nw5c32\STARTNET Solaris To customize the driver edit "/kernel/drv/bcme.conf" and update the respective p arameters in this file. The parameters include the following. ⢠ForceSpeedDuplex. Configures link (or instance) to a ce rtain Speed and Duplex. By default, AutoNegotiate (0) is set. The set up is based on the following values: 0âAutoNegotiate 1â10 Mbps speed and Half-Duplex mode 2â10 Mbps speed and Full-Duplex mode 3â100 Mbps speed and Half-Duplex mode 4â100 Mbps speed and Full-Duplex mode 5âForce 1000 Mbps Full-Duplex mod e (Fiber adapter only) 6âAutoNegotiate only 1000 Mbps F ull-Duplex mode 7âAutoNegotiate only 1000 Mbps Half-Duplex mod e 8âAutoNegotiate only 100 Mbps Ful l-Duplex mode 9âAutoNegotiate only 100 Mbps Half-D uplex mode 10âAutoNegotiate only 10 Mbps Full-Duplex mode 11âAutoNegotiate only 10 Mbps Half-Duplex mod e Example: Configure adapters of instance#0 and in stance#3 to 100 Mbps Full-Duplex, and 10 Mbps Half-Duplex. ForceSpeedDuplex=2,0,0,1,0, 0,0,0,0,0,0,0,0,0,0,0; ForceSpeedDuplex=0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0; ⢠FlowControl. Configures the flow control pa rameters of a link. The setu p is based on the following values: 0âBoth Tx and Rx flow control are disabled. 1âTx flow control is enabled. Pause frames will be sent if resource is low, but the device will not process Rx Pause Frame. 2âRx flow control is enabled. If the device receives Pause Frame, it will stop send ing, but the device will not send Pause Frame if resource is low. 3âBoth Rx and TX flow control are enabled. Pause fr ames will be sent if resourc e is low. If the device receives Pause Frame, it will stop sending.
Adapter configurations 123 4âAdvertise both Rx and TX flow control being en able and negotiate with link partner. If link AutoNegotiate is not enabled, then both Tx & Rx Flow Contro l are disabled. Example: FlowControl=0,0,0, 0,0,0,0,0,0,0, 0,0,0,0,0,0; ⢠MaxJumboFrameSize. Configure s the Jumbo Frame feature of a link. The valid range for this parameter is 0 to 9000. If the value configured is less than 1500, then the Jumbo Frame feat ure is disabled. BCM5705 adapters do not support jumbo fr ame and this parameter will be ignored by the driver Example: MaxJumboFrameSize=0,0, 0,0,0,0,0,0,0,0, 0,0,0,0,0,0; ⢠TxPacketDescCnt. Configures the number of Tx packet de scriptors. The valid range is 32 to 512. More system memory reso urces are used for larger number of Tx Pack et Descriptors. The default value is 200. Example: TxPacketDescCnt=200; ⢠RxStdDescCnt. Configures the number of Rx packet descriptors. The valid range for this parameter is 32 to 512. More system memory resourc es are used for larger number of Rx Packet Descriptors. The default value is 500. Example: RxStdDescCnt=500; ⢠RxJumboDescCnt. Configures the number of Rx Jumbo packe t descriptors. The valid value is 32 to 256. More system memory resourc es are used for la rger number of Rx Jumbo packet descriptors. This parameter is only used if the jumbo frame feature is enabled. The default value is 50. Example: RxJumboDescCnt=50; ⢠RxCoalescingTicks. Configures the number of Rx Host Co alescing Ticks in microseconds. This determines the upper-boun dary of the time interval th at the device will generate interrupts if one or more frames are received. The default va lue is 150. Example: RxCoalescingTicks=150; ⢠RxMaxCoalescedFrames. Configures the number of Rx Maximum Coalesced Fram es parameters. This determines the upper-boundary of maximum number of Rx buffer descriptors that the device processes before it will generate an interrupt. The default value is 10. Example: RxMaxCoalescedFrames=10; ⢠TxCoalescingTicks. Configures the number of Tx Host Co alescing Ticks in microseconds. This determines the upper-boun dary of the time interval th at the device will generate interrupts if one or more frames are sent. The default value is 500. Example: TxCoalescingTicks=500; ⢠TxMaxCoalescedFrames. Configures the number of Tx Maximu m Coalesced Frames parameters. This determines the upper-boundary of maximum number of Tx buffer descriptors that the device processes before it will generate an interrupt. The default value is 80. Example: TxMaxCoalescedFrames=80; ⢠RxCoalescingTicksDuringInt. Configures the number o f Rx Host Coalescing Ticks in microseconds during interrupt. This determines the upper-boundary of the time interval that the device will gen erate interrupts if one or more fr ames are received du ring interrupt handling. The default value is 75. Example: RxCoalesci ngTicksDuringInt=75; ⢠TxCoalescingTicksDuringInt. Configures the number of Tx Host Coalescing Ticks in microseconds during interrupt. This determines the upper-boundary of the time interval that the device will gen erate interrupts if one or more fr ames are received du ring interrupt handling. The default value is 75.
Adapter configurations 124 Example: TxCoalesci ngTicksDuringInt=75; ⢠RxMaxCoalescedFramesDuringInt. Configur es the number of Rx Maximum Coale sced Frames parameters during interrupt handling. This determin es the upper-boundary of maximum number of Rx buffer descriptors that the device processes before it will generate an interrupt during interrupt handling. The default value is 10. Example: RxMaxCoalescedFramesDuringInt=10; ⢠TxMaxCoalescedFramesDuringInt. Configures the n umber of Tx Maximum Coalesced Frames parameters during interrupt handling. This determin es the upper-boundary of maximum number of Tx buffer descriptors that device proc esses before it will generate an interrupt during interrupt handling. The default value is 10. Example: TxMaxCoalescedFramesDuringInt=10; ⢠StatsCoalescingTicks. Configures how often adapter statis tics are DMAed to host memory in microseconds. Default is 1000000. Example: StatsCoalescingTicks=1000000; ⢠DoubleCopyTxBufferSize. Configures double copy Tx buffer size. If a packet to be transmitted is less than this parameter and spanned more than one fragment, these fragments of this packet will be combined into one fragment. Example: DoubleCopyTxBufferSize=64; Customize driver configuration via ndd command Driver configurations can also be temporarily changed with the Solaris ndd command. Any changes made with the ndd command are temporary and will be lost when you reboot the system. To make configuration changes survive after reboo t, modify bcme.conf instead. ⢠To display parameters that are configurable via ndd: ndd /dev/bcme '?' The system should ret urn the following: ? (read only) Instance (read and write) ForceSpeedDuplex (read and write) FlowControl (read and write) TxPacketDescCnt (read and write) RxStdDescCnt (read and write) RxCoalescingTicks (read and write) RxMaxCoalescedFrames (read and write) TxCoalescingTicks (read and write) TxMaxCoalescedFrames (read and write) RxCoalescingTicksDuringInt (read and write) RxMaxCoalescedFramesDuringInt (read and write) TxCoalescingTicksDuringInt (read and write) TxMaxCoalescedFramesDuringInt (read and write) StatsCoalescingTicks (read and write) DoubleCopyTxBufferSize (read and write) DoubleCopyTxBufferSize (read and write) BlinkLeds (write only) ⢠To configure a particular adapter, the Instance paramete r must be set to the instance associated with a particular adapter. For example, to force adapte r of instance 1 to 100 Mbps Full-Duplex: ndd -set /dev/bcme Instance 1 ndd -set /dev/bcme ForceSpeedDuplex 3
Adapter configurations 125 ⢠To query current configuration of Flow Control of instance 3: ndd -set /dev/bcme Instance 3 ndd -get /dev/bcme FlowControl ⢠To blink all LEDs for 10 seconds of adapter of instance 5: ndd -set /dev/bcme Instance 5 ndd -set /dev/bcme BlinkLeds 10
Team properties 126 T eam pr oper tie s In this section Adapter te aming ................................................................................................................ .................. 126 Multifunction gigabit se rver adapter teaming ................................................................................... ........ 129 Windows 2000 and Wind ows NT4 te aming .......................................................................................... 1 30 NetWare te aming ................................................................................................................ ................ 139 Adapter teaming ⢠Adapters must all be on the same Laye r 2 (Broadcast) network to be teamed. ⢠The maximum configuration is eight teams with ei ght adapters in ea ch team for a maximum of 64 ports. ⢠For information about teaming Multifunction Gigabit Server adapters, refer to Multifunction gigabit server adapter teaming (on page 129 ) in this guide. Teaming types for HP Fast Ethern et and Gigabit Server Adapters NOTE: Dual Channel Network Fault Tolerance te aming capabilities are not supported on systems running Windows® 2000. ⢠Automatic (Recommended). Default setting. The system automati cally selects the best team type o r combination of team features based on the config uration of the adapters in the team. 802.3ad will be used to place team members into port trunks/channels if possible. All transmit packets are load balanced among all team members based on a Lo ad Balancing algorithm in the teaming device driver. The receive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are lo ad balanced among the remaining adapters. Currently, Automatic is supported in Window s® 2000, Windows Server 2 003, and Windows Server 2003 x64. ⢠802.3ad Dynamic Dual Channel Load Balancing (INP). ProLiant Essentials Intelligent Ne tworking Pack (INP) only. There must be a minimum of two adapters in a dual channel team. All team members within each group are dynamically placed into a port- trunk/channel by dynamic Link Aggregation Contro l Protocol (LACP) agreement with the switch. A failure of a team member to synchronize joining a po rt-trunk or channel with the switch is treated as an error and that team member is considered fail ed. All transmit packets are load balanced among all team members based on a Load Balancing algo rithm in the teaming device driver. The receive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are load balanced among the remaining adapters. ⢠802.3ad Dynamic Dual Channel Network Fault Tolerance (INP). Available in the ProLiant Essentials INP only. Th ere must be a minimum of two adapters in a dual channel team and a minimum of one adapter per gr oup. All team members within each group are dynamically placed into a port-trunk/channel by dy namic LACP agreement with the switch. A failure of a team member to synchronize joining a port-tru nk or channel with the switch is treated as an error and that team member is considered fail ed. Dynamic Dual Channel Network Fault Tolerance
Team properties 127 (NFT) provides the safety of additional backup li nks between the server and hub/switch. Dynamic Dual Channel NFT is implemented with one channe l team providing a primary adapter and an other channel team providing a secondary, backup adapte r(s). During normal operations, if the adapters in the primary team fail, a link to one of the ad apters in the secondary, backup team automatically takes over. ⢠802.3ad Dynamic Dual Channel NFT and Preference Order (INP). Available in the ProLiant Essentials INP only. Pr ovides the same optio ns as 802.3ad Dynamic Dual Channel NFT except it offers the additional option of setting the priority order of the adapters that are used in the primary and secondary channels. ⢠Switch-assisted Dual Channel Load Balancing (INP). Available in the ProLiant Essentials INP only. There must be a minimum of two adapters in a dual channel team and a minimum of one adapter per group. One group of team members is treated as a Switch-assisted Load Balancing (SLB) group to one switch, and a second group of team members is treated as an SLB gr oup to a second switch. All transmit packets are load balanced among all team members ba sed on a Load Balancing algorithm in the teaming device driver. The rece ive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are load balanced among the remaining adapters. Only Gigabit adap ters can be used for Switch-assisted Dual Channel Load Balancing. When you select Switch-a ssisted Dual Channel Load Balancing, the Team Members window bec omes Team Members Grouping and displays two windows l abeled Group 0 and Group 1. The up and down arrows are used to assign the members to be in Group 0 and Group 1. If a team contains an iSCSI-enabled adapter, you cannot use the Switch-assisted Dual Channel Load Balancing (INP) setting. ⢠Switch-assisted Dual Channel Network Fault Tolerance (INP). Available in the ProLiant Essentials I NP only. There must be a minimum of two adapters in a dual channel team and a minimum of one adapter per group. Switch-assisted dual channel Network Fault Tolerance (NFT ) provides the safety of additional backup links betwee n the server and hub/switch. It is implemented with one channel team providing the primary adapter and a second channel team providing the secondary, backup adapter(s). During normal operations , if the adapters in the primary team fail, a link to an adapter on the Secondary team automatically takes over . Team members can be moved to Group 0 or Group 1 on the Team Properties Teaming Controls tab. ⢠Switch-assisted Dual Channel NFT and Preference Order (INP) . Available in the Pr oLiant Essentials INP only. Provides the same options as Switch -assisted Dual Channel Network Fault Tolerance except it offers the additional opti on to select the priority order of the group (Group 0 or Group 1) used for the primary and secondary channels. ⢠802.3ad Dynamic with Fault Tolerance. All team mem bers are dynamically placed into a port- trunk/channel by dynamic protocol agreement (LACP) with the switch. A failure of a team member to synchronize joining a port-trunk or channel with th e switch is treated as an error and that team member is considered failed. All transmit pa ckets are load balanced among all team members based on a Load Balancing algorithm in the teamin g device driver. The receive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are load balanced among the remaining adap ters. If a team contai ns an iSCSI-enabled adapter, you cannot use the 802.3ad Dyna mic with Fault Tolerance setting. ⢠Switch-assisted Load Balancing with Fault Tolerance (SLB). SLB increases a server's throughput using Fast Ethernet or Gigabit Ethernet ports. SLB lo ad balances all traffic. Unlike TLB, SLB can be configured to increase both transmitting and rece iving channels between your serv er and switch. For example, an SLB team containing four HP Fast Ethernet adapters configured for full-duplex operation would provide an aggregate maximum transmit rate of 400 Mbps a nd an aggregate maximum receive rate of 400 Mbps, resulting in a total bandwidth of 800 Mb ps.
Team properties 128 SLB works only with SLB-en abled switches, such as the Catalyst 600 0 series. The SLB software continuously analyzes loading on each adapter an d balances network traffic across the adapters as needed. Adapter teams configured f or SLB also provide the benefits of NFT. To use SLB, you must have at least two HP Fast Et hernet or Gigabit adapters installed in your server and linked to the same SLB-enabled switch. For SLB, you can only team 100 Mbps adapters together (no 10 Mbps or Gigabit), with the exception of the NC71xx and NC77xx server adapters, which are Gigabit and can run at 10/100 Mbps. ⢠Transmit Load Balancing with Fault Tolerance (TLB). TLB creates a team of adapters to increase a server's throughput. If you configure TLB, NFT is automatically configured. TLB wor ks with any industry-standard 10BASE-T, 100BASE-TX, or Gigabi t switch. All transmit IP frames are load balanced among all team members based on a Load Balancing algorithm in the teaming device driver. The Current Primary adapter transmits all non- IP frames, and receives all frames for the team. If a failover event occurs, one of the non-Primar y adapters assumes the role of Current Primary adapter, and transmit IP packets are load balanced among all remaining team members. If a failure occurs in any of the non-Primary adapters, the pa ckets are load balanced among all remaining team members. To use TLB, you must have at least two NCxxxx a dapters (ports) capable of the same speed and linked to the same Layer 2 (Broadcast) sw itch insta lled in your server. They cannot be connected to the same switch and/or be members of different VLANs. With TLB, as you add adapters to your server, you can group them in teams to provide a si ngle virtual adapter with increased transmission bandwidth. For example, a TLB team containing four HP Fast Ethernet adapters configured for full- duplex operation would provide an aggregate maximum transmit rate of 400 Mb ps and a 100 Mbps receive rate resulting in a total bandwidth of 500 Mbps. TLB load balances IP traffic under Window s NT® and IP and IPX traffic under NetWare. ⢠Transmit Load Balancing with Fault Tolerance and Preference Order. All transmitted IP packets are load balanced among all team members based on a Load Balancing algorithm in the teaming device driver. The Current Primary adapter transmit s all non-IP packets, and receives all frames for the team. If a failover event occurs, one of the non-Primary adapters assumes the role of Current Primary adapter, and transmitted IP packets are load balanced among all remaining team members. If a failure occurs in any of the non-Primary a dapters, the packets are load balanced among all remaining team members. The user s ets the priority of the adapters relative to each other within the team. The priority of the adapters is an additi onal consideration when determining the Primary adapter. ⢠Network Fault Tole rance Only (NFT). NFT provides the safety of an additional backup link between the server and hub/switch. NFT is implemented wi th a Primary adapter and a Secondary (backup) adapter(s). During normal operations, if the Primary adapter fails, the link to the Secondary adapter automatically takes over. o NFT can occur between two TOE-enabled adapters ⢠Network Fault Tolerance only with Preference Order. NFT provides the safety of an additional backup link between the server and hub/switch. NF T is implemented with a Primary adapter and a Secondary (backup) adapter(s). During normal operatio ns, if the Primary adapter fails, the link to the Secondary adapter automatically takes over. The user sets the priority of the adapters relative to each other within the team. The priority of the adapters is an additional consideration when determining the Primary adapter. Currently, NFT is supported in Windows® 200 0, Windows Server 2003, and Windows Server 2003 x64, Windows NT® 4.0, NetWare 6.x, and UnixWare 7.x. Transmit Balancing Method
Team properties 129 Select or change the Transmit Balancing Method for th e team. If the team type is Network Fault Tolerance Only with Preference Order or Network Fault Toleranc e Only (NFT), the Transmit Balancing Algorithm list disabled and Automatic is selected. o Automatic . (Recommended) Default settin g. Teaming driver selects the load balancing mechanism based on the packet type. o TCP Connection . Load balances transmit TCP packets using the TCP connection. o Destination IP Address . Load balances transmit IP packets using the last four bits of the destination IP Address. o Destination MAC Address . Load balances transmit IP packets using the last four bits of the destination MAC Address. o Round Robin (Packet order not guaranteed) . Load balances transmit IP packets among all team members. A packet is sent on one team member, the next packet is sent out on the next team member, and so on. When the last team member is utilized, the rota tion begins again. UnixWare 7.x and OpenServer 6.x teami ng notes UnixWare 7.x and OpenServer 6.x operating systems support Network Fault Tolerance (NFT) but not Transmit Load Balancing (TLB) or Switch-A ssisted Load Balancing (SLB). Multifunction gigabit server adapter teaming If a TOE- or iSCSI-enabled adapter is included as one of the following team types, the team type is changed to TransmitLoadBalancing. ⢠Automatic ⢠802.3ad Dynamic Dual Channel Load Balancing (INP) ⢠Switch-assisted Dual Channel Load Balancing (INP) ⢠802.2ad Dynamic with Fault Tolerance ⢠Switch-assisted Load Balancing with Fault Tolerance (SLB) The following are limitations for a TOE-en abled team: ⢠TOE teaming performance is reduced if a switch or network device is configured with the round robin aggregation load balancing method. Configure the aggregation load balancing method on the switch or network device to somethin g else ot her than round robin and then re-start the team (disable the team and enable the team for all VLANs). ⢠The 802.3ad Dynamic Dual Channel Load Balancing (INP) and Switch-assisted Dual Channel Load Balancing (INP) team types are not supported fo r TOE-enabled adapters. Select a TOE-supported team type and apply the change, if necessary, before enabling TOE. It is necessary to apply the new team type, if you are unable to enable TOE. When creating a team with all TOE-capable adapters: ⢠If at least one adapter is TOE enabled, the team is created with TOE enabled and all adapters on the team are set to TOE enabled. ⢠If none of the adapters are TOE enabled, th e team is created with TOE disabled. ⢠If TOE is disabled on a team of TOE-capable adapters, enabling TO E for the team changes all of the adapters to TOE enabled. When creating a team with TOE-capable and non-TOE capable adapters:
Team properties 130 ⢠If the new team includes at least one TOE-en abled adapter and at least one non-T OE capable adapter, the team is created with TOE disabled and TO E is disabled on all of the individual teamed adapters. ⢠If a non-TOE capable adapter is added to a TO E-enabled team, the team is changed to TOE disabled. TOE functionality for ea ch teamed adapter is disabled. Disabling or dissolving a TOE-enabled team: ⢠If TOE teaming is disabled on an existing TOE-en abled team, TOE is disabled on all of the tea med adapters. ⢠If a team is dissolved, the TOE functionality for ea ch adapter that was on the team does not change. RSS teaming characteristics When creating a team with all RSS-capable adapters: ⢠If at least one adapter is RSS-enabled, the team is created with RSS-enabled and all adapters on the team are set to RSS-enabled. ⢠If none of the adapters are RSS-enabled, th e team is created with RSS-disabled. ⢠If RSS is disabled on a team of RSS-capable adapters, enabling RSS for the team changes all of the adapters to RSS-enabled. When creating a team with RSS-ca pable and non-RSS capable adapters: ⢠If the new team includes at least one RSS- en abled adapter and at least one non-RSS capable adapter, the team is created with R SS-disabled and RSS functionality fo r each adapter that is in the team is disabled. ⢠If a non-RSS capable adapter is added to an R SS-enabled team, the team is changed to RSS- disabled and RSS functionality for each a dapter that is in the team is disabled. Disabling or dissolving an RSS-enabled team: ⢠If RSS teaming is disabled on an existing RSS-enabled team, the RSS functionality for each adapter that is in the team is disabled. ⢠If a team is dissolved, the RSS fu nctionality for each adapter that was in the team does not change. Windows 2000 and Windows NT4 teaming Windows 2000 teaming NOTE: The installation of adapter drivers from ot her vendors is not supported by HP. To avoid unpredictable results when using the HP Teaming function or other features supported by the HP Network Configuration Utility, install only HP adapter driver s. Upgrading teams to Windows® 2000 The Windows® 2000 Primer Utility from HP lets you re move unsupported Windo ws® 2000 utilities before or after upgrad ing your OS. These uns upporte d utilities w ere designed to run on Windows NT® 4.0 and may have been installe d using the following media: ⢠Any version of the HP SSD for Windows NT® Setup program ⢠Any version of HP Integration Maintenance Utility and HP System Partition Upgrad e Utility installed from HP SmartStart
Team properties 131 ⢠Any HP Software and Documentation CD Version 5.0 or earlier shipped with an HP 10/100 or Gigabit adapter The HP Primer Utility is supported on the following oper ating systems: ⢠Microsoft® Windows NT® 4.0 ⢠Microsoft® Windows® 2000 For PRIMER.EXE, refer to the HP Softpaq #10652 at the HP support website ( http://h18000.www1.hp.com/s upport/files/index.html ) Before running HP Primer Utility Before running the HP Primer Utility, you should be aware of the following: ⢠Retaining Teaming Config uration Settings in Windows NT® 4.0 HP recommends that you record all Windows NT® 4. 0 teaming configuration s ettings. After you run Primer and upgrade the OS to Windows® 2000, yo u can install the new HP Network Configuration Utility for Windows® 2000 and reconfig ure th e adapter teams using the saved settings. ⢠Windows NT® 4.0 components removed by Primer The Primer utility removes the install ed files, th e program links, and the registry entries for the previously installed components. In addition, general cleanup is pe rformed to remove program l inks and common files that may not belong to any specific component. ⢠Adapter configur ation components remove d by Primer o HP Advanced Network Control Utility o HP Network Configuration Utility o HP TLAN Network Teaming and Configuration Utility When Primer removes these network uti lities, it dissolves all Windows NT® 4.0 teams. It also removes the teaming drivers that support the utility. As a result, the adapters that were managed by this utility are "unteamed" and function as single, stand-alone adapters. ⢠Other components also removed by Primer o HP Array Configuration Utility o HP Cluster Verification Utility o HP Insight Management Foundation Agents o HP Insight Management Server Agents o HP Insight Management Storage Agents o HP Integrated Management Display Utility o HP Integrated Management Log Viewer Utility o HP Integration Maintenance Utility o HP NT Management o HP Online Storage Controller Recovery Utility o HP PCI Hot Plug Utility o HP Power Down Manager o HP Power Supply Viewer o HP Remote Monitor Service o HP System Management Services
Team properties 132 o HP System Partition Upgrade Utility Using HP Primer Utility from the graphical user interface To remove the Windows NT® components either befo r e or after upgrading to Windows® 2000, follow these steps: 1. From the location where you have downloaded Primer, double-click PRIMER.EXE. 2. Click CLEANUP. The utility will now remove the previously-in stalled Windows NT® components. 3. Exit the Windows® 2000 Primer Util ity from HP. 4. Reboot your system for the changes to take effect. Using HP Primer Utility from the command line You may run this utility in silent mode to perform th e cleanup without us ing the graphical user interface by following these step s: 1. Open the MS-DOS command prompt. 2. Go to the directory where the Windows® 200 0 Prim er Utility from HP, PRIMER.EXE, is located. 3. Type primer s and press Enter. The utility will run in silent mode. Viewing the HP Primer Utility log results After clicking the Cleanup button or performing a command-line in stall using "primer s," a log file is generated that allows you to view the uninstalled comp onents and any non-critical errors received during cleanup, such as components that the utility was unable to uninstall. You can easily view the results by clicking the View Log button or you can go to the directory where the log file resides, which is usually th e current working directory (where the utility is run). However, if the media from which you are running the utility is writ e-protected, the log file will reside in the %systemdir% directory. NOTE: You must have Administrator privileg es to use the Log. Teaming considerations when upgrading to Windows® 2000 Under Windows NT® 4.0, adapter teaming ma y be managed using CPQNTAC (HP Network Configuration Utility) for HP NCxxxx Ethernet , Fast Ethernet, and Gigabit adapters Because different utilities m anage teaming under Wi ndows NT® 4.0, in upgrading to Windows® 2000, certain steps are necessary to maintain teaming for redundant adapter pairs. Removing previous forms of teaming on Windows NT® 4.0 Both CPQNTAC.SYS and CPQTTAC.SYS are NDIS4 Inte rmediate drivers that do not work with Windows® 2000. For a Windows® 200 0 upgrade, yo u must remove them with one of the following methods: Method 1: Use HP Primer Utility for Windows® 2000 1. Run the HP Primer Utility. Primer remov es the installed files, program links, registry e ntries, and generally cleans files that may not belong to an y specific component. For further information about Primer, refer to HP Softpaq #10652 provided at HP support website ( http://h18000.www1.hp.com/s upport/files/index.html ). 2. Upgrade the OS to Windows® 2000. 3. Install the new HP Network Configuration Utility (NCU) for Windo ws® 2000.
Team properties 133 Method 2: Manual Remova l 1. Dissolve all teams and controller-pairs wit h their respective utilities. 2. Upgrade the OS to Windows® 2000. 3. Install the new HP Network Configuration Utility (NCU) for Windo ws® 2000. Installing the teaming driver NOTE: Before installing the Teaming protocol (T eaming driver), you may want to back up the Windows® 2000 Registry. To install the teaming driver: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. 8. Select the win2000, win2003 or win2003x64 folder that is appropriate for your operating system. 9. Run the HP ProLiant Network Configuration Ut ility Component Package (cpxxxxxx.exe file). The component is a self-extracting executable with a file name based on the componen t number. 10. Click the Install button to proceed with the installation. 11. After the drivers have been installed, you may delete the HP SoftPaq file. If this is a first-time installation, proceed to step 1 in the "Configuring (or Adding) a Team" section. Additional documentation can be found in Soft Paq SP24889 available at the HP support web site ( http://h18000.www1.hp.com/s upport/files/index.html ). Configuring (or adding) a team NOTE: You can configure a maximum of eight teams with eight adapters in each team for a maximum of 64 ports. NOTE: If you are adding adapters to an existing team, you do not need to install the Teaming driver. To configure a team or add an adapter to an existing team: 1. Launch the HP Network Configuration Utility by doing one of the following: o Double-click the HP Network tray icon. o Double-click the HP Network icon in the Control Panel window. 2. Select the adapters to be teamed in the NCU Main window. 3. Click Team . 4. Configure the team as needed and click OK . 5. Click Yes to confirm the configuration changes.
Team properties 134 6. A message box appears with a warning that the protocol information for the primary adapter, such as the IP address must be configured properly. Click OK to exit this message box. 7. Restart the HP NCU and verify the adapter protocol s were configured properly. Select the adapter and click Properties . Uninstalling teamed adapters If you need to completely uninstall a teamed adapter (n ot simply replace it), perform these steps to ensure proper operations: 1. Select the Team and click Dissolve . 2. Click OK on the NCU Main window to accept the changes. 3. Physically remove the adapter from the server. Configuring HP adapters The Properties dialog for the HP Network Team Loca l Area Connection contains the following tabs and buttons. General Tab ⢠Connect usingâ HP Network Team. Components that are selected are used by this connection. These include: o Client for Microsoft® Networks o HP Network Configuration Utility o Network Load Balancing o File and Printer Sharing for Microsoft® Networks o Network Monitor Driver o Net BEUI Protocol o Internet Protocol (TCP/IP) NOTE: Other protocols may apply. ⢠Descriptionâ Describes the selected c omponent ⢠Show icon in taskbar when connectedâ Enables/Disables the HP connection tray icon Buttons ⢠Configure âDisplays Windows® 2000 device properties o General (Troubleshooter)âDisable Device option o Driver (Name, Version number, Driver D etails; Install/Uninstall/Update driver buttons) ⢠Install âInstalls the selected protocol ⢠Uninstall âUninstalls the selected protocol ⢠Properties âAccesses dialogs on the protocol for the selected device. Sharing Tab Sharing allows other servers or your local network to access external resource s through this connection. The options on the Sharing tab are:
Team properties 135 ⢠Shared Access (Enable) âIf selected, it allows Internet connection sharing. You can share Applications and Protocols. ⢠Applications tab âDisplays network applications to be enabl ed for servers sharing this connection. ⢠Services tab âDisplays services to be provid ed to the remote network. HP NCxxxx Local Area Connection The HP NCxxxx Local Area Connection window contains the following tabs and buttons: General Tab The options on the General tab are as follows: ⢠Connect using âName of the adapter or module connection. Components that are checked are used by this connection. These include: o Client for Microsoft® Networks o HP Network Configuration Utility o Network Load Balancing o File and Printer Sharing for Microsoft® Networks o Network Monitor Driver o Net BEUI Protocol o Internet Protocol (TCP/IP) NOTE: Other protocols may apply. ⢠Descriptio nâDescribes the select ed component ⢠Show icon in taskbar when connected âEnables/Disables the HP tray icon Buttons ⢠Configure buttonâ Windows® 2000 device properties device. o General Tab Device status Troubleshooter Device usage o Advanced Tab The following options may vary with different adapters. 802.1p QOS (Enable/Disable) Checksum Offload Ethernet@WireSpeed Flow Control Media Connector (AUI/BNC or UTP) Locally Administered Address Number of Receive Descriptors Number of Transmit Descriptors Speed/Duplex
Team properties 136 WakeUp Capabilities WOL Speed ⢠Driver Tab o Name, Version number, Driver Details o Install o Uninstall o Update driver ⢠Resources Tab o Resource settings o Conflicting device list ⢠Install button âInstalls the selected protocol ⢠Uninstall button âUninstalls the selected protocol ⢠Properties button âAccesses dialogs on the protoc ol for the selected device. Sharing Tab Sharing allows other servers or your local network to access external resource s through this connection. The options on the Sharing tab are: ⢠Shared Access (Enable) âIf sele cted it allo ws Inte rnet conne ct ion sharing. You can share applications and protocols. o Applications âNetwork applications to be enabled for servers sharing this connection o Services âServices to be provided to the remot e network Windows NT Teaming The Fast Ethernet and Gigabit Ethernet adapters prov ide the followi ng options for increasing throughput and fault tolerance when running Windows NT® 4.0. ⢠Network Fault Tole rance (NFT). NFT provides the safety of an additional backup li nk between the server and hub/switch. NFT is implemented with a Primary adapter and a Secondary (backup) adapter(s). During normal operations, if the Primary adapter fails, the link to the Secondary adapter automatically takes over. ⢠Transmit Load Balancing (TLB). TLB creates a team of adapters to increase a server's throughput. If you configure TLB, NFT is automatically configured . TLB works with any industry-standard 10BASE-T, 100BASE-TX, or Gigabit switch. All transmit IP fr ames are load balanced among all team members based on a Load Balancing algorithm in the team ing device driver. The Current Primary adapter transmits all non-IP frames, and receives all frames for the team. If a failover event occurs, one o f the non-Primary adapters assumes the role of Current Pr imary adapter, and transmit IP packets are load balanced among all remaining team members. If a fa ilure occurs in any of the non-Primary adapters, the packets are load balanced amon g all remaining team members. To use TLB, you must have at least two NCxxxx a dapters (ports) capable of the same speed and linked to the same Layer 2 (Broadcast) sw itch insta lled in your server. They cannot be connected to the same switch and/or be members of different VLANs.With TLB, as you add adapters to your server, you can group them in teams to provide a si ngle virtual adapter with increased transmission bandwidth. For example, a TLB team containing four HP Fast Ethernet adapters configured for full- duplex operation would provide an aggregate maximum transmit rate of 400 Mb ps and a 100 Mbps receive rate resulting in a total bandwidth of 500 Mbps.
Team properties 137 ⢠TLB load balances IP traffic under Windows NT® and IP and IPX traffic under NetWare. ⢠Switch-assisted Load Balancing (SLB). SLB increases a server's through put using Fast Ethernet or Gigabit Ethernet ports. SLB load balances all traffic . Unlike TLB, SLB can be configured to increase both transmitting and receiving ch annels between your server and switch. For example, an SLB team containing four HP Fast Ethern et adapters configured for full-duplex operation would provide an aggregate maximum transmit rate of 400 Mbps and an aggregate maximum receive rate of 400 Mbps, resulting in a tota l bandwidth of 800 Mbps. SLB works only with SLB-en abled switches, such as the Catalyst 500 0 series. The SLB software continuously analyzes loading on each adapter an d balances network traffic across the adapters as needed. Adapter teams configured f or SLB also provide the benefits of NFT. To use SLB, you must have at least two HP Fast Et hernet or Gigabit adapters installed in your server and linked to the same SLB-enabled switch. For SLB, you can only team 100 Mbps adapters together (no 10 Mbps nor Gigabit), with the exception of the NC71xx and NC77xx server adapters, which are Gigabit and can run at 10/100 Mbps. NOTE: Configuring Switch-assisted Load Balancing typically requi res pairs of ports (two or four ports). Refer to the documentation that sh ipped with your SLB-c apable switch for more information. After your switch is correctly set up, you can configure the adapters in the server with the HP Networking Configuration Utility. Before configuring adapter teaming Before you configure th e HP adapter for adapter teaming, you need to install the adapters and make the appropriate connections to hubs and switches. Installing adapters 1. Install at least two adapters in a Windows NT® 4.0 system as describe d in Windows NT® 4.0 Installation Notes. 2. After installation is comple te, restart Windows NT® 4.0. NOTE: Windows NT® 4.0 Service Pack 5 or late r is required for implementing adapter teaming properly. Install the service pa ck before configuring adapter teaming. Connecting to hubs and switches ⢠For Fast Ethernet adapters, if connected to a hub, each adapter in a team must be connected to a port that is in the same collision domain. ⢠For any adapter connected to a switch, each adapter in a team must be connected to a port which is on the same Layer 2 (Broadcast) network. Installing and configuring teaming NOTE: You can configure a maximum of eight teams with eight adapters in each team for a maximum of 64 ports. 1. Double-click the Network Icon in the Control Panel. This launches the Network Control Panel applet. 2. Select the Adapter tab and double-click the adapter found in the Network Adapters list box. This will launch the CPQNTAC utility. NOTE: To modify the networking configuratio n in Windows NT® 4.0 through CPQNTAC, you must be logged on to the system as an Adminstrator.
Team properties 138 3. Select two to eight of the listed adapters by single -clicking them. Individual stand-alone adapters will remain selected as you click them. After you have selected all the desired ad apters for a team, click the Team button on the upper-right of the main dialog window. IMPORTANT: If the proper Service Pack level (SP5) has not been installed on the system, a popup message displays warning you of this condit ion. If this message is displayed, install the correct Service Pack before proceeding. 4. The Teaming Property Sheet appears for the newly created team. The first page displayed is the Teaming Controls Sheet. The default mode of oper ation for a team is Faul t Tolerant mode. Change this to the desired mode of operation by selecting either Fault Tolerant or Load Balancing at the top of the page. 5. For a Fault Tolerant team, select among the following modes: Manual. Causes a failover when the user clicks Switch Now . In this mode, a failover will also o ccur if the Current Primary adapter is dropped from th e team. Recommended only for troubleshooting purposes. Fail on Fault (Default). If a failover event occurs, a non-Primary adapter takes on the role of the Current Primary adapter. This new Current Primary adapter remains primary unless another failover event occurs, or the system is rebooted. Smart Switch. If a failover event occurs, a non-Primary ad apter takes on the role of Current Prim ary adapter. This new Current Primary adapter remains primary until one of the following occurs. The Preferred Primary adapter becomes available, and a failback occurs, in which case the Preferred Primary adapter resumes its role as Current Primar y adapter. Or, another failover event occurs, in which case, a non-Primary adapter takes on the role of Current Primary adapter. If this mode is selected, make sure the desired Pr imary adapter is displayed in the Preferred Primary NIC list box. A single Fault Tolerant team CAN contain a mixture of Fast Et hernet and Gigabit adapters. 6. For a Load Balancing team, select among the following modes: o Transmit Load Balancing o Switch-assisted Load Balancing 7. Select the Settings tab at the top of the Teaming Property Sheet. On the Settings tab, you may configure the following it ems: Team Network Address. Used to "Locally Administer" the team's un ique receive node address on the network like you would locally administ er a single adapter's node a ddress. Enable Heartbeats. This check box is used to enable/d isable teaming Heartbeats. Heartbeats provide an additional level of adapter failure de tection above simple li nk failure detection. Heartbeat detection, if enabled, will cause Heartb eat packets to be transmitted between the Primary adapter and Secondary adapters to ensure proper transmit and receive functionality among the adapters in the team. Heartbeat timer interval. This value specifies th e amount of time to wait before transmitting Heartbeat packets between adapters in the team. Heartbea ts are not transmitted between the adapters if normal network transmit/receive traffic is passing through the adapters. However, in the absence of network traffic, after the specified Heartbeat time out period expires (the timer interval value), Heartbeat packets are sent. This value MUST be greater than 3000 ms (3 seconds). 8. Click OK on the Teaming Property Sheet dialog to complete the configuration of the team. 9. Repeat Steps 3â8 to create additional teams as desired. 10. Click Close at the bottom of the Main CPQNTAC dial og to finish configuration of teams.
Team properties 139 11. Click Close at the bottom of the Network Control Pane l applet to finish configuration of network devices. 12. Binding analysis will be done to bind newly installed adapte rs and teams to all installed communication protocols. You may be prompted to enter protocol specific information (s uch as IP information) for newly installed adapters and teams. 13. When prompted, click Yes to reboot Windows NT® . NetWare teaming IMPORTANT: Teaming NC150, NC32x, NC37x, NC1020, NC6770, or NC77 xx adapters with NC110x, NC310, NC31xx, NC340x , NC360x, NC364x, NC61xx, or NC71xx adapters is not supported for NetWare. The HP Advanced Network Services driver name is CPQANS.LAN. CPQANS requirements : ⢠Use with NetWare 5.1 and 6.x ⢠Always install the latest Novell Support Pa cks ⢠Always install the latest HP NSSD CPQANS teams with the following HP server adapters : ⢠HP NC110T PCI Express Gigabit Server Adapter ⢠HP NC310F PCI-X Gigabit Server Adapters ⢠HP NC340T PCI-X Quad-Port Gigabit Server Adapter s ⢠HP NC360T PCI Express Dual Port Gigabit Server Ad apter ⢠HP NC364T PCI Express Quad Port Gigabit Server Adapter ⢠HP NC61xx Gigabit Server Adapters ⢠HP NC31xx Fast Ethernet Server Adapters ⢠HP NC71xx Gigabit Server Adapters CPQANS requires the following Novell settings : ⢠SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) ⢠SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) ⢠SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) Options for increasing throughput and fault tolerance HP Gigabit, Fast Ethernet, and Ethernet server adapters teamed with CPQANS provide the following options for increasing thr oughput and fault tolerance for NetWare 5.1, 6.0, and 6.5. ⢠Network Fault Tole rance (NFT). Creates a team of adapters with the Primary adapter in the active state and the remaining adapters in a standby state. Primary specifies which adapter in the team is the Preferred Primary adapter. If the Preferred Prim ary adapter fails and is later replaced, or if a cable is reattached to fix a failure on the pr imary adapter, the secondary adapter automatically passes control back to the Preferred Primary. Works with any Ethernet switch or hub. NFT is the only team mode that allows for a team of different speed adapters. Refer to the discussion on mixed teams in the "Important teaming notes" section for more informat ion.
Team properties 140 o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=NFT NOTE: NFT is the default mode. Issuing a "LOAD CPQANS COMMIT" command without specifying a "MODE" will result in a team configured for the Network Fault Tolerance mode. ⢠Adaptive Load Balancing (ALB). Creates a team of two or more adapters to increase transmission throughput while providing adapter redundancy. HP recommends that all adapte rs in an ALB team be the same speed (either all Gigabit or all Fast Ethernet). In an ALB team, the primary adapter is used for receiving and transmitting, while all ot her team members transmit. ALB w orks with any Ethernet switch. No performance gain is made with a hub or repeater. o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=ALB ⢠Fast EtherChannel (FEC). Creates a team of HP Fast Ethernet server adapters to increase receive and transmit throughput using a switch with Fast Ethe rChannel technology. As with NFT, FEC provides redundancy. An FEC team uses the switc h FEC tec hnology to perform the load balancing rather than the server. The switch algorithm determin es which ports are used for transmits and receives. All adapters are capable of receiving and transmitting at the same time. o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=FEC NOTE: FEC teaming requires a switch with Fa st EtherChannel technology. Refer to the "Important teaming notes" section for more information about configuring your switch for FEC or GEC teams. ⢠Gigabit EtherChannel (GEC). Creates a team of HP Gigabit server adapters to increase receive and transmit throughput using a switch with Giga bit EtherChannel technolo gy. As with NFT, GEC provides redundancy. A GEC team uses the switch GEC technology to perform the load balancing rather than the server. The switch algorithm determin es which ports are used for transmits and receives. All adapters are capable of receivin g and transmitting at the same tim e. o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=GEC NOTE: GEC teaming requires a switch with Giga bit EtherChannel technology. Refer to the "Important teaming notes" section for more information about configuring your switch for FEC or GEC teams. ⢠802.3ad (Link Aggregation Control Protocol â LACP). Creates a team of HP Gigabit server adapters to increase receive and transmit throughput us ing a switch that supports 802.3ad with LACP protocol. o Syntax : LOAD CPQANS MODE=xxx o Example : LOAD CPQANS MODE=AD Important teaming notes ⢠Configuring Fast EtherChannel (FEC) or Gigabit Et herChannel (GEC) typically requires pairs of ports (two or four ports). However, review the documentation that comes with your FEC- or GEC-capable switch for more information. After yo ur switch is correctly configured, yo u can configure your GEC or FEC team in accordance with your switch. You sh ould always configure your switch first then your server for GEC or FEC tea ms.
Team properties 141 ⢠Every member of a CPQANS team must be connected to the same network segment. If you have multiple CPQANS teams, each team must be on a separate network segment. If you are connected to a VLAN-capable switch, each HP server adapter in a team must be connected to a port with the same VLAN ID number (same network segment). NOTE: VLAN is not supported on heterogeneous teaming. ⢠For mixed teams (different speed adapters), Networ k Fault Tolerance (NFT) is the only team mode that allows for a team of different speed adapters. Such a team can contain any combination of HP NC310x Gigabit Server Adapters, NC61xx Giga bit Server Adapters, NC71xx Gigabit Server Adapters, NC31xx Fast Ethernet Server Adapters an d so on. With a mixed team, you can designate the highest-performance adapter as the preferred primary adapter. To set the preferred primary adapter, use the "PRIMARY" keyword when binding CPQANS to the base driver for the Ethernet server adapter. Adapter performance hierarchy: Gigabit server ad apter (1000 Mbps) is the fastest followed by the Fast Ethernet (10/100 Mbps) server adapter. Additional information The following are some commands that can be help ful in configuring or troubl eshooting teaming . ⢠CPQANS STATUS o Description: This command displays th e current status for the CPQANS team. o Syntax: LOAD CPQANS STATUS o Example: LOAD CPQANS STATUS If you have more than one team, then you can is sue the same command with the team's ID number. o Syntax: LOAD CPQANS STATUS TEAM=<IDnumber> o Example: LOAD CPQANS STATUS TEAM=2 ⢠CPQANS PRIMARY o Description: This command cr eates a preferred primary. o Syntax: BIND CPQANS <adapter name> PRIMARY o Example: BIND CPQANS N100_1_E82 PRIMARY In this example, the adapter "N100_1_E82" is the preferred primary for the CPQANS team. PRIMARY specifies which adapter in the team is the Preferred Primary adapter. If the Preferred Primary adapter fails and is later replaced, or if a cable is reattached to fix a failure on the primary adapter, the secondary adapter automa tically passes control back to the Preferred Primary. For NFT mode teams, if you have a server with an HP Gigabit Server Adapter and an HP Fast Ethernet Server Adapter, you would want the Gigabit Server Adapter to be the Preferred Primary. In this scenario, if the Gigabit Server Adapter fails, the Fast Ethernet S erver Adapter will take over. Then, when the Gigabit Server Adapter is replaced, it will automatically revert to being the Primary Adapter in the team. NOTE: This command is only supported during the BIND command.
Team properties 142 Error conditions If you receive the error messa ge "Failed to create new team," check the following: ⢠All adapters have the same frame types. ⢠The CPQANS team has the same frame type as the adapters. ⢠That CPQANS is loaded once for each team. ⢠A "BIND CPQANS <adapter name>" command has b een issued for each adapter in the team. ⢠All adapters in a team are connect ed to the same network segment. ⢠CPQANS is not being bound to an unsupported adapter. NOTE: The "CONFIG" command will not reflect CPQANS until the command "LOAD CPQANS COMMIT" has be en successfu lly issued. Setting NetWare probe and balance variables NetWare provides several probe and balance variables that provide the optimum performance for most configurations. However, you can modify these to me et your special configuration needs. Each of these variables is described below. You can use the followin g formula to determine th e relative range of many of these parameters: probe_send_interval<(probe_check_interval probe_recheck_interval*probe_retry_count) ⢠load CPQANS balance_interval=xxx [team=nnn] o Default = 180 o Range = 18â1080 o Description: Changes the team's balance interv al. XXX counts 1/18 second ticks. Only even numbers are supported. Odd numbers are round e d down. This can also be used after the COMMIT command. ⢠load CPQANS balance_set_default [team=nnn] Description: Restores the te am's balance interval to the factory default setting. ⢠load CPQANS probes=[on | off] [team=nnn] o Default = Probes enabled o Description: Enables and disables probe packet generation. Probes are not supported in FEC, GEC, and IEEE 803.3ad modes. This can al so be used after the COMMIT command. ⢠load CPQANS probes=[broadcas t|multicast] [team=nnn] o Default = Probes sent in broadcast o Description: Hangs probe addressin g to broadcast or multicast. This cannot be used after the COMMIT command. ⢠load CPQANS probe_retr y_count=xxx [team=nnn] o Default = 7 o Range = Relative to other pa rameters o Description: Changes the team's probe retry coun t value. XXX is the number of retries to send probes, in case the probes are not received by other members of this team. This can also be used after the COMMIT command. ⢠load CPQANS probe_burs t_size=xxx [team=nnn]
Team properties 143 o Default = 3 o Minimum = 1 o Description: Changes the team's number of probes to send in case of a retry. This can also be used after the COMMIT command. ⢠load CPQANS probe_send_i nterval=xxx [team=nnn] o Default = 18 o Range = Relative to other pa rameters o Description: Changes the time interval sending of regular probes. XXX coun ts 1/18 second ticks. Only even numbers are supported. Odd numbers ar e rounded down. This can also be used after the COMMIT command. ⢠load CPQANS probe_check_ interval=xxx [team=nnn] o Default = 2 o Range = Relative to other pa rameters o Description: Changes the time interval betw een probes and checks them. XXX counts 1/18 second ticks. Only even numbers are supported. Odd numbers are rounded down. This can also be used after the COMMIT command. ⢠load CPQANS probe_recheck_ interval=xxx [team=nnn] o Default = 2 o Range = Relative to other pa rameters o Description: Changes the time interval between probe retries and checks them. XXX counts 1/18 second ticks. Only even numbers are supported. Odd numbers are rounded down. This can also be used after the COMMIT command. ⢠load CPQANS probe_set_default [team=nnn] Description: Restores the team's pr obe settings to the factory default settings. (Does not affect On /Off or Multicast/Broadcast settings.) ⢠load CPQANS tx_ecbs_to_use=xxx [team=nnn] o Default = 64 o Minimum = 0 o Description: Number of Tx EC Bs to allocate when the Virtual Adapter is created. ⢠load CPQANS RESET LBN=xxx [team=nnn] Description: Resets a bound adapt er. This accesses the adapter by its Logical Board Number. Supported keywords For a list of supported keywords, see Adv a nced Server Program for NetWare (on page 110 ). Examples of NetWare teaming The following are examples of load and bind statemen ts that you should add to your AUTOEXEC.NCF file to set up adapter Teaming. HP recommends that you try some of these examples from the command line before writing them to the AUTOEXEC.NCF file. EXAMPLE 1: Simple NFT team The following is an example of an NFT Team that co nsists of an HP Gigabit server adapter and an HP Fast Ethernet server adapter. For performance reasons the Gigabit server adapter is the primary adapter and the Fast Ethernet server adapter is the backup.
Team properties 144 Note that ONLY an NFT mode Team can combine an HP Gigabit server adapter (server adapters that us e N1000.LAN) and an HP Fast Ethernet ser ver adapter (use N100.LAN). (Exception: The NC71xx Gigabit over copper server adapters can be teamed with adapters of different speeds.) See Overview of NetWare teaming (" NetWare teaming " on page 139 ) for information about mixing adapters. Note the following: ⢠For simplicity, in this example, only the default frame type 802.2 is used. By default, CPQANS uses the frame type 802.2, so you do not need to specify 802.2 on the command line. ⢠By default, CPQANS uses the NFT mode. In this example, only one instance of each adapter is present in the system, so you do not need to specify a s lot number. ⢠This example checks the status of the Team after each command. Example: 1. Load CPQANS.LAN to form the basis of a Team. LOAD CPQANS FRAME=ETHERNET_802.2 Next, check the status of the Team. LOAD CPQANS STATUS 2. Load the base driver for each adapter and frame type for the Team. As this example uses only 802.2, each adapter has its base driver loaded only once. LOAD N1000 LOAD N100 Next, check the status of the Team. LOAD CPQANS STATUS 3. When you load CPQANS.LAN, it looks like a protocol , so you can bind it to the physical adapter. Bind CPQANS.LAN to each physical adapter in the Team. Notice the key word "PRIMARY." For performance reasons the Gigabit server adapter should be the primary adapter. BIND CPQANS N1000 PRIMARY BIND CPQANS N100 Next, check the status of the Te am. Notice that the Team now has two adapter instances with the Team. Also notice that the Team is in the "BEFORE COMMIT" state. The HP Gigabit server adapter is the "PRIMARY. " LOAD CPQANS STATUS 4. Finalize the Team with the "COMMIT" command. Sinc e we have not chosen the mode of the Team, it is NFT by default. LOAD CPQANS COMMIT Next, check the status of the Team . Notice that NFT is now the Team mode and t hat the Team is now in the "COMMIT-DONE" state. No changes to th e Team can be made when the Team is in the "COMMIT-DONE" state. LOAD CPQANS STATUS 5. Now that the Team is in the "COMMIT-DONE" st ate, the CPQANS Team appears as a virtual adapter. You may now bind a protocol to CPQA NS. ALWAYS ensure that your adapters and CPQANS are using the same frame typ e. BIND IPX CPQANS NET=102 Next, check the status of the Team. LOAD CPQANS STATUS Example 2: TLB Team with Ethernet_II frame type This example is slightly more complicated. In this example, we will only use one frame type, Ethernet _II. Since Ethernet_II and TLB are not the CPQANS defaults , you must specify them both at the command line.
Team properties 145 This example also differs from th e previous example in that the Te am will be in the Transmit Load Balancing (TLB) mode. This example assumes a server with two HP Fast Ethern et server adapters, one in PCI Slot 1 and the other in PCI Slot 2. Note that ONLY an NFT mode Team can combine an HP Gigabit server adapter (server adapters that use N1000.LAN) and an HP Fast Ethe rnet server adapter (server adapters that use N100.LAN). See Overview of NetWare teaming (" NetWare teamin g " on page 139 ) for more information about mixing adapters. Use the "LOAD CP QANS STATUS" command to view the Team status after each of the following commands. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS.LAN to form the ground work for the Team. Since 802.2 is the default frame type we need to specify the frame type Ethernet_II. In a ddition, TLB should be specified, since NFT is the default Team "MODE." LOAD CPQANS FRAME=ETHERNET_II 3. Load the base drivers for each adapter in the Te am. Just as 802.2 is the default frame type for CPQANS, 802.2 is also the default frame type for HP Fast Ethernet. Thus, Ethernet _II must be specified during the driver load. In addition, since two different HP Fast Ethernet server adapters are in use, each one needs to be assi gned a name for later reference. LOAD N100 SLOT=1 FRAME=ETHERNET_II NAME=PRIMARY_N100_II LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=SECONDARY_N100_II 4. Bind CPQANS.LAN to each physical adapter usin g the names you defined when the base drivers were loaded. BIND CPQANS PRIMARY_N100_II BIND CPQANS SECONDARY_N100_II 5. If you have not specified your Te am as an Transmit Load Balance (TLB) Team, you may do it now. LOAD CPQANS MODE=ALB 6. If you forgot to choose the frame type or Te am mode during the or iginal "LOAD CPQANS" command, you can specify the desired frame type and Team mode any time before the "COMMIT" command. 7. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 8. Bind both IPX and IP as if CP QANS were a physical adapter BIND IPX CPQANS NET=112 BIND IP CPQANS ADDRESS=192.168.1.1 MASK=255.255.255.0 Example 3: Two teams (one NFT and one SLB) This example creates two Teams. The first is an NFT Team consisting of one HP Gigabit server adapter (N1000.LAN) and one HP Fast Ethernet server adap ter (N100.LAN). The second Team has four HP Fast Ethernet server adapters (N100.LAN) configured as a Switch-assisted Load Balancing (SLB) Team. SLB is the implementation of port aggre gation. This example uses IPX on Ethernet 802.2 and IP with the Ethernet II frame type. See Overview of NetWare teaming (" NetWare teamin g " on page 139 ) for more information about SLB Team implementation. Use the "LOAD CPQANS STATUS" command to vi ew the Team status after each command. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS.LAN for the first Team. By default yo u do not need to include the first Team's ID number on the command line. However, you will need to include it for every Tea m after the first. If
Team properties 146 you do not include "TEAM =1" CPQANS defaults to T eam 1. Thus, the "TEAM" command line switch is optional for the first Team. LOAD CPQANS FRAME=ETHERNET_802.2 NAME=GIG_FT_E82 TEAM=1 MODE=NFT LOAD CPQANS FRAME=ETHERNET_II NAME=GIG_FT_EII TEAM=1 3. Load the base drivers for the first Team. LOAD N1000 SLOT=1 FRAME=ETHERNET_802.2 NAME=GIG_E82 LOAD N1000 SLOT=1 FRAME=ETHERNET_II NAME=GIG_EII LOAD N100 SLOT=2 FRAME=ETHERNET_802.2 NAME=100_E82 LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=100_EII 4. Bind CPQANS to the base driver for each adapter and each frame type. Bind to the name of the base driver instance (assigned when the driver was loaded) and specify the Team ID number. BIND CPQANS GIG_E82 TEAM=1 BIND CPQANS GIG_EII TEAM=1 BIND CPQANS 100_E82 TEAM=1 BIND CPQANS 100_EII TEAM=1 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT TEAM=1 6. Bind IPX to 802.2 and bind IP to Ethernet II. BIND IPX GIG_FT_E82 NET=102 BIND IP GIG_FT_EII ADDR=128.181.11.1 MASK=255.255.255.0 7. Load CPQANS with the second team. Name each frame type with the "NAME" keyword and specify the second Team with the "TEAM=2" keywor d. You can also specify the Team "MODE" FEC. LOAD CPQANS FRAME=ETHERNET_802.2 NAME=FEC_TEAM_E82 TEAM=2 MODE=FEC LOAD CPQANS FRAME=ETHERNET_II NAME=FEC_TEAM_EII TEAM=2 8. Load the base drivers for the second Team. LOAD N100 SLOT=6 FRAME=ETHERNET_802.2 NAME=FEC_1_E82 LOAD N100 SLOT=6 FRAME=ETHERNET_II NAME=FEC_1_EII LOAD N100 SLOT=5 FRAME=ETHERNET_802.2 NAME=FEC_2_E82 LOAD N100 SLOT=5 FRAME=ETHERNET_II NAME=FEC_2_EII LOAD N100 SLOT=4 FRAME=ETHERNET_802.2 NAME=FEC_3_E82 LOAD N100 SLOT=4 FRAME=ETHERNET_II NAME=FEC_3_EII LOAD N100 SLOT=3 FRAME=ETHERNET_802.2 NAME=FEC_4_E82 LOAD N100 SLOT=3 FRAME=ETHERNET_II NAME=FEC_4_EII 9. Bind CPQANS to the base driver for each adapter and each frame type. Bind to the name of the base driver instance (assigned when the driv er was loaded) and specify the Team number. BIND CPQANS FEC_1_E82 TEAM=2 BIND CPQANS FEC_1_EII TEAM=2 BIND CPQANS FEC_2_E82 TEAM=2 BIND CPQANS FEC_2_EII TEAM=2 BIND CPQANS FEC_3_E82 TEAM=2 BIND CPQANS FEC_3_EII TEAM=2 BIND CPQANS FEC_4_E82 TEAM=2 BIND CPQANS FEC_4_EII TEAM=2 10. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT TEAM=2 11. Bind IPX to 802.2 and bind IP to Ethernet II. BIND IPX FEC_TEAM_E82 NET=202 BIND IP FEC_TEAM_EII ADDR=128.181.22.1 MASK=255.255.255.0 Example 4: One NC series server adapter and two VLANs NOTE: VLAN is not supported on heterogeneous teaming.
Team properties 147 For CPQANS.LAN to work in VLAN mode, the adap ter (or adapters) needs to be connected to a "tagged" port (a "trunk port") on an IEEE 802.1Q compliant switch. This example uses 802.1Q VLAN, which is suppo rted by all HP server adapters that support CPQANS.LAN. (Exception: The NC6134, NC6132, an d NC6133 adapters do not support V LANs.) This example uses an HP Fast Ethernet server adapter an d is configured for two VLANs (VLAN 5 and VLAN 6). It also uses all four NetWare frame types and binds both the IPX and IP protocols. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS and specify the VLANID for each fram e type and each VLAN the server is supposed to be part of. With a single Te am configuration, "Team=1" is optional, since that is the default setting. a. First, load CPQANS for VLAN 5, assign ing a name of 5xx to each instance: LOAD CPQANS VLANID=5 FRAME=ETHERNET_802.2 NAME=582 TEAM=1 LOAD CPQANS VLANID=5 FRAME=ETHERNET_802.3 NAME=583 TEAM=1 LOAD CPQANS VLANID=5 FRAME=ETHERNET_SNAP NAME=5SP TEAM=1 LOAD CPQANS VLANID=5 FRAME=ETHERNET_II NAME=5II TEAM=1 b. Next, load CPQANS for VLAN 6, assign ing a name of 6xx to each instance: LOAD CPQANS VLANID=6 FRAME=ETHERNET_802.2 NAME=682 TEAM=1 LOAD CPQANS VLANID=6 FRAME=ETHERNET_802.3 NAME=683 TEAM=1 LOAD CPQANS VLANID=6 FRAME=ETHERNET_SNAP NAME=6SP TEAM=1 LOAD CPQANS VLANID=6 FRAME=ETHERNET_II NAME=6II TEAM=1 3. Load the base driver for each frame type. LOAD N100 SLOT=5 FRAME=ETHERNET_802.2 NAME=E82 LOAD N100 SLOT=5 FRAME=ETHERNET_802.3 NAME=E83 LOAD N100 SLOT=5 FRAME=ETHERNET_SNAP NAME=ESP LOAD N100 SLOT=5 FRAME=ETHERNET_II NAME=EII 4. After loading CPQANS for all the desired VLANs, bind CPQANS to the base driver using the name you defined when the base driv ers were originally lo aded. BIND CPQANS E82 TEAM=1 BIND CPQANS E83 TEAM=1 BIND CPQANS ESP TEAM=1 BIND CPQANS EII TEAM=1 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 6. Bind the protocols to each instance (VLAN and fram e type) of CPQANS in the same manner as if it were a physical adapter. BIND IPX 582 NET=502 BIND IPX 583 NET=503 BIND IPX 5SP NET=505 BIND IPX 5II NET=511 BIND IPX 682 NET=602 BIND IPX 683 NET=603 BIND IPX 6SP NET=605 BIND IP 6II ADDR=192.168.62.1 MASK=FF.FF.FF.0 Example 5: TLB team with multiple VLANs
Team properties 148 NOTE: VLAN is not supported on heterogeneous teaming. You can implement Teaming and VLAN at the same ti me. Because the script gets large quickly, this example only uses one frame type. To have both IP an d IPX bound to it, the Ethernet_I I frame type will be used instead of Novell's default 802.2. A homogene ous Team of two load-balanced HP Fast Ethernet server adapters configured as members of five different VLANs will be used here. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS.LAN with the VLANID key word once fo r each VLAN that the server is intended to communicate with. Assign a name so each instance can be identified later. LOAD CPQANS FRAME=ETHERNET_II VLANID=10 NAME=V10 TEAM=1 MODE=ALB LOAD CPQANS FRAME=ETHERNET_II VLANID=20 NAME=V20 TEAM=1 LOAD CPQANS FRAME=ETHERNET_II VLANID=30 NAME=V30 TEAM=1 LOAD CPQANS FRAME=ETHERNET_II VLANID=40 NAME=V40 TEAM=1 LOAD CPQANS FRAME=ETHERNET_II VLANID=50 NAME=V50 TEAM=1 3. Load the base drivers for each adapter in the Team. LOAD N100 SLOT=1 FRAME=ETHERNET_II NAME=PRIMARY_II LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=SECONDARY_II 4. Bind CPQANS to each physical adapter. BIND CPQANS PRIMARY_II BIND CPQANS SECONDARY_II 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 6. Bind both IPX and IP to CPQANS for each VLAN. 7. Bind IPX first. BIND IPX V10 NET=1011 BIND IPX V20 NET=2011 BIND IPX V30 NET=3011 BIND IPX V40 NET=4011 BIND IPX V50 NET=5011 Then bind IP. BIND IP V10 ADDR=192.168.1.1 MASK=FF.FF.FF.0 BIND IP V20 ADDR=192.168.2.1 MASK=FF.FF.FF.0 BIND IP V30 ADDR=192.168.3.1 MASK=FF.FF.FF.0 BIND IP V40 ADDR=192.168.4.1 MASK=FF.FF.FF.0 BIND IP V50 ADDR=192.168.5.1 MASK=FF.FF.FF.0 Example 6: IEEE VLAN NOTE: VLAN is not supported on heterogeneous teaming. This example creates two IEEE VLANs on a single HP Fa st Ethernet server adapter, loads two frame types (Ethernet_802.2 and Ethernet_II) on each VLAN, bind s IPX to each loaded frame type, and binds IP to each instance of the Ethernet_II frame type. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP
Team properties 149 2. Load CPQANS twice for each frame type, once for each VLAN. As sign the VLANID number keyword for each instance. LOAD CPQANS FRAME=ETHERNET_802.2 NAME=E82_Vl7 TEAM=1 VLANID=7 LOAD CPQANS FRAME=ETHERNET_II NAME=EII_Vl7 TEAM=1 VLANID=7 LOAD CPQANS FRAME=ETHERNET_802.2 NAME=E82_Vl8 TEAM=1 VLANID=8 LOAD CPQANS FRAME=ETHERNET_II NAME=EII_Vl8 TEAM=1 VLANID=8 3. Load the base driver for each frame type: LOAD N100 SLOT=2 FRAME=ETHERNET_802.2 NAME=PRI_E82 LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=PRI_EII 4. Bind CPQANS to each frame type on the physical adapter. BIND CPQANS PRI_E82 TEAM=1 BIND CPQANS PRI_EII TEAM=1 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 6. Bind the protocols to CPQANS for each VLAN and frame type. BIND IPX E82_Vl7 NET=702 BIND IPX EII_Vl7 NET=711 BIND IP EII_Vl7 ADDR=192.168.72.200 MASK=255.255.255.0 BIND IPX E82_Vl8 NET=802 BIND IPX EII_Vl8 NET=811 BIND IP EII_Vl8 ADDR=192.168.82.200 MASK=255.255.255.0
VLAN properties 150 VL AN pr oper tie s In this section VLAN info rmation ............................................................................................................... .................. 150 VLAN information A Virtual LAN (VLAN) is a collection of network node s that share the same broadcast domain regardless of their physical location or connection point to the network. A VLAN serves as a logical workgroup with no physical barriers and allows users to share inform ation and resources as thou gh located on the same LAN. The HP Network Configuration Utility supports VL ANs in Windows S erver 2003 and Windows Server 2003 x64 environments. Before creating VLANs, you should be aware of the following: ⢠This feature supports IEEE 802.1Q VLAN tagging only. ⢠This feature supports a VLAN identifier ra nge of 1 to 4094. ⢠Up to 64 Ethernet VLANs can be defined for a team or a single, non-teamed adapter. ⢠Only one iSCSI VLAN can be defined on a multifunction adapter ⢠VLANs cannot be defined for individual adapters th at are members of a team. VLAN defin ition must be done at the Team level. ⢠If an adapter with VLANs is added to a team, th e VLANs defined for the adapter will be deleted. ⢠When a VLAN is defined for an adap ter, the properties of that adap ter will be modifi ed if necessary to allow proper VLAN functionality. For example, 802.1p QoS must be enabled on an adapter or team with VLANs. Likewise, when an adapter is ad ded to a team that has VLANs defined, then that adapter will have 802.1p QoS enabled. ⢠The VLAN defined as the âDefault VLANâ cannot be deleted if more than one VLAN is defined for a team. For example, if VLAN 1 needs to b e deleted and it is defined as the âDefault VLANâ, chan ge the âDefault VLANâ to a different VL AN. VLAN 1 can then be deleted. ⢠When creating a VLAN for a team, the Fast Path VLAN ID is disabled if the Spanning Tree Protocol is set to IEEE 802.1D. ⢠When the last VLAN is removed from an adapter or the adapter is removed from a team with VLANs, the user will be given the option to disabl e 802.1p QoS on the adapter or allow it to remain enabled. For more in formation: ⢠Removing a VLAN
ProLiant Essentials 151 Pr oL iant E ss ent ials In this section Networking pack installa tion ................................................................................................... .............. 151 Intelligent Networking Pack-Windows Edition .......................................................................................... 151 Intelligent Networking Pack-Linux Edition ...................................................................................... ........... 152 Accelerated iSCSI for stand-alone and c-Class BladeSystem Multifunctio n server ad apters ............................ 153 Accelerated iSCSI for embedded HP Multifunction serv er adapt ers............................................................. 153 Networking pack installation Included with the networking software that shippe d with your HP server adapter are a number of advanced features that extend the basic networking capabilities of your eq uipment. You can activate these advanced features through the purchase and in stallation of the foll owing ProLiant Essentia ls networking pack licenses: ⢠Intelligent Networking PackâWindows Edition ⢠Intelligent Networking PackâLinux Edition ⢠Accelerated iSCSI Pack for Embedded Multifunction Server Adapters Installing a networking pack license from the NCU The HP ProLiant Essentials Network License Manager a llows you to install and track network licenses on your system. To access the License Mana ger dialog: 1. From the NCU Main window, click License Manager. The License Manager dialog appears. 2. Click Help to access the License Manage r help file. NOTE: The HP Accelerated iSCSI feature comes st andard with stand-alone HP Multifunction Server Adapters and Multifunction Se rver Adapters for HP BladeSystem. Intelligent Networking Pack-Windows Edition The HP Network Configuration Utility (NCU) allows you to configure and monitor HP server network adapters running under Windows® 2000, Windows Server 2003, or Windows Server 2003 x64. Although the NCU provides an extensive list of features through the Foundation Pack, additional advanced networking capabilities are available with the installation of an HP ProLiant Essentials Intelligent Networking Pack license. The Intelligent Networking Pack (INP)âWindows Edition includes: ⢠Virus Throttle, a network pack et-filtering feature that helps slow down the spread of viru ses on your system. Virus Throttle monitors all outbound connect ion requests and counts the number of un ique connections. It detects abnormal ( âvirus-likeâ) behavior in the re quests, and slows down excessive connection requests to new hosts un til you can determine if they are viral in nature and take action. ⢠Advanced networking capabilities that allow ProLia nt servers to adapt and change the network path to achieve maximum reliability and performance. These features detect and analyze network
ProLiant Essentials 152 bottlenecks or broken network linkages, and steer traffic to the optimum route based on network conditions. Advanced networ king features include: o Fast Path FailoverâAllows a ProLiant ser ver to use the quickest available path to the core network for all server traffic. o Active Path FailoverâAllows a ProLiant server to detect blocked paths an d to redirect data along an unblocked path to the core network. o Router Path FailoverâProvides red undancy by grouping routers into one router g roup. One router serves as the active router. If that router fails, a second ro uter assumes rout er functionality. ⢠Dual Channel Load Balancing, a special type of te aming that allows the creation of two teams, called groups, inside of a single team. Each group is assigned one or more teamed ports and can be connected to a different switch to provide swit ch fault tolerance. Full inbound and outbound load balancing is provided across both groups. Should an y group completely fail, caus ed by a failure of all teamed ports in the group or by a failure of the groupâs switch, the team remains available through the other group. These types of Dual Channel Load Balancing teams can be configured : o Switch-assisted Dual Channel Load Balancing o Switch-assisted Dual Channel Network Fault Tolerance o Switch-assisted Dual Channel NFT and Preference Order o 802.3ad Dynamic Dual Channel Load Balancing o 802.3ad Dynamic Dual Channel Network Fault Tolerance o 802.3ad Dynamic Dual Channel NFT and Preference Order NOTE: Dual Channel Network Fault Tolerance te aming capabilities are not supported on systems running Windows® 2000. ⢠Cisco Discovery Protocol (CDP) and Link Layer Discov ery Protocol (LLDP) are both layer two protocols that are neighbor device discovery protocols that run on supported networ king devices such as routers, access servers, bridges, and switch es. Using CDP or LLDP, an HP adapter team can passively discover configuration information about directly connected networ k devices that have CDP or LLDP enabled. NOTE: LLDP is not supported on systems runnin g Windows® 2000. For additional information about INP features and how to install a license on your system, refer to the HP ProLiant Essentials Intellig ent Networking Pack License â Windows Edition User Guide located on the HP website ( http://www.hp.com ). For information on how to purchase an HP ProLia nt Essentials Intelligent Networking PackâWindows Edition license, go to the HP website ( http://www.hp.com ) or contact your HP reseller. Intelligent Networking Pack-Linux Edition The Intelligent Networking PackâLinux Edition includes Virus Throttle, a network packet-filtering feature that helps slow down the spread of virus es on your sy stem. Virus Throttle monitors all outbound connection requests and counts the number of unique connections. It detects abnorma l (âvirus-likeâ) behavior in the requests, and slows down excessive co nnection requests to new hosts until you can determine if th ey are viral in nature and take action.
ProLiant Essentials 153 To enable Virus Throttle for Linux, an Intelligent Netw orking PackâLinux Edition lice nse must be installed on the system. For additional information about Virus Throttle and how to install an Intelligent Networking Pack license on your system, refer to the HP ProLiant Essentials Intelli gent Networking Pack License â Linux Edition User Guide located on the HP website ( http://www.hp.com ). For information on how to purchase an HP ProLiant Essentials Intelligent Netw orking PackâLinux Edition license, go to the HP website ( http://www.hp.com ) or contact your HP reseller. Accelerated iSCSI for stand-alone and c-Class BladeSystem Multifunction server adapters The HP Accelerated iSCSI feature comes standard with HP stand-alone Multifunction server adapters and Multifunction server adapters for c-Class BladeSystem. The Accelerated iSCSI feature runs under Wind ows® 2000, Windows Server 2003, and Windows Server 2003 x64 and allows you to complete the following: ⢠Enable an iSCSI device ⢠Adjust the SRB Timeout Delta for an adapter ⢠Configure an iSCSI VLAN For additional information about the HP A ccelerated iSCSI feature, refer to the HP Accelerated iSCSI for Multifunction Network Adapters User Guide located on the HP website ( http://www.hp.com ). Accelerated iSCSI for embedded HP Multifunction server adapters The HP Accelerated iSCSI capability is available for embedded HP Multifunction server adapters through the purchase and installation of a ProLiant Essent ials Accelerated iSCSI Pack license. The ProLiant Essentials Accelerated iSCSI Pack li cense is part of the ProLiant Ess entials software family and is integrated with HP Systems Insight Manager (SIM), Sm artStart, and HP Management agents. It is included with the networking software (releas e 8.35 and higher) found on the HP website ( http://www.hp.com ). NOTE: The HP Accelerated iSCSI feature comes st andard with stand-alone HP Multifunction Server Adapters and Multifunction Se rver Adapters for HP BladeSystem. The Accelerated iSCSI feature runs under Wind ows® 2000, Windows Server 2003, and Windows Server 2003 x64 and allows you to complete the following: ⢠Enable an iSCSI device ⢠Adjust the SRB Timeout Delta for an adapter ⢠Configure an iSCSI VLAN for a Multif unction adapter For additional information about the HP A ccelerated iSCSI feature, refer to the HP Accelerated iSCSI for Multifunction Network Adapters User Guide located on the HP website ( http://www.hp.com ). For information on how to purchase HP ProLiant Essentials value pack li censes for networking, go to the HP website ( http://www.hp.com/servers/proliantessentials ) or contact your HP reseller.
Diagnostics 154 Di agnos ti c s In this section PCI shared interrupts .......................................................................................................... ................... 154 HP NC-Series Intel adapter diag nostics ......................................................................................... .......... 154 HP NC-Series Broadcom adapter diag nostics ...................................................................................... .... 156 HP NC-Series Broadcom Multifunction adapter di agnostics ....................................................................... 1 67 PCI shared interrupts Shared interrupts are normal in PC I servers, but not all PCI device dr ivers support them. Although these adapters support shared interrupts, other PCI drivers used by your server may not. If you are having problems with a shared interrupt, try one of the following: ⢠Use the BIOS setup program for your ser ver to ch ange the interrupt assign ed to this adapter. ⢠Be sure that the adapter device driver loads AFTER the driver or any other device sharing this interrupt. ⢠Contact the manufacturer of the other adapter or add -in card that is sharing this interr upt for a new driver. HP NC-Series Intel adapter diagnostics Overview The diagnostic software lets you test adapters to see i f there are any problems with the adapter hardware, cabling, or network connections. HP recommends that you run diagnostics every time you in stall an adapter. Prerequisites The diagnostic utility runs under MS-DOS® v6.22 or later. It will not run from a DOS window in IBM OS/2® or any Microsoft Windows prod ucts. This util ity is designed to test hardware operation and confirm the adapter's ability to communicate with an other adapter in the same network, which primarily confirms cabling status. This diagostic utility i s not a throughput measurement tool. The utility can test the adapter wheth er or not there is a responder on the network; however, to co nduct a thorough test you should set up a second system on the network as a responder prior to starting a test. Starting the DIAGS.exe test s NOTE: If there is a DOS network driver present, such as NDIS2 or DOS-ODI, the t est utility and/or the network driver could become unstable. You should reboot and ensure that there are no network drivers loaded.
Diagnostics 155 NOTE: The DOS diagnostic utility, DIAGS.EXE, wi ll always use auto-negotiation to determine the link type listed under the "View Adapter Co nfiguration" section. If the network device's port to which your adapter is connected is set to auto-negotiate, then the link type will reflect an auto-negotiated value. If the network device's port has its dupl ex value forced to full, then the link type will show a half-duplex value. 1. If you have not already done so, download and extract the HP SoftPaq file and then run MAKEDISK.BAT to create driver diskettes. 2. Boot to DOS or the EFI shell. 3. From the DOS prompt navigate to the \APPS\DIAGS\N100X directory. 4. Type DIAGS and press the Enter key. The test utility program automatically scans the hardware and lists all NC110X, NC310F, NC340x, NC36 0x, NC364X, NC31xx, NC61xx, and NC71xx adapters. If you have a multiport adapter, each port is listed separately starting with "Port 1." 5. Using the arrow keys, select the adapter to test then press the Enter key. The main menu of the diagnostic utility appears. Viewing adapter configuration Selecting the View Adapter Configuration option displays the adapter configuration screen, which lists the adapter properties such as MAC address, media type, bus number and so on. 1. Press the F5 key to view additional information on th e PCI slot occupied by the adapter. This is primarily used for troublesh ooting by technical support. 2. Press the Enter key twice to return to the main menu. Accessing the test adapter menu Selecting Test Adapter from the main menu displays the test a dapter menu. The test me nu allows the user to begin the adapter tests, change test options, di splay diagnostic log files, and access the networking submenu. Beginning adapter tests Selecting Begin Adapter Tests brings up the test scr een. While tests are being performed, a rota ting spinner is shown letting the user know that the applic ation is still running. The result s of the tests are displayed as each test is performed. If multiple test passes are selected, then the results contain a count of test failures. A list containin g zeros means that all tests have passed. If there is no responder on the network, then the Ne twork Test will indicate a failure. To correct this situation, set up a system on the network to act as a responder then re-run the test. Changing test options The test setup screen allows you to select and conf igure the specific t ests desired. Toggle betwee n options by using the arrow keys and pressing the Enter key. The utility displays only those test options that are supported by your adapter. The following is a list of the available test options. ⢠Number of test passes âThis option determines how many times a test runs. While the test is running, the information on the screen is continuously updated. When you pres s the Esc key, the test is cancelled and control is returned to the Test Adapter menu. ⢠Diagnostic Log âThis test is disabled by default. If enab led, the program will ask for a directory for the log file. The file it places here is named DIAGS.LOG.
Diagnostics 156 ⢠Device Registers âTest patterns are written, read, and verified through the adapter's device registers to check proper functionality. ⢠FIFO âThis option writes test bit patterns to the a dapter's FIFO buffer s to check the FIFOs proper functionality. ⢠EEPROM âThis option tests both the readability of th e EEPROM as well as the integrity of the data stored in the EEPROM. It reads EEPROM and calculates the checksum. This checksum is then compared to the checksum stored in the EEPROM. If the values are not identical, the test r eports failure. ⢠Interrupt âThis option tests the adapter's ability to ge nerate an interrupt and have it propagated through the system to the Programmable Interrupt Contr oller (PIC). The test triggers an interrupt by setting the interrupt cause register and then verifies that an interrupt has been triggered. On EFI, interrupts are not supported , Therefore, this test sets the interrupt cause regist er and reads the interrupt set register, which verifies the c a rd internally registered an interrupt. ⢠Loopback Tests âThese options are internal loopback test s. These tests set the adapter in the appropriate loopback mode and send packets back through the adapter's receive circuitry and logic. ⢠Link âThis option checks to see whether or not the adapter has link. ⢠Network Test âThis option tests netwo rk communication. It looks for a responder a nd then sends packets. If no responder is found, then the test re ports failure. If packets are received back from the responder, the test reports succe ss. Displaying Diagnostic Log The Display Diagnostic Log option allow s you to view a detailed report of the tests you just ran. When Diagnostic Log is enabled, test re sults are recorded in a log file named DIAGS.LOG. If it does not already exist, the test utility creates it. If it already exist s , new data is appended to it. The DISPLAY DIAGNOSTIC LOG command displays the contents of the log file for your convenience. Each entry in the log fil e is time stamped. The test run banner identifies the test ed adapter according to its bus slot address. Accessing the networking submenu The networking submenu allows you to setup the adapter as a responder and to de tect a spanning tree on the network. ⢠Setting up as a responder âThis allows the user to set up th e adapter as a responder so another system can perform the continuous network test. Sele cting this option displays the tr ansmit/receive screen. This test will fail if the adapter does no t have link. Although you can use a variety of adapters as responders, and either connect directly (with a crossover cable) or through a switch, ideal results are obtained with a same-type adapter. When you press the Esc key, the responder operation is cance lled and control is immediately returned to the Test Adapter menu. ⢠Detecting spanning tree âThis allows the user to detect if a spanning tree is used on the network. HP NC-Series Broadcom adapter diagnostics Overview The diagnostic software lets you te st adapters to see if there are any problems with the adapter hardware, cabling, or network connections. HP recommends that you run diagnostics every time you in stall an adapter.
Diagnostics 157 Prerequisites The diagnostic utility runs under MS-DOS® v6.22 or later. It will not run from a DOS window in IBM OS/2® or any Microsoft Windows prod ucts. This util ity is designed to test hardware operation and confirm the adapter's ability to communicate with an other adapter in the same network, which primarily confirms cabling status. This diagostic utility is not a throughput measurement tool. The utility can test the adapter whether or not there is a responder on the network; however, to co nduct a thorough test you should set up a second system on the network as a responder prior to starting a test. These diagnostics use pmode/w DOS Extender V1.33. ⢠Input file list: The following files sho uld be foun d in the same location as the q57diag.exe file. o 5701.bin o 5703.bin o config.sys o cpu05.bin o cpu.bin o cpu14a.bin o cpu14b.bin o cpudg05.bin o cpudiag.bin o cpusj.bin o cpumem.bin o diagcfg.bin o flshdg05.bin o flshdg5x.bin o flshdiag.bin o flshd14a.bin o flshd14b.bin o flshdgsj.bin o q57diag.exe o ump14a.bin o ump14b.bin Do not use emm386.exe. Himem.sy s and emm386.exe are both memory managers. If they are both loaded, the diags aborts and the following error may occur: Exception: OE Error Code: 00000000 About the Q57DIAG.exe diagnostic tests The tests are divided into four group s: Register tests, Memory test s, Miscellaneou s tests, and Data tests. They are identified as group A, B, C, and D. Group A: Register tests ⢠A1. Indirect register test
Diagnostics 158 Using indirect addressing method, this test write s increment data into the MAC Hash Register table and reads back for verification. Th e memory read/write is done 100 times while incrementin g test data. ⢠A2. Control register test Each register specified in the configurati on contents are defined as read only bit and read/write bit. The test writes zero and one into the test bits to ensure the read only bits are not changed and read/write bits are changed accordingly. ⢠A3. Interrupt test This test verifies the interrupt functionality. It enables interrupt and then waits for the interrupt to occur. It waits for 500ms and reports an erro r if it could not generate interrupts. ⢠A4. Built-In-Self test This test initiates Hardware Built-In-S elf-Test (BIS T) and then waits for the test result returned by hardware. The hardware could not generate interrupts. ⢠A5. PCI Cfg register test This test verifies the access inte grity of the PCI config register s. Group B: Memory tests ⢠B1. Scratch pad test This tests the scratch pad SRAM on boar d. The following tests are performed: o Address test: Writes each address with unique increment data. Reads back data t o ensure data is correct. After filling the entire data with th e unique data, the program reads back data again to ensure data stays the same. o Walking one bit test: For each address, data one is written and read back for testing. Then the data is shifted left one bit, so the data becomes two and the same test is run again. It repeats for 32 times until the test bit is shifted out of test da ta. The same test is repeated for the entire test range. o Pseudo random data test: A pre-calculated pseudo random data is used to write a unique data into each test RAM. After the first pass of th e test, the program reads back one more time to ensure data stays correct. ⢠B2. BD SRAM test This tests the BD SRAM by performing th e tests as described in test B1. The Scratch pad test. ⢠B3. DMA SRAM test This tests DMA SRAM by perfo rming the tests described in test B1. The Scratch pad test. ⢠B4. MBUF SRAM test This tests DMA SRAM by perfo rming the tests described in test B1. The Scratch pad test. ⢠B5. MBUF SRAM via DMA test Eight test pattern data are used in the test. They are described below. A 0x 1000 sized data buffer is used for this test. Before each pattern test, the buff er is initialized and filled with the test pattern . It then, performs size 0x1000 transmit DMA from host buffer to adapter MBUF memory. It verifies the data integrity in MBUF against host memory and repe ats the DMA for the entire MBUF buff er. Then it performs receive DMA from adapter to host. The 0x 1000-byte test buffer is cleared to zero before each receive-DMA. It verifies the data integrity an d the test is repeated for the enti re MBUF SRAM range. Test Pattern Description:
Diagnostics 159 "16 00's 16 FF's" Fill the entire host DMA buffer with 16 bytes of 00's and then 16 bytes of FF's. "16 FF's 16 00's" Fill the entire host DMA buffer wi th 16 bytes of FF's and then 16 bytes of 00's. "32 00's 32 FF's" Fill the entire host DMA buffer with 32 bytes of 00's and then 32 bytes of FF's. "32 FF's 32 00's" Fill the entire host DMA buffer wi th 32 bytes of FF's and then 32 bytes of 00's. "00000000's" Fill the entire host DMA buffer with all zeros. "FFFFFFFF's" Fill the entire host DMA buf fer with all FF's. "AA55AA55's" Fill the ent ire host DMA buffer with da ta 0xAA55AA55. "55AA55AA's" Fill the ent ire host DMA buffer with da ta 0x55AA55AA. Group C: Miscellaneous tests ⢠C1. NVRAM test An increment test data is used in the EEPROM test. It fills the test data into th e test range and reads it back to verify the content. After the test, it fills data with zeros to clear the memory. ⢠C2. CPU test This test opens the file cpu.bin. If the file exists and the content is good, it loads code to the Rx and Tx CPU and verifies CPU execution. ⢠C3. DMA test ⢠This tests both high and low prio rities DMA. It moves data from host memory to adapter SRAM, verifies data, and then moves data back to the host memory again to verify data. ⢠C4. MII test This function is identical to A2. Control Register Test. Each Register specified in the configurati on contents is define d as re ad only bit and read/write bit. The test writes zero and one into the test bits to ensure the read only bits are not changed and read/write bits are changed accordingly. ⢠C5. VPD test The content of VPD is saved first before performing the test. After it is saved, th e test writes on e of the five pattern test data, 0xff, 0xaa, 0x55, incremen t data, or decrement data, into VPD memory. By default, increment data pattern is used. It writes and reads back the data for the entire test range, and then it restores the original content. ⢠C6. ASF test The function of this test is as follows: o Reset test. Sets the reset bit and polls for self-clear ing. Verifies the reset value of th e registers. o Event Mapping Test. Sets SMB_A TTN bit by changing ASF_ATTN LOC bits. Verifies the mapping bits in TX_CPU or RX_CPU event bits. o Counter Test. Clear s WG_TO, HB_TO, PA_T O, PL_TO, RT _TO bits by setting those bits. Makes sure the bits are clear. Clears the Timestamp Coun ter. Writes a value 1 into each of the PL, PA, HB, WG, RT counters. Sets the TSC_EN bit. Polls each PA_TO bit and counts up t o 50 times. Checks if PL_TO gets set at the end of 50 times. Continues to count up to 200 times. Checks if all other TO bits are set and verifies th e Timestamp Counter is incremen ted. ⢠C7. Expansion ROM test This tests the ability to enable/disabl e/access the expansion ROM on the device . Group D: Driver associated tests ⢠D1. Mac loopback test
Diagnostics 160 This is an internal loopback data transmit/receive test. It initializes M AC into internal loopback mode, and transmits 100 packets. The data should be routed back to the receive channel and is received by the receive routi ne, which verifies the in tegrity of data. One Gigabit rate is used for this test. ⢠D2. Phy loopback test This test is the same as D1. Mac L oopback Test ex cept the data is routed back via physical layer device. One Gigabit rate is used for this test. ⢠D5. MII miscellaneous test (copper only) This function tests the auto-polling and phy-interrupt capabilities. Th ese are the functionalities of the phy. ⢠D6. MSI test This tests the Testing Message Interrupt Function to se e if it handles t his interrupt correctly. The default is disabled. Starting the Q57DIAG.exe tests 1. If you have not already done so, download and extract the HP SoftPaq file and then run MAKEDISK.BAT to create a driver disk. 2. Boot to DOS or the EFI shell. 3. From the DOS prompt or shell naviga te to the \APPS\DIAGS\Q57 directory. 4. Type Q57DIAG. exe and press the Enter key. The diagnostic te sts run automatically. Command line options The following are the available options in q57diag. They are case sensitive. Usage: q57diag.exe [options] ⢠-c <num> o Specifies UUT device number When more than one device is in the system, the devices are numb ered starting from zero. For example, if there are three devices det ected, the devices are numbered as 0, 1, and 2. In this case, entering the parameter -c 2 will sele ct the last found device as default UUT. In manufacture testing mode, by default, all device s are tested; however, if this opti on is used, only that selected device is tested. o Example: q57diag.exe -c 2 ⢠-smbaddr <num>- o Specifies the ASF SMBus A ddress ⢠-h o Specifies high resolution (80x50) vide o mode o Example: q57diag.exe -h ⢠-l <file> o Specifies the log file in which all diagnostic ou tput is saved. The default is no lo g file. o Example: q57diag.exe -l mylogfile.txt ⢠-p o Specifies Print on Error
Diagnostics 161 o Example: q57diag.exe -p ⢠-npol o Selects negative link polarity in TBI test o Example: q57diag.exe -npol ⢠-q Selects quick diagnostic mode ⢠w <value> o Enables WOL programming in manufacture mode After a successful manufacturing testing, the prog ram will set WOL to either enable or disable mode. Value = 0 to disable WOL and 1 to enable WOL. When -f is entered, software uses eeprom.bin's content for WOL setting. When -w is entered with -f, -w is ignored. o Example: q57diag.exe -w 1 o To disable/enable WOL for LOM design s, Lan On Motherboard, use the Rom Bios Setup Utility, RBSU. ⢠-x <value> o Enables PXE in manufacture mode After a successful manufacturing testing, the prog ram will set PXE to either enable or disable mode. Value = 0 to disable PXE and 1 to enable PXE. When -f is entered, software uses eeprom.bin content for PXE setting. When -x is entered with -f, -x is ignored. o Example: q57diag.exe -x 0 o To disable/enable PXE for LOM de signs, Lan On Motherboard, use the Rom B ios Setup Utility, RBSU. ⢠-asf <value> o Enables asf in manufacture mode o Value = 0 to disable; Value = 1 to enable ⢠-t <id> Disables test ⢠-T <id> o Enables test A certain test is enabled or disabled by default. The user can overwrite the enabling status by those options. The test ID must start with a letter 'A ', 'B', 'C', or 'D' to indicate the group and be followed by test numbers. Each digit of the numb er represents the sub-test numb er. For example, if the user wants to disable test A1 and A3. The option -t A13 should be entered. If no test numbers are entered, all tests in that group are selected. For the tests no t specif ied, the default setting will be used. o Example: q57diag.exe -t A15BC1 -T C4 -t D2 This disables A1, A5, B1, B2, B3, B4, B5, B6, C1, D2, and enables C4
Diagnostics 162 o Default Settings Enabled Tests: A1. Indirect Register Test A2. Control Register Test A3. Interrupt Test A4. Built In Self Test A5. PCI Cfg Register Test B1. Scratch Pad Test B2. BD SRAM Test B3. DMA SRAM Test B4. MBUF SRAM Test B5. MBUF SRAM via DMA Test C1. NVRAM Test C2. CPU Test C3. DMA Test C4. MII Test C5. VPD Test C6. ASF Test C7. Expansion ROM Test D1. Mac Loopback Test D2. Phy Loopback Test D4. MII Miscellaneous Test (Copper only) Disabled Tests: B6. External SRAM Test D3. External Loopback Test ⢠-I <num> o Use this option to specify the number (iteration) of times the tests ar e to be run. The default is to run one time. A number zero indicates loop fore ver. A control-C or control-break key can be used to break the loop. Any error detected will also stop testing after reporting the erro r. o Example: q57diag.exe -I 5 Runs tests five times ⢠-ver o Displays the version of the current test software/eeprom .bin o If this option is entered, it displays the softwa re version number/silkscreen r evision and then exits the program. o Example: q57diag.exe -ver ⢠-bus <num> o Tests UUT location o If this option is specified, the program will test all UUTs on the specified bus number. This option is ignored if -c option is entered.
Diagnostics 163 o Example: q57diag.exe -bus 5 Program tests UUTs on bus 5 ⢠-dpmi o Specifies the use of DPMI memory alloc ation o Uses DPMI memory allocation method to allo cate memory instead of malloc() or free(). ⢠-dir o Displays the version of the IPMI (Inte lligent Platform Ma nagement Interface) ⢠-errctrl <c> Specifies action on error -> a:abort , w:wait, l:loop, c:continue, s:skip ⢠-com <value> Enables com port value (1..4) ⢠-ems <size> Specifies external memory size in HEX to tst ⢠-ref Runs test with reference device ⢠-findref Detects reference device ⢠-fmac Enables input of Mac address from a file ⢠-lbm <n> Sets Mac loopback packets ⢠-lbp <n> Sets Phy loopback packets ⢠-lbe <n:n:n> o Sets external loopback packets o Format: <1000Mbps:100Mbps:10Mbps> ⢠-mba <value> o Enables/disables MBA o Value = 0 to disable; Value = 1 to enable ⢠-mbap <value> o Selects MBA protocol o Value = 0 for PXE; Value = 1 for RPL; Value = 2 for BootP ⢠-mbas <value> o Selects MBA speed o Value = 0 for auto; Value = 1 for 10HD; Value = 2 for 10FD; Value = 3 for 100H D; Value = 4 for 100FD; Value = 6 for 1000FD (fiber ) ⢠-firm <file> Updates device eeprom based on <file> image match
Diagnostics 164 ⢠-firmall <file> Updates device eeprom based on <file> ; Only MAC address will be preserved ⢠-pxe Programs PXE firmware from file ⢠-elog <file> Produces a log file with only error information ⢠-dids Shows PCI DID, VID, SDID, and SVID of present devices ⢠-pasf <file> o Programs ASF firmware o <file> is a text file that contains ASF bin file names ⢠-pxes <value> o Sets PXE speed o Value = 0 for auto; Value = 1 for 10HD; Value = 2 for 10FD; Value = 3 for 100H D; Value = 4 for 100FD; Value = 6 for 1000FD (fiber ) ⢠-hlb <spd> o Sets device host loopback mode o spd = 10/100/1000 o dflt = 1000 ⢠-tr <pkts> <spd> o Sets Tx and Rx pkts packets o dflt Tx pkts = 1e6 o spd = 10/100/1000 ⢠-lbsped <spd> o Runs D3 loopback test at <spd> o spd = t/h/g => 10/100/1000 ⢠-l200 <m200> Runs D4 loopback test with 200*m200 packets ⢠-dispgbend Runs D4 loopback test and displays packets at end ⢠-disppkts Runs D4loopback test and di splays packet count ⢠-errlimit <lim> Runs D4 loopback test. Fail afte r errors greater than <lim>. ⢠-pktf0 Runs D4 loopback test. Packet Data Alternating F 0. ⢠-rsttm <x> o <x> = 0 dis GRC rst. <x>!=0, post rst dly in ms
Diagnostics 165 o dflt = 1 ⢠-rstpl <x> <y> o <x> = post GRC rst poll interval, dly in ms; dflt = 10 o <y> = post GRC rst poll count; dflt = 0 (disable) ⢠-rstpld Above post reset polling counts display enable. ⢠-addref (x) Added a ref card info of dev <x> to refinfo.txt ⢠-blink <x> Blinks the LED on the selected device ⢠-rf <x> Selects a reference device for test D3 ⢠-sil Used before the -firm and -fir mall options to hide warnings ⢠-exe Executes diags commands wi thout a device present ⢠-updateasfcfg Used with the -firmall option to include updating ASF CFG block ⢠-updateumpcfg Used with the -firmall option to include updating UMP CFG block ⢠-pump <file> Option to program UMP firmware ⢠-updatesecfg Used with the -firmall option to include updating SE CFG block ⢠-mfct <file> Programs a new SSID and SVID based on <file> ⢠-fail2 On failure offsets failed message printo ut ⢠-logfl Log file buffer flush per line ⢠-nosz Do not write the NVRAM with the TPM/ NVRAM size Error codes and messages 1. Got 0xx @ 0xx. Expected 0xx 2. Cannot run test while chip is running 3. Invalid adapter device 4. Read only bit %s got changed afte r writing zero at offset 0x%X
Diagnostics 166 5. Read only bit %s got changed afte r writing 1's at offset 0x%X 6. Read/Write bit %s did not get cleared after writing zero at offset 0x %X 7. Read/Write bit %s did not get set after writing 1's at offset 0x %X 8. BIST failed. 9. Could not generate interrupt 10. Test aborted by user 11. Tx DMA:Got 0xx @ 0xx. Expected 0xx 12. Rx DMA:Got 0xx @ 0xx. Expected 0xx 13. Tx DMA failed 14. Rx DMA failed 15. Data error, got 0xX at 0xX, expected 0xX 16. Second read error, got 0xX at 0xX, expected 0xX 17. Failed writing EEPROM at 0xX 18. Failed reading EEPROM at 0xX 19. EEPROM data error, got 0x08X at 0x04X, expected 0xX 20. Cannot open file %s 21. Invalid CPU image file %s 22. Invalid CPU image size %d 23. Cannot allocate memory 24. Cannot reset CPU 25. Cannot release CPU 26. CPU test failed 27. Invalid Test Address Range Valid adapter address is 0xx-0xx and exclude 0xx-0xx 28. DMA:Got 0xx @ 0xx. Expected 0xx 29. Unsupported PhyId X:X 30. Too many registers specified in the file, max is %d 31. Cannot write to VPD memory 32. VPD data error, got X @ 0x04X, expected X 33. No good link! Check Loopback plug 34. Cannot TX Packet! 35. Requested to Tx %d. Only %d is transmitted 36. Expected %d packets. Only %d good packets are received %d unknown packets have been received %d bad packets have been received 37. %c%d is an invalid Test 38. EEPROM checksum error 39. Error in reading WOL/PXE 40. Error in writing WOL/PXE
Diagnostics 167 41. No external memory detected 42. DMA buffer X is large, size must be less than %0 4X 43. File size %d is too big, max is %d 44. Invalid %s 45. Failed writing 0x%x to 0x%x 46. *1 47. *1 48. *1 49. *1 50. Cannot perform task while chip is not running. (need driver) 51. Cannot open register define file or content is bad 52. ASF Reset bit did not self-cleared 53. ATTN_LOC %d cannot be mappe d to %cX CPU event bit %d 54. %s Register is not cleared to zero after reset 55. Cannot start Register Timer 56. poll_ASF bit did not get re set after acknowledged 57. Timestamp Counter is not counting 58. %s Timer is not working 59. Cannot clear bit %s in %cX CPU event register 60. Invalid "EEPROM_FILENAME" file size, expected %d but only can read %d bytes 61. Invalid magic value in %s, expected x but found x 62. Invalid manufacture revision, expected %c but found %c 63. Invalid Boot Code revision, expected %d.%d but found %d.%d 64. Cannot write to EEPROM 65. Cannot read from EEPROM 66. Invalid Checksum 67. Invalid Magic Value 68. Invalid MAC address, expected X-X-X -X-X-X 69. Slot error, expected an UUT to be found at location X:X:00 70. Adjacent memory has been corrupted while testing block 0xx-0xx Got 0xx @ address 0xx. Expected 0xx *1 Internal Use. Program will not generate this error. HP NC-Series Broadcom Multifunction adapter diagnostics Overview
Diagnostics 168 The diagnostic software lets you te st adapters to see if there are any problems with the adapter hardware, cabling, or network connections. HP recommends that you run diagnostics every time you in stall an adapter. Prerequisites The diagnostic utility runs under MS-DOS® v6.22 or later. It will not run from a DOS window in IBM OS/2® or any Microsoft Windows prod ucts. This util ity is designed to test hardware operation and confirm the adapter's ability to communicate with an other adapter in the same network, which primarily confirms cabling status. This diagostic utility i s not a throughput measurement tool. The utility can test the adapter wheth er or not there is a responder on the network; however, to co nduct a thorough test you should set up a second system on the network as a responder prior to starting a test. The DOS extender PMODE/W is embedded into the e xecutable. About the XDIAG.exe diagnostic tests The xdiag.exe diagnostic tests are di vid ed into four groups: Group A: Register tests; Group B: Memory tests; Group C: Block tests; an Group D: Miscellaneous tests. Group A: Register tests ⢠A1. Register test This tests the ch ip registers, accessible th rough PCI/PCI-E bus, for their read-only and read/write attributes. Some critical register s are not tested as the system and/or the chip becomes unstable when the values change. ⢠A2. PCI configuration test This test checks the functionality of the BAR size configuration by examining the BAR value as the BAR size varies. ⢠A3. Interrupt test This test checks to see if the system (OS) receive s the interrupt artificially generated by the chip and if the software ISR is properly invoked. ⢠A4. Not used ⢠A5. MSI test This test checks for the correct behavior of the MSI, making sure no interrupt is generated other than the message. It also runs the negative test to ma ke sure no message is generated when interrupt is masked off. ⢠A6. Memory BIST This tests all memory modules inside the chip using B uilt-In-Self-Test (BIST). ⢠A7. Network link test This tests the external link connection. For the fibe r medium, this is simply another external loopback test. For the copper medium, this is not applicable. Group B: Memory tests Various patterns (0x55aa55aa, 0xaa55a a55, & address) are used to te st each of the memory blocks. ⢠B1. TXP scratchpad ⢠B2. TPAT scratchpad
Diagnostics 169 ⢠B3. RXP scratchpad ⢠B4. COM scratchpad ⢠B5. CP scratchpad ⢠B6. MCP scratchpad ⢠B7. TAS header buffer ⢠B8. TAS payload buffer ⢠B9. RBUF via GRC ⢠B10. RBUF via indirect access ⢠B11. RBUF Cluster list ⢠B12. TSCH list ⢠B13. CSCH List ⢠B14. RV2P scratchpads ⢠B15. TBDC memory ⢠B16. RBDC memory ⢠B17. CTX page table ⢠B18. CTX memory Group C: Block tests ⢠C1. CPU logic and DMA interface tests The tests check the basic logic functionalities of ea ch of the on-chip CPUs. The tests also c over the DMA interface exposed to the CPUs. These tests r equire the presence of a test fi rmware file inside the âdiagfwâ directory. ⢠C2. RBUF allocation test This tests the Rx buffer allocation interface. ⢠C3. CAM access test This tests read, write, add, modify, and cache hit functionalities of the associative memory. ⢠C4. TPAT cracker test This tests the packet cracking logic block as well as the checksum/CRC offl oad logic. This test requires the presence of a test firmware file inside the â diagfwâ directory. ⢠C5. FIO register test This is another register test dedicated for register interface only exposed to the internal CPUs. This test requires the presence o f the test firmware files in the âdiagfwâ directory. ⢠C6. NVM access and reset-corruption tests This tests the non-volatile memory access (both re ad and write). It also tests for appropriate access arbitration among multiple entities (CPUs). Another test is to check to issue chip reset while NVM block is servicing data to look for any NVM data corruption. This test requires the presence of a test firmware file inside the âdiagfwâ directo ry. ⢠C7. Core-reset integrity test
Diagnostics 170 This test issues multiple chip resets and constant driver lo ad/unload to check for boot ROM appropriately loads the bootcode. This test requires a proper bootcode to be programmed into the non-volatile memory. ⢠C8. DMA engine test This tests the internal DMA engine by performing both DMA reads and writes at various location and various sizes. CRC check is performed to ensure data integrity. DMA write test also checks to ensure that the DMA writes do not corrupt the neighboring host memory. This test re quires the presence of test firmware files inside the âdiagfwâ direc tory. ⢠C9. VPD test This tests the VPD interface used by the bootcode . This test requires a proper bootcode to be programmed into the non-volatile memory. ⢠C10. Not used ⢠C11. FIO Events test This test checks for the event bits in the CPUâs Fa st IO interface, making sure appropriate bits are triggered when a particular event occurs (GP IO bit changes, NVM access, and so on). Group D: Miscellaneous tests ⢠D1. MAC loopback test This test puts the chip in the MAC loopback mode and transmits 5000 layer two packets of various sizes and receives them and checks the packet integrity. ⢠D2. PHY loopback test This test puts the chip in the PHY loopback mode and tr ansmits 5000 layer two packets of various sizes and receives them and checks the packet integrity. ⢠D3. External loopback test This test puts the chip in the PHY loopback mode and transmits various number of layer two packets of various sizes and receives th em and checks the packet integrity. The number of packets is determined by the speed. For 10Base-T, only 1000 packets are used; for 100Base-T, 5000; for gigabit traffic, 20000 packets are used. This test re quires an external loopback terminator to t he traffic to be returned, and is turned off by default. ⢠D4. LSO test This test checks the functionality of the large send offload by submit big TCP packets to the chip and expects the chip to segment them into multiple sm aller TCP packets (based on the MSS). The packets are returned in the MAC loopback mode and their integrity is check ed upon receive. ⢠D5. EMAC statistics test This test checks the basic statistics information maintained by the chip by sending and receiving packets of various sizes. ⢠D6. RPC test (D06) This test checks the receive catch-up path by sendin g packets to a different Tx chai n. The packets will go through the RPC logic and return to th e receive buffers as Rx packets. The integrity of each packet is checked to ensure no data corruption. Starting the XDIAG.exe tests
Diagnostics 171 1. If you have not already done so, download and extract the HP SoftPaq file and then run MAKEDISK.BAT to create a driver disk. 2. Boot to DOS or the EFI shell. 3. From the DOS prompt navigate to the \APPS\DIAGS\MFDIAG directory. 4. Type XDIAG and press the Enter key. The diagnostic tests r un automatically. Command line options The following are the available options in xdiag. They are case sensitive. Usage: xdiag [options] ⢠-t <grps/tests> Disables certain tests/grou ps (for example: a1c2) ⢠-T <grps/tests> Enables certain tests/groups (for example: a1c2) ⢠-c <card#> Selects device number on wh ich tests will be running ⢠-I <iteration#> Specifies how many iterat ions tests need to run ⢠-wol <1|0> Enables (1) or disables (0) magic pkt wol ⢠-pxe <1|0> Enables (1) or disables (0) pxe ⢠-asf <1|0> Enables (1) or disables (0) asf ⢠-cof Allow tests to continue test s on failure ⢠-ver Displays information on devices and xdiag version ⢠-log <logfile> Logs the tests execution into the specified file ⢠-help Prints out the screen
Technical support 172 T ec hni cal suppo r t In this section Before you contact HP.......................................................................................................... ................. 172 HP contact in format ion ......................................................................................................... ................. 172 Free automated custom er support services ....................................................................................... ........ 172 Before you contact HP Be sure to have the following info rmation available before you call HP: ⢠Technical support registrati on number (if applicable) ⢠Product serial number ⢠Product model name and number ⢠Applicable error messages ⢠Add-on boards or hardware ⢠Third-party hardware or software ⢠Operating system type and revision level HP contact information For the name of the nearest HP authorize d reseller: ⢠In the United States, see the HP US service locator webpage ( http://www.hp.com/service_l ocator ). ⢠In other locations, see the Contac t HP worldwide (in English) webpa ge ( http://welcome.hp.com/country/us/en/wwcontact.html ). For HP technical support: ⢠In the United States, for contact options see the Contact HP United States webpage ( http://welcome.hp.c om/country/us/en/contact_us. html ). To contact HP by phone: o Call 1-800-HP-INVENT (1-800-474 -6836). This service is available 24 hours a day, 7 days a week. For continuous quality improvemen t, calls may be recorded or monitored. o If you have purchased a Care Pack (servi ce upgrade), call 1-800-633-3600. For more information about Care Packs, refer to the HP website ( http://www.hp.com ). ⢠In other locations, see the Contac t HP worldwide (in English) webpa ge ( http://welcome.hp.com/country/us/en/wwcontact.html ). Free automated customer support services Free Automated Customer Support Services
Technical support 173 The following sites offer tro ubleshooting information, compatibility notes, and softw are upgrades (including Softpaqs and drivers). HP Worldwide Web Server ⢠Navigate to a specific product, and then look for support information from the list of support resources at the HP support websit e ( http://h18007.www1.hp.com/support/files/serv er ). ⢠For downloadable support software for HP Digital Networking Products, Hu bs, Integrated Access Devices, Modems and ISDN, Adap ters, Remote Access Concentrators/Servers, Software, and Switches, go to the HP software and drivers website ( http://h18007.www1.hp.com /support/files/server ). ⢠All SoftPaqs sorted by SoftPaq number can be found at the HP ftp support website ( ftp://ftp.compaq.com/pub/softpaq/ ). An ASCII version of a SoftPaq can be found by selecting a SoftPaq at the HP ftp support website. ( ftp://ftp.compaq.com/pub/softpaq/ ) ⢠An index of available software sorted by produc t can be found at the HP software and drivers website ( http://h18007.www1.hp.com/support/files/serv er ). HP FTP Server Navigate to a specific product, and then look for supp ort information from the list of support resources at the HP ftp support website ( ftp://ftp.compaq.com/pub/softpaq/ ).
Acronyms and abbreviations 174 A c r on y ms and abbr e v iati ons BIOS Basic Input/Output System DHCP Dynamic Host Configuration Protocol IPL initial program load iSCSI Internet Small Computer System Interface PCI peripheral component interface PDU power distribution unit PXE Preboot Execution Environment ROM read-only memory RSS Receive-Side Scaling SAN storage area network TCP/IP Transmission Control Protocol/Intern et Protocol TOE TCP/IP Offload Engine
Acronyms and abbreviations 175 UNDI Universal Network Driver Interface VLAN virtual local-area network
Index 176 A Adapter teaming 129, 139 Advanced Server Program for NetWare 110 B Boot Agent 69 D diagnostics 155, 157 Diver and software information 7 DXE configuration NC37xx and NC380x 67 F Firmware upgrade 48 H HP ProLiant Essentials value packs for networking 151 I installation 13, 15, 20, 21, 23, 28 K Keywords for B06 NDIS2 driver 63 Keywords for N100 NDIS2 driver 81 Keywords for N1000 NDIS2 driver 87 Keywords for Q57 NDIS2 driver 62 L Linux installation 20 N NCU properties 91 NetWare Client 32 driver installation for HP adapters 120 NetWare installation 28 NetWare teaming 139 P PCI Hot Plug 9 PCI shared interrupts 155 properties 91, 98, 106 PXE configuration HP NC-series Broadcom adapters 64, 67 HP NC-series Intel adapters 89 S Solaris driver information guide 23 supported operating system s 5 T teaming adapters 126 U Unix driver software Gigabit Ethernet Adapter user's guide 21 W Windows 2000 driver software NC150, NC32x, NC37x, NC380, NC1020, NC6770, and NC77xx Gigabit Ethernet Adapter user's guide 98 Windows 2000 installation notes 15 Windows Server 2003 NC-series Intel adapter properties 103 Windows Server 2003 and Windows Server 2003 x64 installation notes 13 Windows Server 2003 NC31xx Fast Eth ernet Adapter user's guide 106 Windows Server 2003 NC -s eries Broadcom adapter properties 98 Inde x
© Copyright 2005-2007 Hewlett-Packard Development Co mpany, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Noth ing herein should be construed as constituting an a dditional w arranty. HP shall not be liable for technical or editor ial errors or omissions contained herein. Confidential computer software. Va lid license from HP required for possession, use or copying. Consistent with FAR 12.211 and 12.212 , Commercial Computer Software, Co mputer Software Documentation, and Technical Da ta for Commercial Items are li censed to the U.S. Government under vendorâs standard commercial license. Microsoft, Windows, and Windows NT are U.S. re gistered trademarks of Microsoft Corporation. Windows Server 2003 is a trademark of Microsoft Corporation. Intel, Pentium, and Itanium are trademarks or registered trademarks of Intel Corporatio n or its subsidiaries in the United States and other countries. UNIX is a registered trademark of The Open Group. Audience assumptions This document is for the person who insta lls, admini sters, and troubleshoots servers and storage syst ems. HP assumes you are qualified in the serv icing of co mputer equipment and trained in recognizing hazards in products with hazardous energy levels.
Contents 3 C on t en ts Overvi ew ..................................................................................................................................... 5 Supported adapters and operating systems....................................................................................... ............ 5 Adapter driver and so ftware info rmation ...................................................................................................... 7 PCI Hot Plug ................................................................................................................... .......................... 9 Duplex support................................................................................................................. ....................... 10 Installa tion ................................................................................................................... .............. 13 Download latest drivers, fi rmware, and do cumentation ........................................................................... ..... 13 Windows Server 2003 and Wi ndows Server 2003 x64 .............................................................................. 13 Windows 2000 an d Windows NT .................................................................................................... ........ 15 Linux.......................................................................................................................... ............................ 20 SCO OpenServer and Unix Ware .................................................................................................... .......... 21 Solaris ........................................................................................................................ ........................... 23 NetWare ........................................................................................................................ ....................... 28 Firmware upgr ade util ity ....................................................................................................... ....... 48 Windows ut ility ................................................................................................................ ....................... 48 c-Class BladeSys tem utility .................................................................................................... .................... 54 Linux ut ility .................................................................................................................. ........................... 54 DOS util ity .................................................................................................................... .......................... 59 Adapter conf igurat ions ......................................................................................................... ....... 62 HP NC-Series Broa dcom adapters ................................................................................................. ............ 62 Keywords for Q5 7 NDIS2 driver .................................................................................................. ... 62 Keywords for B 06 NDIS2 driver .................................................................................................. .... 63 PXE configuration: HP NC-series Broa dcom and Multifun ction adapters ................................................ 64 HP NC-Series Multif unction a dapters............................................................................................ .............. 66 Accelerated iS CSI devices ...................................................................................................... ........ 66 iSCSI boot configuration: Multifunctio n adapters ............................................................................... 67 PXE configuration: HP NC-series Broa dcom and Multifun ction adapters ................................................ 67 HP NC-Series In tel adapters .................................................................................................... .................. 69 Boot Agent ..................................................................................................................... .............. 69 Keywords for N100 NDIS2 driver ................................................................................................. .. 81 Keywords for N10 00 NDIS2 driver ................................................................................................ .8 7 PXE configuration for HP NC-Series Inte l adapters.............................................................................. 89 Windows ........................................................................................................................ ....................... 91 Modifying adapter proper ties using the NCU .................................................................................... 9 1 Modifying HP NC-Series Broadcom and Multifunction adapt er properties ............................................. 98 Modifying HP NC-Series Inte l adapter prop erties ............................................................................. 103 Modifying NC31xx Fast Ethern et adapter pr operties ........................................................................ 106 Modifying adapter properti es in Wind ows NT ................................................................................ 107 Novell Netware ................................................................................................................. ................... 110 Advanced Server Prog ram for NetW are ......................................................................................... 110 NetWare Client 32 driv er installation fo r HP adapters...................................................................... 120 Solaris ................................................................................................................................................. 122 Team prop erties ................................................................................................................ ........ 126 Adapter te aming ................................................................................................................................... 126
Contents 4 Multifunction gigabit se rver adapter teaming ................................................................................... ......... 129 Windows 2000 and Wi ndows NT4 teaming ........................................................................................... .1 3 0 NetWare te aming ................................................................................................................................. 139 VLAN prop erties ................................................................................................................ ....... 150 VLAN info rmation .................................................................................................................................. 150 ProLiant E ssentia ls ............................................................................................................ ......... 151 Networking pack installation ................................................................................................... ............... 151 Intelligent Networking Pack-Windows Edition.................................................................................... ........ 151 Intelligent Networking Pack-Linux Edition ...................................................................................... ............ 152 Accelerated iSCSI for stand-alon e and c-Class BladeSystem Multif unction server adapters .............................. 153 Accelerated iSCSI for embedded HP Multifunction serv er adapters .............................................................. 15 3 Diagnost ics .................................................................................................................... .......... 154 PCI shared interrupts .............................................................................................................................. 154 HP NC-Series Intel ad apter diagnostics ......................................................................................... ........... 154 HP NC-Series Broadcom adapter diagnostics ...................................................................................... ...... 156 HP NC-Series Broadcom Multif unction adapter diagnostics ........................................................................ .1 6 7 Technical support .............................................................................................................. ........ 172 Before you contac t HP ............................................................................................................................ 172 HP contact in format ion ........................................................................................................................... 172 Free automated customer support services ....................................................................................... ......... 172 Acronyms and abbreviati ons ..................................................................................................... . 174 Index .......................................................................................................................... ............. 176
Overview 5 Ov er v ie w In this section Supported adapters and operating systems ....................................................................................... .......... 5 Adapter driver and so ftware info rmation ........................................................................................ ............. 7 PCI Hot Plug ................................................................................................................... ......................... 9 Duplex support................................................................................................................. ...................... 10 Supported adapters and operating systems Supported adapters The Network Configuration Utility (NCU) supports th e following HP Fast Ethernet and Gigabit Server Adapters. Standup adapters: ⢠HP NC110T PCI Express Gigabit Server Adapter ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC310F PCI-X Gigabit Server Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC340T PCI-X Quad Port Gigabit Server Adapter ⢠HP NC360T PCI Express Dual Port Gigabit Server Ad apter ⢠HP NC364T PCI Express Quad Port Gigabit Server Adapter ⢠HP NC370T PCI-X Multifun ction Gigabit Server Adapter ⢠HP NC370F PCI-X Multifun ctio n Gigabit Server Adapter ⢠HP NC373F PCI Express Multifunction Gigabit Server Adapter ⢠HP NC373T PCI Express Multifunction Gigabit Serv er Adapter ⢠HP NC380T PCI Express Dual Port Multifunction Giga bit Server Adapter ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC3123 Fast Ethernet Server Adapter * ⢠HP NC3132 Fast Ethernet Upgrade Module (D ual 10/100) for the NC3134 Server Adapter * ⢠HP NC3133 100FX Upgrade Module f or the NC3134 Server Adapter * ⢠HP NC3134 Fast Ethernet Server Adapter (Dual Ba se 10/100) * ⢠HP NC3135 Fast Ethernet Upgrade Module (Dual 10 /100) for the NC3134 Server Adapter * ⢠HP NC6132 1000 SX Gigabit Upgrade Module f or the NC3134 Server Adapter* ⢠HP NC6133 1000 LX Gigabit Upgrade Modu le for the NC3134 Server Adapter * ⢠HP NC6134 Gigabit Server Adapter (PCI 1000 SX)
Overview 6 ⢠HP NC6136 Gigabit Server Adapter (PCI 1000 SX) ⢠HP NC6170 Dual Port PCI- X Gigabit Server Adapter ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7131 Gigabit Ethernet Server Adapter ⢠HP NC7132 Gigabit Ethernet Upgrade Module for the NC31 34 Server Adapter * ⢠HP NC7170 Dual Port PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC320i PCI Express Gigabit Server Adapter ⢠HP NC324i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC325i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC326i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC370i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC371i PCI-X Multifunction Gigabit S erver Adapter ⢠HP Embedded NC373i Multifunct ion Gigabit Server Adapter ⢠HP NC3161 Fast Ethernet Server Adapter * ⢠HP NC3163 Fast Ethernet Server Adapter * ⢠HP NC7760 PCI-X Gigabit Server Adapter ⢠HP NC7761 PCI-X Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gigabit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter Mezzanine adapters: ⢠HP NC320m PCIe Gigabit Server Adapter ⢠HP NC325m PCI Express Quad Port 1Gb Server Ad apter for c-Class BladeSystem ⢠HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem ⢠HP NC373m PCI Express Dual Port Multifunction Giga bit Server Adapter for c-Class BladeSyst em ⢠HP NC374m PCI Express Dual Port Multifunction Giga bit Server Adapter * Not supported in Windows Server 2003 x64 Supported operating systems These server adapters are supported on the followin g operating systems. Microsoft ⢠Windows Server 2003 x64 Editions ⢠Windows Server 2003
Overview 7 ⢠Windows® 2000 (NDIS5) ⢠Windows NT® 4.0 (NDIS4) (only supported on adapters that shipped befor e January 1, 2003) ⢠DOS NDIS2* NetWare ⢠NetWare 6.5 Server ⢠NetWare 6.6 Server ⢠NetWare Open Enterprise Server Linux 32 ⢠Red Hat Enterprise Linux 5 ⢠Red Hat Enterprise Linux 4 ⢠Red Hat Enterprise Linux 3 ⢠SUSE Linux Enterprise Server 10 Service ⢠SUSE Linux Enterprise Server 9 Service Linux 64 ⢠Red Hat Enterprise Linux 5 for AMD64 and Intel EM64T ⢠Red Hat Enterprise Linux 4 for AMD64 and Intel EM64T ⢠Red Hat Enterprise Linux 3 for AMD 64 and Intel EM64T ⢠SUSE Linux Enterprise Server 10 for AMD64 and Intel EM64T ⢠SUSE Linux Enterprise Server 9 for AMD64 and Intel EM64T UNIX ⢠Unixware 7.x ⢠OpenUnix 8.x ⢠Open Server 5.x/6.x Solaris ⢠Solaris 10 ⢠Solaris 9 * Unattended Install Only. Also, DOS ND IS2 is not supported for NC3133 adapters. Adapter driver and software information To obtain driver and software information for Micro soft® Windows Server 2003 x64, Microsoft® Windows Server 2003, Microsoft ® Windows® 2000, Microsoft® Windows NT® (4.0), Novell NetWare, SCO OpenServer, UnixWare, Linux, and Solaris: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system.
Overview 8 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. Drivers Drivers are found in the following folder s: ⢠Microsoft® Windows Ser ver 2003 \WIN2003 ⢠Microsoft® Windows Ser ver 2003 x64 \WIN2003x64 ⢠Microsoft® Windows® 2 000 \WIN2000 ⢠Microsoft® Windows NT® 4.0 \NT4 ⢠Novell NetWare \NETWARE\SERVER ⢠Linux \LINUX ⢠UnixWare 7.x and OpenServer 6.x \UNIX\UNIXWARE7_OPENSERVER6 ⢠Solaris 9 and Solaris 10 \SOLARIS ⢠Unattended Install \DOS Supplemental applications in \APPS ⢠Boot agent ⢠Diagnostic tests including MFDIAG.exe, N100X.ex e and Q57.exe ⢠Online firmware upgrade utility ⢠Multifunction firmware upgrade utility ⢠c-Class BladeSystem firmware upgrade utility ⢠Linux firmware upgrade utility Device-specific information See \docs\files.txt for a descriptive li st of all files in the SoftPaq. Release notes and other documentation in \docs ⢠HP Accelerated iSCSI for Multifunct ion Network Adapters (AiSCSIUG.pdf) ⢠HP Network Adapter License Util ity for Windows (NALICNSE.pdf) ⢠HP Network Adapter Scripting Utility (NICSCRPT.pdf) ⢠HP Network Configuration Agent (ncuasr.pdf) ⢠HP ProLiant Essentials Intelligent Networking PackâLinux Edition (PEINPL.pdf) ⢠HP ProLiant Essentials Intelligent netw orking PackâWindows Edition (PEINPW.pdf) ⢠HP Virus Throttle for Windows License Utility (VTLICNS E.pdf) ⢠HP NC510x PCIe 10 Gb Software Configuration Guide (NC5 10x PCIe 10 Gb Software Guide.pdf) ⢠Network adapter user guides (\docs\hw) ⢠Release notes (Relnotes.txt) ⢠SetLACState Utility (SetLAC.pdf)
Overview 9 Linux documentation ⢠Linux readme.txt file is located in the \LINUX directory ⢠HP iSCSI Boot User Guide for Linux in the \docs directory PCI Hot Plug PCI Hot Plug is an industry-standard solution that provides increased system availability and non-stop serviceability in business-cri tical computing environm ents. PCI Hot Plug enables PCI networking a dapters to be replaced or added to a server without taking the system down. HP pioneered this technology and has worked closely with the PCI Special Interest Group (SIG) and several key industry partners to accelerate its adoption as an industry standard. PC I Hot Plug technology also has a broad range of industry support from leading operating system suppliers incl uding Microsoft® a nd Novell. HP server adapters support PCI Hot Plug technology. To take advantage of this solution, you must use the adapter in an HP or third-party system that suppor ts PCI Hot Plug. To implement PCI Hot Plug, use the configuration instructions that come with your sy stem. If you use an HP system that supports PCI Hot Plug, refer to the documentation for configuring and insta lling PCI Hot Plug devices in the HP SmartStart application, which ships with HP systems. If you use a third-party system that supports PCI Hot Plug, refer to the documentat ion that came with your system or contact technical support for the vendor. Support for enabling PCI Hot Plug Different operating syst ems support PCI Hot Plug in diff erent ways. The following list describ es how different operating systems enable PCI Hot Plug devices. ⢠Windows® 2000, Windows Server 2003, and Windows Serv er 2003 x64 o Can replace like item for like item o Can add a new adapter to an empty slot o Can add an option module to an existing NC3134 adapter ⢠Windows NT® 4.0 o Can replace like item for like item o Cannot add a new adapter to an empty slot ⢠NetWare 6.x o Can replace like item for like item o Can remove an adapter o Can add a new adapter to an empty slot o Can add an option module to an existing NC3134 adapter (This feature is not supported on an NC3134 adapter that has a pers istent driver installed.) ⢠UnixWare 7.x and OpenServer 6.x Requires hardware that implement s PCI Hot Plug fa ilover and a DDI 8 MDI driver that implements the CFG_SUSPEND and CFG_RESUME subfunctions to th e entry point routine. This lets the system administrator suspend the driver, power down the slot, hot-swap or hot-replace the adapter, power up the slot, and resume the driver, all with out bringing the machine down. For further information about PCI Hot Plug support, refer to your UNIX documentation. ⢠SCO OpenServer 5.x
Overview 10 No support ⢠Linux No support Upgrading the adapter with a persistent driver To upgrade the adapter with a persistent driver: 1. Manually remove (unload) the driver. 2. Power down the slot and remove the adapter. 3. Upgrade the adapter (add the option module). 4. Reinstall the upgraded adapter. 5. Power up the slot to let NetWare automa tically detect the (upgrad ed) adapter. Duplex support Duplexing is a performance option that allows you to choose how the adapter sends and receives packets over the network. Duplex options include : ⢠Auto. (Auto-negotiation) The adapter negotiates with the full-duple x switch to send and receive packets at the highest rate. The switch must suppor t auto-negotiation. If the switch does not support auto-negotiation, the adapter defaults to half-duplex. NOTE: An exception is the HP NC3133 adap ter, which defaults to full-duplex. ⢠Full-duplex. (Requires f ull-duplex switch) The adapter se nds and receives packets simultaneously by using a different wire to perform each function. This feature can improve adapter performance by eliminating collisions. Set duplex mod e to full-duplex ONLY if you have a hub/switch that supports full-duplex. ⢠Half-duplex. The adapter performs one operation at a ti me (because send and receive lines are shared); it either sends or receives. Configuring for full-duplex If you have an auto-negotiating switch, the adapter au tomatically matches the duplex setting at the switch. If you do not have an auto-negotia ting switch, you must manually configur e the adapter for full-duplex support. Configuration is specific to the driver yo u are loading for your network operating system (NOS). To set up the duplex mode, refer to the section below that corresponds to your operating system. Note th at performance may suffer if your switch is not full-dup lex and you configure the adapter to full-duplex. Leave the adapter on half-duplex if you are not sure wh at type of switch to which you are connected. Windows NT® 4.0 To set the duplex options: 1. Double-click the HP Network icon in the Control Panel. Or, right-click on Network Neighborho od and select Properties. Then select the Adapters tab. Double-click on an adapter or team to launch the CPQNTAC utility. 2. Select an adapter from the list and click Properties.
Overview 11 3. Set Speed/Duplex on the adapter to Auto/Auto, 10/Half, 10/Full, 1 00/Half, 100/Full, or 1000/Full. 4. Click OK until prompted to restart. 5. Restart Windows NT® 4.0. Windows® 2000, Windows Server 2 003, and Windows Server 2003 x64 For new installations, Windows® 20 00 automatically installs the adapter drivers. Exception: For the NC150T, NC310F, NC32xx, NC37xx, NC340x, NC1020, NC61 36, NC6170, NC71xx and NC77xx adapters, you must install the appropriat e driver components located in the \WIN2000 directory in the HP SoftPaq. Complete the fo llowing steps to download the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. NOTE: If you need to update existing Windows® 2000 drivers, refer to the instructions in Windows 2000 installation notes (" Windows 2000 and Windows NT " on page 15 ). To set the duplex options: 1. Launch the Network Configuration Utility. 2. Select an adapter from the list in the NCU Main window and click Properties . The Adapter Properties window appears. 3. On the Settings tab, set the Speed/Duplex Setting for the adapter to Auto/Auto, 10/Half, 10/Full, 100/Half, 100/Full, or 1000/Full . 4. Click OK twice to accept the configuration changes. NetWare Server To set the duplex options, use the following syntax for FORCEDUPLE X and SPEED. NOTE: The NC340T, NC1020, NC71xx, and NC77xx adapters support full-duplex and half- duplex. The fiber Gigabit adapters support fu ll-duplex mode only and cannot be changed. ⢠Syntax: FORCEDUPLEX=n Where n = 0âauto-negotiation (You must have an auto-negot iating switch/hub (an Nway* swit ch) to get full- duplex support with the FORCEDUPLEX par ameter set to 0 (auto-negotiation). 1âhalf-duplex 2âfull-duplex NOTE: To set the adapter to half- or full-duplex, yo u must set the SPEED parameter to either 10 or 100. Default = auto-n egotiate
Overview 12 Examples: o 100 Mbps full-duplex: FORCEDUPLEX=2 SPEED=100 o 10 Mbps full-duplex: FORC EDUPLEX=2 SPEED=10 ⢠Syntax: SPEED= n (n = 10 or 100 or 1000) Specifies the speed the driver uses. If you do not use this parameter, the driver aut omatically detects the network speed. If unable to detect the network sp eed, the driver defaults to 10 Mbps. If you use this parameter, the driver operates at the specified speed instead of auto-d etecting network speed. NOTE: The NC3133 adapter supports 100 Mbps on ly. It does not default to 10 Mbps. You must set the SPEED parameter to either 10 or 100 if you want to set the FORCEDUPLEX parameter to either half or full. The SPEED parameter is for N100.LAN driver and N1000.LAN. You cannot change the speed for the Gigabit (fiber) ad apters. You can, however, change the speed for the NC7131 and NC7132 Gigabit (copper) adapters to 10 Mbps or 100 Mbps. Default = The adapter automatically senses speed. Other NDIS 2.x-based clients To set the duplex options in other NDIS 2.x-base d clients, edit the PROTOCOL.INI file. Add the FORCEDUPLEX and SPEED parameters to the N100 sect ion, or the SPEEDUPLEX parameter to the N1000 section. For information on these parameters, refer to K eywords for N100 NDIS2 driver (on page 81 ) , Keywords for N1000 NDIS2 driver (on p age 87 ), or Keywords for Q57 NDIS2 driver (on page 62 ).
Installation 13 Inst allati on In this section Download latest drivers, fi rmware, and documentation ........................................................................... .... 13 Windows Server 2003 and Wi ndows Server 2003 x64 ............................................................................ 13 Windows 2000 an d Windows NT .................................................................................................... ....... 15 Linux .......................................................................................................................... ........................... 20 SCO OpenServer and UnixWare .................................................................................................... ......... 21 Solaris ........................................................................................................................ .......................... 23 NetWare ........................................................................................................................ ...................... 28 Download latest drivers, firmware, and documentation HP updates networ king softwar e freq uently to include new functionality and features. For the latest driver, firmware, and documentation updates go to the HP website ( http://h18004.www1.hp.com/products/ servers/networking/ index.html ). Windows Server 2003 and Windows Server 2003 x64 The following provides inform ation about Windows Server 200 3 and Windows Server 2003 x64 support for HP NC-series network adapters. Windows Server 2003 The following Windows Server 2003 network adap ter drivers are used in this release: ⢠N100325.sys for NC31xx Fast Ethernet server adapters ⢠N1000325.sys for NC310xx, NC340x, NC61xx, an d NC71xx server adapters ⢠Q57XP32.sys for NC150x, NC32xx, NC10xx, NC 67xx, and NC77xx server adapters ⢠N1E5132.sys for NC110T, NC360x, NC364x server adapters ⢠BXND51X.sys for all Multifunction server adapters ⢠NXP2NIC.sys for NC510x 10 Gb server adapters Components for these drivers are located in the \W IN2003 directory in the HP SoftPaq. Complete the following steps to downloa d the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system.
Installation 14 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. Windows Server 2003 x64 IMPORTANT: NC31xx, NC6132, NC6133, and NC7132 adapters are not supported in Windows Server 2003 x64 . The following Windows Serv er 2003 x64 network adapter drivers are used in this relea se: ⢠N1G5132E.sys for NC310F, NC340T, NC61xx, and NC71xx server adapters ⢠Q57AMD64.s ys for NC150T, NC32x, NC1020, NC6 7xx, and NC77xx server adapters ⢠BXVBDA.sys for Multifunction server adapters ⢠N1E5132E.sy s for NC110T, NC360T, NC364x server adapters ⢠NXP2NIC.sys for NC510x 10 Gb server adapters Components for these drivers are located in the \WIN2003x 64 directory in the HP SoftPaq. Complete the following steps to do wnload the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. Installing network drivers (new installations) When you install the adapter in the server for the first ti me: 1. Connect the cable from the adapter to your network. 2. Provide power to your server. 3. Start Windows Serve r 2003 or Windows Se rver 2003 x64 . o For the NC150T, NC310F, NC32xx, NC340x, NC3 7xx, NC380x, NC1020, NC6136, NC6170, NC6770, NC71xx, and NC77xx server ad apters, you must install the appropriate driver component located in the \WIN2003 or \WIN2003x 64 directory in the HP SoftPaq file. o For all other adapters, Windows® automatically de tects and installs a driver for the adapter from its own library of drivers. Updating network drivers To update the network driver, follow these instructions: Locate the driver on the CD, diskette, or directory on your hard driv e where you copied it from an updated source. See Technical Support for a list of online sources for software and driver updates.
Installation 15 1. Run the appropriate Driver Component Package (lis ted as CPxxxxxx.exe) in a directory on your hard drive. The component is a self-extracting execut able with a file name based on the component number. 2. Click the Install button to proceed with the installation. 3. After the drivers have been installed, you may delete the Component Package that you copied. Removing the driver software To remove the driver software: 1. Before physically removing an adapter from your system, first remove the adapter driver software. 2. Start Windows Server 2003 or Windows Server 2003 x64 and log in. You must have Netwo rk Administrator privileges to re move the driver software. 3. Open the Control Panel and double-click the System icon. In the System Properties dialog box, select the Hardware tab. 4. Click Device Manager . The Device Manager window appears. 5. Click the plus ( ) sign to expand the Network adapter s. All network adapters appear. 6. Right-click the adapter to be removed and choose Uninstall . NOTE: Not all driver files are removed as part of this procedure. The driver and adapter can be removed via Hot Plug, if supported. Windows 2000 and Windows NT The following provides inform ation about Windows 2000 support for HP NC-series network adapters. The following Windows 2000 netw ork adapter drivers are used in this release: ⢠N100NT5.sys for NC31xx Fast Ethernet serv er adapters ⢠N1000NT5.sys for NC310x, NC340x, NC61xx, and NC71xx Gigabit Ethernet server adapters ⢠Q57W2K.sys for NC150x, NC32xx, NC10xx, NC67xx, and NC77xx Gigabit Ethernet server adapters ⢠BXVBDX.sys for Multifunction Gigabit server adapters ⢠N1E5032.sys for NC360x and NC364T Gigabit Ether net server adapters Components for these drivers are located in the \WIN2000 dir ectory in the HP SoftPaq file. Installing network drivers (new installations) When you install the adapter in the server for the first ti me: 1. Connect the cable from the adapter to your network. 2. Provide power to your server. 3. Start Windows Serve r 2003 or Windows Se rver 2003 x64 . o For the NC150T, NC310F, NC32xx, NC340x, NC3 7xx, NC380x, NC1020, NC6136, NC6170, NC6770, NC71xx, and NC77xx server ad apters, you must install the appropriate driver component located in the \WIN2003 or \WIN2003x 64 directory in the HP SoftPaq file. o For all other adapters, Windows® automatically de tects and installs a driver for the adapter from its own library of drivers. Updating network drivers from disk
Installation 16 To update the network driver, follow these instructions: Locate the driver on the CD, diskette, or directory on your hard driv e where you copied it from an updated source. See Technical Support for a list of online sources for software and driver updates. 1. Run the appropriate Driver Component Package (lis ted as CPxxxxxx.exe) in a directory on your hard drive. The component is a self-extracting execut able with a file name based on the component number. 2. Click the Install button to proceed with the installation. 3. After the drivers have been installed, you may delete the Component Package that you copied. Windows NT4 unattended install (push) The following provides inform ation about Windows NT support for HP NC-series network adapters. Overview The primary objective of the Push Installation is to get the client properly connected to the Windows NT® server before the installation begi ns using the NDIS2 driver, and to keep the same client connected using the NDIS4 driver after the Windows NT® server is installed without any end-user action. The difficulty is that the list of supported adapter s on the Windows NT® 4.0 CD-ROM is r estricted and special steps must be performed to update the installati on source files and modify the scripts to automate the process for the adapters not on the list. Main steps of the process 1. Create a Network Installation Startup disk from the Windows NT® server and modify it for your adapter. 2. Create a shared directory for the Windows NT® server installation files on the Wi ndows NT® server. 3. Modify the UNATTEND.TXT installation script to automate the procedure and add the proper adapter installation and configuration information. The sample UNATTEND.TXT file, provided in Item 4 of the "Samples and Templates" section, indicates how to override default registry paramet ers for the adapter in the [CPQN ic1ParamSection] section. The override parameters sho uld not be used unless the def ault parameters will not work properly for a specific installation. Th e sample parameters are commented out. 4. Integrate the adapter driver files into th e Windows NT® installa tion source files. 5. Perform the Unattended Installation bootup with th e prepared startup disk from the client system. Performing an unattended installation on Windows NT 4.0 1. Prepare a Windows NT® 4.0 server: a. Create a folder on the server hard disk for the installation files (for example, NT40.INS). Make this folder shared (for example, with shar ed name NT40.INS). b. Create a user (for example, USERID with a password password) and grant read and view rights for the NT40.INS folder. 2. Use a CD-ROM drive at the Windows NT® server an d copy the i386 folder with Windows NT® 4.0 installation files to the shared folder NT40.INS on this server (for ex ample, using NT Explorer). 3. Create or modify the installation answer file UNA TTEND.TXT as required by your specific syst em and desired Windows NT® config urat ions. UNATTEND.TXT may be re ad-only and may need to be made write-able before editing (for example, a ttrib -r unattend.txt). Refer to the sample UNATTEND.TXT file with comments. 4. Create the \CPQANC direc tory: NT40.INS\i386\DRVLIB.NIC\CPQANC.
Installation 17 5. Copy driver installation files: Xcopy the HP Configuration and Drivers Disk to the proper location under Adapter Device Driver folder (NT40. INS\i386\DRVLIB.NIC\CPQANC fol der on the Windows NT® server). This enables the installati on program to find the same file structure as on the Drivers disk. Be sure to use th e /s /e switches on xcopy. 6. On the Windows NT® server, prepare the Microso ft® Client Installation disk using the Windows NT® Network Client Administrator: a. Prepare a system diskette (for example, DOS 6. 22: format a: /s). Do not use the Windows NT® system disk. b. Copy UNATTEND.TXT from \NT40.INS\i386 to th e diskette as specified in item 3 above (preferably to the root (a:\)). The best location to pu t this file is in the root of your install disk, but the only requirement is to specify the full path to this file in the AUTOEXEC.BAT for the WINNT command. c. Use the Windows NT® Administrat ive Tools (Co mmon), start Network Client Administrator and select Make Network Installation Startup Disk. d. Set "Existing Path" to use the previously install ed software for MS Cl ient or, if it is the first time, create the shared directory on the server' s ha rd drive, using the Windows NT® 4.0 Server CD- ROM as a source (for example, by copying from \client on the CD-ROM to c:\client). e. Choose Network Client v3.0 fo r MS-DOS and Windows. f. Choo se any of the Network Adapters from the list (for example, NE2000 compatible). g. Set Computer name, User name, Domain and Network protocol when prompted. h. Select OK and wait while files are copied to the disk. i. Copy the proper ndis2 dos driver (N100.DOS, N1000.DOS, or Q57.DOS) to this disk in the \NET directory. N100.DOS, N1000.DOS, Q57.DO S can be found in \DOS\NDIS2\ in the HP SoftPaq file. j. Modify A:\NET\SYSTEM.INI : netcard=N100.DOS (or N1000.DOS or Q57.DOS). k. Modify A:\NET\PROTOCOL .INI : drivername =N100$ (or N1 000$ or Q57$ ). l. Add a line: CACHEFLUSH=1 just after the "drivername = N100 $" line. m. Verify or modify A:\AUTOEXEC.BAT to have the following commands: NET USE W: \\ServerName\NT40.INS W:\WINNT /s:W:\ /u:a:\unattend.txt Note that W is an example for the logical drive mapped to the share on the Windows NT® server. n. Remove the invocation of setu p.exe from A:\AUTOEXEC.BAT. 7. Insert the startup disk and boot up your DOS clie nt. After connecting to the network. do one of the following: o Input USERID as a User name when prompted. Input password as a password when prompted. o Modify the NET START line in the AUTOEXEC .BAT to NET LOGON USERID PASSWORD /YES. Samples and templates ⢠PROTOCOL .INI [network.setup] version=0x3110 netcard=ms$ne2clone,1,MS$NE2CLONE,1 ;transport=ms$ndishlp,MS$NDISHLP
Installation 18 ;transport=ms$netbeui,MS$NETBEUI ;lana0=ms$ne2clone,1,ms$netbeui ;lana1=ms$ne2clone,1,ms$ndishlp transport=tcpip,TCPIP lana0=ms$ne2clone,1,tcpip [ms$ne2clone] drivername=N100$ CACHEFLUSH=1 ; CACHEFLUSH line is required ONLY IF NDIS2 does not shut down ; properly ;INTERRUPT=3 ;IOBASE=0x300 ;SlotNumber=1 [protman] drivername=PROTMAN$ PRIORITY=MS$NDISHLP ;[MS$NDISHLP] ;drivername=ndishlp$ ;BINDINGS=ms$ne2clone ;[ms$netbeui] ;drivername=netbeui$ ;SESSIONS=10 ;NCBS=12 ;BINDINGS=ms$ne2clone ;LANABASE=0 [tcpip] NBSessions=6 DefaultGateway0= SubNetMask0=255 255 0 0 IPAddress0=137 65 10 12 DisableDHCP=1 DriverName=TCPIP$ BINDINGS=ms$ne2clone LANABASE=0 ⢠SYSTEM.INI [network] filesharing=no printsharing=no autologon=yes computername=COMPUTERNAME lanroot=A:\NET username=Administrator workgroup=DOMAIN reconnect=no directhost=no dospophotkey=N lmlogon=0 logondomain=DOMAIN preferredredir=full autostart=full maxconnections=8
Installation 19 [network drivers] netcard=N100.DOS ;transport=ndishlp.sys,*netbeui transport=tcpdrv.dos,nemm.dos devdir=A:\NET LoadRMDrivers=yes [Password Lists] *Shares=a:\net\Share000.PWL USERID=A:\NET\USERID.PWL ⢠AUTOEXEC.BAT path=a:\net a:\net\net start rem a:\net\net logon userid password /yes net use W: \\ServerName\nt40.ins W:winnt /s:W:\ /u:a:\unattend.txt ⢠UNATTEND.TXT [Unattended] OemPreinstall = yes NoWaitAfterTextMode = 1 NoWaitAfterGUIMode = 1 FileSystem = LeaveAlone ExtendOEMPartition = 0 ConfirmHardware = no NtUpgrade = no Win31Upgrade = no TargetPath = winnt OverwriteOemFilesOnUpgrade = no OemSkipEULA = yes [GuiUnattended] OemSkipWelcome = 1 OEMBlankAdminPassword = 1 TimeZone = "(GMT-08:00) Pacific Time (US & Canada)" [UserData] FullName = "Userid" OrgName = "Organization" ComputerName = COMPUTER ProductId = "xxx-xxxxxx" [Display] ConfigureAtLogon = 0 BitsPerPel = 8 XResolution = 640 YResolution = 480 VRefresh = 60 AutoConfirm = 1 [Network] InstallAdapters = CPQAdapters InstallProtocols = ProtocolsSection
Installation 20 InstallServices = ServicesSection JoinWorkgroup = workgroup [ProtocolsSection] TC = TCParamSection [TCParamSection] DHCP = yes [CPQAdapters] CPQANC.1 = CPQNic1ParamSection, \i386\drvlib.nic\CPQANC ;CPQANC.2 = CPQNic2ParamSection, \i386\drvlib.nic\CPQANC ;CPQANC.3 = CPQNic3ParamSection, \i386\drvlib.nic\CPQANC ;CPQANC.4 = CPQNic4ParamSection, \i386\drvlib.nic\CPQANC [CPQNic1ParamSection] ;NetAddress = "" ;SpeedDuplex = 0 ; 0 means autonegotiation ;MaxFrameSize = 0 ; 0 is use card default (1514) [ServicesSection] References ⢠Microsoft® Window s NT® Resource KitâMic rosoft® Press 19NT ⢠Microsoft® TechNet CD-ROM, April 1997 ⢠Microsoft® support web page at http://support .microsoft.com/support Article ID : Q155197, Q156795 Linux The following Linux drivers are available for HP adapters: ⢠tg3âSupports NC150x, NC320x, NC324x, NC32 5x, NC326x, NC1020, NC67xx, and NC77xx server adapters ⢠e1000âSupports NC1 10T, NC310F, NC340x, NC 360x, NC364T, NC61xx, and NC71xx server adapters ⢠hp-pelâSupports ProLiant Essentials Licen sing package ⢠hp-vtâSupports Virus Throttle for NCxxxx adapters ⢠iSCSI-bootâSupports the HP iSCSI boot f eature that allows booting a server over iSCSI ⢠unmâSupports HP NC510x PCIe 10 Gigabit Adapter driver The drivers are distributed in source RPM format and are available in the HP SoftPaq file in the \Linux directory. Complete the following steps to download the HP SoftPa q. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number.
Installation 21 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. In addition, the Linux (open source) drivers can be found at the HP software and drivers website ( http://h18000.www1.hp.com/s upport/files/index.html ). HP ROMPaqs for specific s ervers can be found at the HP support website ( http://www.hp.com/support ). The source RPM is suitable f or use on any system that has the RPM utility installed and a properly configured kernel source tree. SCO OpenServer and UnixWare SCO OpenServer 5.x, 6.x, and UnixWare 7.x driv ers are provided in the following locations: ⢠HP SoftPaq file at the HP website ( http://www.hp.com ). ⢠HP EFS 7.64A for UnixWare 7.x and OpenServer 6.x, and HP EFS 5.64A for SCO OpenServer 5.x at the HP website ( http://www.hp.com ). HP Extended Feature Supplement (EFS) is a colle ction of device drivers that provide maximum performance for HP servers. T he EFS also includes several utilit ies to monitor and increase the performance of your HP server. SCO OpenServer 5.x This procedure describes the installation of the SCO O p enServer driver. This driver is released as a media image file containing the driver package. The media image file can be copied to the target ma chine directly for installation or from an in stallation diskette that you can create. Creating a diskette 1. Copy the file VOL.000.000 to a Calder a system. 2. Create a diskette using: dd if =VOL.000. 000 of=/dev/rfd0135ds18. Installing the driver To install the driver: 1. Use custom or scoadmin software to instal l the SCO OpenServer driver from the media image or from the diskette created above. 2. Use netconfig to add the new network adapter. 3. Modify the hardware configuration in Advanced Options to change the Line Speed and Flow Control if desired. The settings for th ese parameter s are listed below: o Line Speed AutoNegotiate (defa ult) HalfDuplex10 FullDuplex10 HalfDuplex100 FullDuplex100 FullDuplex1000 o Flow Control Off (default) AutoNegotiate (Symmetric Paus e advertised)
Installation 22 RxPause TxPause RxPause/TxPause NOTE: 1000 Mbps (1 Gbps) full-duplex fixed speed is only valid for fiber connections. For copper, twisted-pair connections , 1 Gbps can only be set through auto-negotiation with a 1 Gbps partner. NOTE: Auto-negotiation of Flow Control is only valid when the Line Speed is set to AutoNegotiate. NOTE: A kernel relink and reboot is required before the new configuration will take effect. For more detailed information and tuning par ameters, at the command prompt enter: man bcme Jumbo Frames and other advanced parameters Jumbo MTU sizes and other advanced, tunable paramet ers for the controller are located in the file space.c in the directory /etc/conf/pack.d/bcme. A description for each parameter is contained in space.c. Modify the desired parame ter in space.c, rebuild the kernel, and reboot the syst em. The MTU sizes can be individually set for each adapter in the system (Jumbo Frames is not supported on the NC1020). All other parameters apply globally to all adapters. UnixWare 7.x and OpenServer 6.x The following procedure describe s the installation of the UnixWare driver. The driver is release d as a PKG file. Installing the driver locally 1. Copy the file bcme-<version>.pkg to a UnixWare 7.x system. 2. Enter the following command: pkgadd -d <path><filename> where <path> is the directory to wh ich the PKG file was copied. Creating a diskette 1. Copy the file bcme-<version> .pkg to a Caldera system. 2. Create a diskette using the following command: dd if=bcme-<vers ion>.pkg of=/d ev/fd0 Installing the driver from a diskette 1. Use pkgadd -d diskette1 to install the bcme package. 2. Use netcfg or scoadmin network to add the new network adapter. 3. When prompted, select the Line Speed and then select Advanced Option for Flow Control, MAC Address, and Jumbo MTU Size settings, if desir ed. The settings for these parameters are li sted below: o Line Speed Auto Ne gotiati on (de fault ) 10 Mbps half-duplex 10 Mbps full-duplex 100 Mbps half-duplex
Installation 23 100 Mbps full-duplex 1000 Mbps full-duplex o Flow Control Disabled (default) Auto Neg otiation ( Symmetr ic Paus e advertised) Receive Pause Transmit Pause Receive & Transmit Pause o MAC Address No Override (default) - a user-administered M AC address entered with a colon separating each hexadecimal byte (for example, 12:34:56:78:9a:bc). o Jumbo MTU Size 1500â9000 (default i s 1500) For more detailed information, at the command prompt, enter: man bcme NOTE: 1000 Mbps (1 Gbps) full-duplex fixed speed is only valid for fiber connections. For copper, twisted-pair connections , 1 Gbps can only be set through auto-negotiation with a 1 Gbps partner. NOTE: Auto-negotiation of Flow Control is only valid when the Line Speed is set to AutoNegotiate. Solaris The following describes the Solaris driv er for HP 10/100/1000 Mbps Netwo rk Server Adapters. The driver is released in two formats: ⢠BRCMbcme.pkg: Datastream format ⢠bcmedu.img: Driver Update (DU) binary imag e which is used to create DU diskette. Installing the driver 1. Change directory to wher e BRCMbcme.pkg resid es. 2. pkgadd -d BRCMbcme.pkg 3. Execute prtconf to determine instance number of the adapter. 4. ifconfig bcme [instance_number] plumb 5. ifconfig bcme [instance_nu mber] ip_address netmask .... To make these changes permanent: 1. Use your favorite text editor (for ex ample, VI), and create a file named hostname.bcme[instance_number] in /etc directory. Add the IP address of the interface to this file, save and exit. 2. Add a proper subnet mask to the file /etc/netmasks.
Installation 24 In Solaris 7.0 (Intel platform), the operating system only allocates 36 pages of 4 K physically contiguous memory. The driver needs approximately 130K of physic ally contiguous mem ory per adapter. In order to use more than one adapter, the operating system must allocate more memory. This can be done by setting the operating system variable, "lomempages" in /etc/ system. For example, if four adapters are installed in a So laris 7 system, physically contiguous memory is calculated as follows: 4 adapters * 130K = 520K 130 pages of 4K is required. Since this memory might be used by another driver in the system, 200 of 4K memory is allocated. Add the following line in the /etc/system file: set lomempages=200 Uninstalling the driver To uninstall the driver: 1. ifconfig bcme [instance_number] down 2. ifconfig bcme [instance_number] unplumb 3. pkgrm BRCMbcme Driver update (DU) diskette instruction (Intel® Platform Only) This area contains the Solaris diskette image files, and the instruction s to create diskettes from these image files. 1. Insert a blank diskette into the di skette drive of your system, and type one of the following commands to format it: o If you are using DOS, type: format A: o If you are using the Solaris operating environment, type: fdformat -Ud 2. Check to see if Volume Management is running: a. Do one of the following: If you are using DOS, type: dd filename A: If you are using the Solaris operating environment, type: volcheck ls -l /vol/dev/aliases/floppy0 b. If you see a message similar to this: lrwxrwxrwx 1 root 34 Jan 21 17:28 /vol/dev/aliases/floppy0 -> /vol/dev/rdiskette0/unnamed_floppy Type: dd if=bcmedu.img of=/vol/dev/aliases/floppy0 bs=36k eject floppy0 c. If you see this message: /vol/dev/aliases/floppy0 not found Type:
Installation 25 dd if=bcmedu.img of=/dev/rdiskette bs=36k Installing Solaris DU dskettes You can use Solaris DU diskettes in the following ways: ⢠To use new drivers to install or upgrade t he Solaris operating system environmen t on a system with new hardware. ⢠To add new drivers to support new hardware on an already installed and booted system. Installing Solaris using DU diskettes To install Solaris (Intel Platform Edition) using driver s on the DU diskette: 1. Insert the appropriate configuration Assistant diskette , made from the included bcmedu.img file, into your system's disk drive. Also, insert the S olari s Installation CD-ROM or, for network installation, verify with your system administrator that the Sola ris network installation im age is available on your network. 2. Power on your system. 3. When the Configuration Assistant scr een is displayed, select the F4 option (on ver sion 2.6, it is labeled F4_Driver Update; on Solaris 7, it is la beled F4_Add Driver). Th e message, "Enumerating buses...." is displayed. 4. Remove the Configuration Assistant diskette from the disk drive and insert the first Solaris DU diskette. 5. Select F2_Continue. The Select Solaris System Ve rsion screen is displayed. 6. Select the appropriate Solaris OS and select F2_Continue. The Loading Driver Update Software screen is disp layed, along with a prog ress bar that indicates the percentage of drivers that have been extracte d from the diskette. Drivers are read into mem ory and survive long enough for the system to successf ully boot to its installation program. When all the new drivers on the diskette have been processed, th e Contin ue Driver Update Installation screen is displayed. 7. If you have additional DU diskettes to install, remove the DU diskette from the dis k drive, insert the next DU diskette. 8. Select F2_Continue. Again, the Loading Driver Update Software screen is displayed, along with a progress bar that shows the percentage of drivers that have been extracted from the diskette. Drivers are read into memory and survive long enough for the syst em to successfully boot to its installati on program. When all the new drivers on the diskette have been processed, the Contin ue Driver Update Installation screen is displayed. 9. Repeat step 7 and 8 until all desired DU diskettes are i nstalled. 10. When all the drivers are processed, remove the DU diskette from the disk drive and reinsert the Configuration Assistant diskette. IMPORTANT: Do not remove the Configuration Assistan t diskette from the disk drive until you see the following message display: "If you want to bypass the device configuration and boot screens when the system reboots, eject the Device Configuration Assistant/Boot diskette now." 11. Select F2_Continue. The Solaris Device Configuration Assistant scr een is displayed. 12. Select F2_Continue. The message, "Enumerating buses...." is displayed, followed by the Scanning Devices screen. System devices are scan ned. When scanning is complete, the Identified Devices screen is displayed.
Installation 26 13. Select F2_Continue. The message "Loading driv er...." is displayed, followed by messages about the drivers that are required to boot your system. After a few seconds, th e Boot Solaris screen is displayed. 14. At the Boot Solaris screen, select the device controller attached to the device th at contains your install medium. 15. Select F2_Continue. Drivers display for the device controller you sele cted. Your system boots to run the install program. The install program starts and your system begins booting the complete Solaris operating system environment. Then the follow ing messages are displayed, prompting you to in sert each of the Solaris DU diskettes required to install on your system: Installing unbundled device driver support Extracting driver list from tree.. <DU diskette name> driver-name... Please insert the Driver Update diskette labeled <DU diskette name> driver-name. Press <ENTER> when ready. 16. Remove the Configuration Assistant diskette and rein sert the first DU diskette into the disk drive. 17. Press the Enter key. Packages, patches, or both that cont ain the new drivers are installed from the diskette onto your system. Me ssages about each inst allation are displayed. If drivers on other D U diskettes are required for your sy stem, the following prompt is displa yed. Please insert the Driver Update diskette labeled <DU diskette name> Press <ENTER> when ready. Otherwise, this prompt is displayed: If you have additional Update diskettes to install (such as video), please insert diskette now. Additional Update diskettes to install? (y/n) [y] 18. If drivers on other DU diskettes are requir ed, remove the DU diskette from the diskette drive, insert the next DU diskette you are prompted to insert. Press t he Enter key. Repeat until all the drivers you need are installed. 19. Press the Enter key. When installation is complete, the messag e "Installation complete" is displayed. 20. Remove the diskette from the disk drive. 21. Reboot the system. When the Solaris operating system is fini shed booting and running, the new devices whose drive rs you installed are available for use. Adding a DU diskette to an existing Solaris system Before adding new or updated drivers, the newly su pported hardware devices should be installed and configured according to the instructions in the corre sponding Device Reference Page, if any. Refer to the Solaris (Intel® Platform Edition) Device Configuration Guide. When the Solaris (Intel® Platform Edition ) software is already installed, the simplest way to add new or updated drivers is to install the DU diskett es as patches on your system, as follow s: 1. Become the superuser on your system. 2. Check to see if Volume Man a gement is running on the system you are updating: ps -ef | grep vold For more information about managing diskette and drives, refer to the System Ad ministration Guide.
Installation 27 3. If Volume Management is running, temporarily stop it by typing: # /etc/init.d/volmgt stop 4. Insert the DU diskette into the disk drive. 5. Mount the DU diskette at the /mnt mount point: # mount -F pcfs /dev/diskette /mnt At this point, you must mount the DU diskette in the file structure to up date your system successfully. 6. Execute the install script on the diskett e, using th e appropriate Solaris release directory (currently sol_26 for Solaris 2.6, sol_27 for Solar is 7, and so on). For example: # /mnt/DU/sol_27/i86pc/ Tools/install.sh -i The install.sh script searches for all new or updat ed drivers on the diskette. When a new or updated driver is found, the following prompt is displayed: Unconditionally installing DUs <DU driver names> Install patch driver-name? [y] 7. If the driver is the one you want to install, at the prompt, type y for yes or press th e Enter key. If the driver is not the one you want to install, type n for no. If you select "yes," then the install.sh script inst alls the driver you selected, as well as the bootmod and bootbin patches. 8. When you have finished, and the install.sh script exits, unmount th e diskette: # cd / # umount /mnt 9. Remove the DU diskette from the disk drive. 10. Reboot the system. # touch /reconfigure # reboot 11. Power down your system, add the new hardware, and power up the system. 12. When the autoboot sequence prompt is displayed, quickly press the Esc key. The au toboot sequence is interrupted, and the Configuration Assistant screen is displayed. 13. Select F2_Continue. The message "Enumerating buses ..." is displayed, followed by the Scanning Devices screen. System devices are scan ned. When scanning is complete, the Identified Devices screen is displayed. 14. Select F2_Continue. The message "Loading driver com.bef ..." is displayed, followed by the Boot Solaris screen. 15. On the Boot Solaris screen, select the device controller attached to the devi ce that contains your install medium, in this case the main system disk. T he /etc/bootrc script is displayed. 16. At the prompt, type: b -r Your system boots. You can now use your new hardware.
Installation 28 NetWare A network device driver must be in stalled before the Gigabit Ethernet adapter can be used with your Novell NetWare system. Before you c an successfully in stall the adapter driver for Novell NetW are, the adapter card must be physically installed in the serv er and, typically, NetWare OS software must already be running on the server. Make sure that your server meets the hardware and operating system software requirements. For an adapter installation with an existing NetW are server, NetWare will automatically detect the new adapter and attempt to load the appropriate driver. Ensu re that your CD-ROM is mounted and select the proper driver. To enable the Gigabit Ethernet adapter to function corr ectly, you need to install the latest support pack files. The latest support pack can be found at the Novell website ( http://www.novell.com ). Netware install program A commonly used method to install a driver on a NetWare server is through NWCONF IG. The following drivers are supported. ⢠Driver names: N1000e.LAN, N1000.LAN, and N100.LAN ⢠HP N1000e.LAN driver requirements o Use with an HP NC110T, NC360T, an d NC364T Gigabit Server Adapter o Always install the latest Novell support pack o Always install the latest HP NSSD o SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) o SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) o SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) ⢠HP N1000.LAN driver requirements o Use with any HP NC310F, NC340T, NC61 xx, and NC71xx Gigabit Server Adapter o Always install the latest Novell support pack o Always install the latest HP NSSD o SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) o SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) o SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) ⢠HP N100.LAN driver requirements o Use with any HP NC31xx Fast Ethernet Server Adapter o Always install the latest Novell support pack o Always install the latest HP NSSD o SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) o SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) o SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) Using NWCONFIG or HDETECT to install the adapter
Installation 29 NOTE: Your NetWare server software must be th e latest version of the support pack before installing the HP driver. The latest support pack is available from the Novell support site: http://support.novell.com/misc/patlst.htm. Before installing, use the HP SoftPaq to create a floppy disk on a different computer. Refer to the README.TXT file located in the HP SoftPaq for det ails on creating a floppy disk. Use the NetWare Install programs to install the adapter driver. See the Novell Installation Guide for specific instructions on installing device drivers. Packet Receive Buffers Setting is Too Low error message When a server is first installed or has been upgraded, the number of MAXIMU M PACKET RECEIVE BUFFERS and MINIMUM PACKET RECEIVE BUFFERS are usually too low. The adapter requires a minimum of 64 Packet Receive Buffers per ad apter port during driver initialization. This initial load of the driver may fail because of the two above mentioned se ttings. The followin g error message is displayed: *** Packet Rece ive Buffers setting is too low. *** To correct: 1. Load MONITOR. 2. Select SERVER PARAMETE RS. 3. Select COMMUNICATIONS. 4. Set MAXIMUM PACKET RE CEIVE BUFFERS = 10000 (or larger). 5. Set MINIMUM PACKET RE CEIVE BUFFERS = 2000 (or larger). 6. Set MAXIMUM PHYSICAL RECEIVE P ACKET SIZE = 2048 (or larger). 7. Restart NetWare to complete. This can also be set by using the NWCONFIG an d selecting the Startup.NCF file. See the Netware Installation Guide for specific instructions. The maximum and minimum settings of 10000 and 2000 respectiv ely are suggested values only. You may need to request a larger number for MAXI MUM PACKET RECEI VE BUFFE RS depending on the number of supported adapter ports in the system. Th is message is also displayed on the console if the driver cannot allocate the required number of Receive Buf fers during driver load. Driver configuration parameters Parameter Options Description CheckSum = Default = ON Selections are: ON, OFF, Tx, Rx Enables or disables the transmit and receive checksum off-loading feature. Checksum is supported under NetWare 5.x only. If you want to enable the Ch eckSum parameter, you need to load it on the first instance. Frame = type Valid types are: Ethernet_802.2, Ethernet_802.3, Ethernet_II, Ethernet_SNAP Defines the frame type being used by this load instance. Ethernet_802.2 and Ethernet_II are the default values. node = NNNNNNNNNNNN Specifies a no de address in this field to override the default Media Access Controller (MAC) address (also known as the Locally Administered Address) name = text Displays the name assigned to this adapter
Installation 30 Parameter Options Description PDriver = Default = OFF Selections are: OFF, ON Allows driver to operate in persistent driver mode. Persistent driver mode is supported under NetWare 5.x only. Use only if adapter is placed in a Hot Plug PCI slot and only if required to swap with an exact board. RxBuffers = Default = 200 Recommended Min = 32 Max = 512 Min = 1 when used with DOS Client32 and when Keywrod P3-1 or 2. Pre-allocates ECB resources on the adapter for receiving packets TxDescriptors = Default = 200 Recommended Min = 100 Max = 512 Min = 1 when used with DOS Client32 and when Keywrod P3-1 or 2. Pre-allocates ECB resources on the adapter for transmitting packets. RxFlow = Default = OFF Selections are: ON, OFF Allows enabling/disabling of RxFlow control. TxFlow = Default = OFF Selections are: ON, OFF Allows enabling/disabling of TxFlow control. Slot = n Identifies the slot number for the specific adapter currently being configured. This parameter is not necessary if only a single adapter is installed. Speed = n If link negotiation has been disabled, specifies port speed to be either Auto, 10HD or 10FD, 100HD or 100FD. Jumbo = Set maximum physical receive packet Size = 18000 in the STARTUP.NCF. Choices are Jumbo = 1514â9000. T his keyword is only supported on NetWare 6.x. Enables/disables Jumbo Frame support. When enabled, jumbo packets of up to 9000 bytes are supported. Not supported on NC1020 adapters. Link= Default=FORCE Selections are: AUTO, FORCE Only used to allow the adapter to negotiate a specific or forced line speed with a switch that is not forced, but instead setup for auto- negotiation. It is best to allow for auto- negotiation of the card and switch by not setting this keyword or the speed keyword. Only use this keyword if the speed keyword is set to something other than AUTO. RxTicks= Default = 360 Min = 0, disabled Max = 5000000, 5 second s Units are in micro seconds Enables the use of batching receives within a specific time period.
Installation 31 Parameter Options Description TXPacketsPer Default = 64 Min = 0, disabled Max = 100 Enables the use of batching transmits to a specific amount of packets. NOTE: With Jumbo Frames, the first frame must be Ethernet_ii. Post installation After NetWare 6.x has been successfully installed, set the minimum packet receive buffers param eter in the startup.ncf file to 1500 for each adapter in th e system. Set the maximum packet receive buffers to three times the minimum packet receive buff ers. Typica lly one MB of RAM is required per 1000 receive buffers. For more information, refer to the "Configuring Driver Param eters" section. In the autoexec.ncf file, delete the packet receive buffers param eter (RxBuffers=32) in the l oad statement for this adapter. Deleting the receive b uffers phrase from the load statement res ets the receive buffers parameter to the default value of 200 fo r this adapter. NOTE: You must reboot the server for the new configuration. Example: The default maximum number of rec eive buffers for the system is 500; the default minimum is 128. Edit the startup.ncf file to have the following entries. Th e actual numbers will be a func tion of the number of adapters in the system. ⢠set maximum packet receive buffers = 30000 ⢠set minimum packet receive buffers = 1 0000 ⢠set maximum physical receive packet size = 2048 Verifying or modifying adapter properties When an adapter configuration is saved, the NetWare install program adds load and bind statements to the autoexec.ncf file. By accessing this file , you can verify the parameters configured for each adapter, modify them, or enter additional parameters. NOTE: The Novell monitor program and the CONFIG command are also us eful for verifying driver configuration. For information on how to use these programs, ref er to the utilities reference in your Novell NetW are online documentation. The parameters that can be defined in the load statem ents are described in NetWare server dri ver LOAD line parameters for HP server adapters below. A valid au toexec.ncf file is shown below. One set of load and bind commands is added for each frame type the adapter is configured to support. Valid Autoexec.ncf file Set Time Zone = PST8PDT set Daylight Savings Time Offset = 1 set Start Of Daylight Savings Time = (APRIL SUNDAY FIRST 2:00:00 AM) set End Of Daylight Savings Time = (OCTOBER SUNDAY LAST 2:00:00 AM) set Default Time Server Type = SINGLE set Bindery Context = O=LAN
Installation 32 # WARNING!! file server name NOVELLSERVER51 # WARNING!! # If you change the name of this server, you must update # all the licenses that are assigned to this server. Using # NWAdmin, double-click on a license object and click on # the Certificate Assignments button. If the old name of # this server appears, you must delete it and then add the # new server name. Do this for all license objects. ServerID 1C8EE2C LOAD ODINEB.NLM LOAD TCPIP LOAD Q57 SLOT=2 FRAME=Ethernet_802.2 NAME=Q57_1_E82 BIND IPX Q57_1_E82 NET=FAFD3D25 LOAD Q57 SLOT=2 FRAME=Ethernet_802.3 NAME=Q57_1_E83 BIND IPX Q57_1_E83 NET=5A2D8D6D LOAD Q57 SLOT=2 FRAME=Ethernet_SNAP NAME=Q57_1_ESP BIND IPX Q57_1_ESP NET=477A35BD LOAD Q57 SLOT=2 FRAME=Ethernet_II NAME=Q57_1_EII BIND IPX Q57_1_EII NET=C3C8F2E4 BIND IP Q57_1_EII ADDR=172.16.1.1 MASK=ff.ff.0.0 mount all SEARCH ADD SYS:\JAVA\BIN SEARCH ADD SYS:\JAVA\NWGFX IMPORTANT: If you modify any adapte r parameters, you must reboot the system before the changes will take effect. If you make changes and do no t reboot, you may experience configuration problems. Removing drivers from Autoexec.ncf To remove the drivers from the Autoexec.ncf, locate the LOAD and BIND command lines as sociated with the driver and remark them out by inserting the poun d (#) symbol at the beginning of each command line. Example: # LOAD Q57 SLOT=2 FRAME=Ethernet_802.2 NAME=Q57_1_E82 # BIND IPX Q57_1_E82 NET=FAFD3D25 # LOAD Q57 SLOT=2 FRAME=Ethernet_802.3 NAME=Q57_1_E83 # BIND IPX Q57_1_E83 NET=5A2D8D6D # LOAD Q57 SLOT=2 FRAME=Ethernet_SNAP NAME=Q57_1_ESP # BIND IPX Q57_1_ESP NET=477A35BD # LOAD Q57 SLOT=2 FRAME=Ethernet_II NAME=Q57_1_EII # BIND IPX Q57_1_EII NET=C3C8F2E4 # BIND IP Q57_1_EII ADDR=172.16.1.1 MASK=ff.ff.0.0 NetWare server driver LOAD line parameters This following lists the N etWare server driver LO AD line parameter settings for HP serve r adapters for N100, N1000, and N1000e. ⢠FORCEDUPLEX. This keyword specifies one of the following dupl ex modes: o Auto-negotiate. The adapter negotiates with the switch and hub how to send and receive packets at either full- or half-duplex speed. If uns uccess ful at negotiating the duplex, HP server adapters default to half-duplex. Yo u must have an auto-n egot iating switch/hub (an Nway switch) to get full-duplex support with the FORCEDUPLEX parameter set to 0 (auto-negotiation).
Installation 33 o Full-duplex. The adapter sends and receives packets at the same time. This improves the performance of your adapter. Set duplex mode to full-duplex ONLY if you have a hub/switch that supports full-duplex. o Half-duplex. The adapter performs one operation at a time. It either transmits or receives. NOTE: The HP NC31xx, 340T, and NC71xx Server Adapters support full-duplex and half- duplex. The HP NC61xx fiber Gigabit Server Ad apters support full-duplex mode only and cannot be changed. NOTE: The NC3133 default to full-duple x. NOTE: To set the adapter to half- or full-duplex, yo u must set the SPEED parameter to either 10 or 100. o Syntax: FORCEDUPLEX=n Where n = 0âauto-negotiate 1âhalf-duplex 2âfull-duplex o Default = auto-negotiate o Examples: 100 Mbps full-duplex FORCEDUPLEX=2 SPEED=100 10 Mbps full-duplex FORCEDUPLEX=2 S PEED=10 ⢠SLOT. (Required) For PCI adapters, SL OT is derived from bus number and device location as defined by the PCI specification. One way to determine th e slo t number is to load the driver from the command line. You will be prompted with valid de vice number(s) for the adap ter(s). Select one of them. o Syntax: SLOT=n Where n = 1, 2, 3, 4,... ⢠SPEED. This keyword specifies the speed the driver use s. If you do not use this parameter, the driver automatically detects the network spee d. If unable to detect, the driver defaults to 10 Mbps. If you use this parameter, the driver operates at the specified speed instead of auto detecting netw ork speed. NOTE: The NC3133 adapter supports 100 Mbps on ly. It does not default to 10 Mbps. You must set the SPEED parameter to either 10 or 100 if you want to set the FORCEDUPLEX parameter to either half or full. The SPEED parameter is for N100.LAN, N1000.LAN, and N1000e.LAN drivers. You cannot change the spee d for the Gigabit (fiber) adapters. You can, however, change the speed for the NC7 170, NC7131, and NC7132 Gigabit (copper) adapters to 10 Mbps or 100 Mbps. o Syntax: SPEED=n Where n = 10 or 100 or 1000 o Default = The adapter automatically detects network speed.
Installation 34 ⢠NODE. This keyword specifies a locally administered address (LAA) unique to each adapter. Use this option to provide your own unique node address for the adapter. The node address is a 12-digit hexadecimal number. The second digit must be one of the followin g digits: 2, 6, A, E. o Syntax: NODE=xnxxxxxxxxxx Where n = 2, 6, A, E Where x = hexadecimal number o Default = The adapter's assigned address ⢠FRAME. This keyword indicates one of fo ur valid frame types the system is transmitting and rec eiving. o Syntax: FRAME=n Where n = Ethernet_802.2, Ethernet_802.3, Ethernet_II, Ethernet_SNAP o Default = Ethernet_802.2 ⢠POLL. This keyword is used to enable polling mode in the driver with interrupt backup. If polling is enabled, interrupts will be reduced, allowing th e processor to spen d more time performing other functions. In polling mode, interr upts will occur only when receive re sources have been reduced to less than half. If polling is no t enabled, the driver will perfor m in traditional interrupt mode. o Syntax: POLL=n o Where n = 0 (interrupt mode), 1 (polling mode with int errupt backup) o Default = 0 (interrupt mode) NOTE: The POLL parameter is for the N100.LA N, N1000.LAN, and N1000e.LAN drivers. ⢠SPURIOUS. This keyword is used to reduce the number of spurious interrupts reported by the OS on the driver interrupt line. This automati cally loads for 10, 100, and 1000 adapters. o Syntax: SPURIOUS=n Where n = 0 or 1 0âDriver operates in norm al mode 1âDriver operates to reduce the number of spur ious interrupts o Default = 1 NOTE: The minimum threshold for acceptable number of spurious int errupts is 200 interrupts per second, which is the default value of the environmental variable "set display spurious interrupt alerts threshold." ⢠RXCHECKSUM. This keyword is used to enable or disable the offlo ad of Receive Packet Checksum verification to the adapter. This au tomatically loads for 10/100 adapters. o Syntax: RXCHECKSUM=n Where n = 0 or 1 0âDoes not offload Rx Checksum ver ification to the adapter 1âOffloads Rx Checksum verification to the adapter o Default = 0 NetWare custom statistics NetWare custom statistics appear when you run monitor NLM and load the LAN driver. One custom statistic counter is the duplex mode, where 1 = half-duplex and 2 = full-dupl ex.
Installation 35 The following describes the adapter Status Counters and Software Counters an d their acceptable ranges. N100 custom counters ⢠Link Status This counter contains the link status, as read from the PHY, or general status register of the ASIC. It is updated once a second. A value of 0 is link up. 1 is link down. ⢠Duplex This counter holds the current duplex as read directly from the PHY, or general status register of the ASIC. A value of 0 is not supported or unknown, 1 is half-duplex, and 2 is full-duplex. Unsupported is usually due to an unsupported chip type, or link not being present. ⢠Current Speed This counter holds the current speed as read directly from the PHY, or general status register of the ASIC. A value of 0 is not supported or not available, 10 for 10 Mbps, 100 for 100 Mbps, and 1000 for 1000 Mbps. Unsupported is usually due to an unsupported chip type, or link not being present. ⢠Hardware Status A counter of non-zero indicates failure. ⢠Tx Good Frames This counter contains the number of frames that we re transmitted properly on the link. It is updated only after the actual transmission on the link is completed, and not when the frame was read from memory as is done for the TxCB status. This value va ries in accordance to the amount of frames sent, and can reach billions very quickly. ⢠Tx Maximum Collisions This counter contains the number of frames that were not transmitted because they encountered the configured maximum number of collisions. This counter increments when the net work is heavily saturated with traffic. If this va lue is consiste ntly growing, you may want to segment your network with switches to change the collision domain, so that more time is spent sending data instead of collisions. ⢠Tx Late Collisions This counter contains the number of frames that were not transmitted because they encountered a collision outside of the normal collision window. No rmally, late collision events should occur only rarely, if at all. A high number of late collisio ns is of ten a sign that there is some problem with the network topology. Check the topology to be sure that all cables m eet IEEE specifications (Category 5 for 100 Mbps TX), and that all distance requirements for cabling are met. Also, check to be su re that all of the network hubs and switches are configured for the proper duplex mode. ⢠Tx DMA Underruns This counter contains the number of frames that we re either not transmitted or re-transmitted due to a Tx DMA under-run. If the device is configured to retransmit on under-run, this counter may be updated multiple times for a single frame. Un der-runs occur because a lack of PCI bandwidth resulted in the device's internal transmit FIFO running dr y during the transmission of a frame. The adaptive algorithms within the software keep this nu mber under control. If a high number of under- runs are seen, you may want to load the driver with TxThreshold=200 on the command line. Example: Load N100 slot=2 Frame=Ethernet_802.2 TxThreshold=200 For additional information about the adaptive algori thm , refer to the "CurrentThreshold" section.
Installation 36 ⢠Tx Lost Carrier Sense This counter contains the number of frames that were transmitted by the device despite the fact that it detected the de-assertion of Carrier Sense durin g the transmission. ⢠Tx Frames Deferred This counter contains the number of frames that were deferred before transmission due to activity on the link. This is a normal condition in busy co nditions, and the number can grow quite large. ⢠Tx OK with Single Collisions This counter contains the number of transmitted fram es that encountered one, and only one, collision before being successfully sent. This is a normal condition in busy conditions. If sin gle collisions accumulate quickly, you may want to consider segm enting your network more . To combat collis ions, the software can adjust the interframe ga p to a llow more room between fr ames. More room between frames allows more frames on av erage. Refer to the "CurrentIFSValue" section for more details. ⢠Tx OK with Multiple Collisions This counter contains the number of transmitted frames that encountered more than one collision. This only gets incremented once, even though one frame m ay have had several collisions before being successfully s ent. This is a normal condition in busy conditions. If mu lti-collisions accumulate quickly, you may want to consider segmenting your netw ork more. To find the average number of multi- collisions a frame had to endure before be ing successful, use the following formul a: (TxTotalCollisions - TxSingle Collisions - TxMaxCo llisions) / TxMultiple Col lisions If the average is close to 16 (the Ethernet max co llisions before a fra me will no longer be retried), then you will need to segment your network to avoid losing frames and network efficiency to collisions. ⢠Tx Total Collisions This counter contains the total number of collisi ons that were encountered while attempting to transmit. This count includes late collisions and collisions from frames that encountered MAXCOL(16). If this counter is gro wing quickl y, you need to segment your network more. ⢠Rx Good Frames This counter contains the number of frames that we re received properly from the link. It is updated only after the actual reception from the link is completed and all the data bytes are stored in memory. ⢠Rx Crc Errors This counter contains the number of aligned frames discarded because of a CRC error. The CRC_ERRORS counter i s mutually exclusive to the ALIGNMENT_ERRORS and SHORT_FRAMES counters. CRC errors are usually ca used by late collisions whe re another station has sent data in the middle of a frame. This number should be very low. ⢠Rx Alignment Errors This counter contains the number of frames that are both mis-aligned (not an even length) and contain a CRC error. The ALIGNMENT_ERRO RS counter is mutually exclusive to the CR C_ERRORS and SHORT_FRAM ES counters. ⢠Rx No Resource Errors This counter contains the number of good fram es discarded because there were no resourc es available. Frames intended for a host whose Receive Unit is in the " No Resources" state fall int o this category. The number of resource errors can be re duced by increasing the number of receive buffers
Installation 37 that software uses, or increasing the perf ormance of the host machine's proce ssing power (which would allow for faster processing of the r esources). ⢠Rx DMA Overrun Errors This counter contains the number of frames known to be lost because the device's internal receive FIFO overflowed. This can occur if the device is unable to get the necessary bandwidth on the system (PCI) bus. If the overflow condition persists for more than one frame, the frames that follow th e first could also be lost. Howeve r, because there is no lost frame indicator they may not be counted. A frame that was counted as an overrun will not be counted in other error counters (SHORT_FRAMES, CRC_ERRORS , or ALIGNMENT _ERRORS). To c ombat th is, a faster bus machine, o r fewer devices on the bus will be needed. ⢠Rx Collisions Detect Errors This counter contains the number of frames that encountered collisions during frame reception. ⢠Rx Short Frame Errors This counter contains the number of received fram es that are shorter than the minimum frame length. The SHORT_FRAMES coun ter is mutually exclusive to the ALIGNMENT_ERR ORS and CRC_ERRORS counters and has a higher priority (a short frame will always increment only the SHORT_FRAMES counter). ⢠Flow Control TX Pause This counter contains Xon and Xoff frames transmitted. ⢠Flow Control RX Pause This counter contains Xon and Xoff frames received. ⢠Flow Control Unsupported This counter contains the MAC Control frames re ceived that are not Flow Control Pause frames. These frames contain the predefined MA X control type value but contain an unsupported opcode. ⢠TCO Rx (or Tx) Frames This counter is incremented at each received/transmitted TCO packet. ⢠Tx Timeouts This is a debug timer and not currently used. It should be zero. ⢠Tx Dynamic Chaining Count This is a performance evaluation counter an d not currently used. It should be zero. ⢠Current IFS Value This is the current Inter Frame Spacing (IFS) value that the driver is using. IFS is an Ethernet specification term for clock time spaces between fr ames sent by an individual card in a burst. Normally a card sends frames back to back with only the minimum frame spacing between each frame. Under low load conditions, this produc es maximum throughput. But under heavy load conditions, this means less time for other stations to send safely; thus, more colli sions and less throughput. By adaptively controlling the IFS, more stations can safely send, lo wering collisions and helping to keep throughput high. The larger this number is, the more collisions your n etwork is experiencing. ⢠Current Threshold As under-runs occur, the software can change the amount of time the hardware will wait before starting to transfer data onto the adapter's hardware buffers. By doing so, it can slow or stop the
Installation 38 amount of under-runs. The higher th is number, the more time the adapter is waiting befor e starting to transmit data onto the wire. Poor bus performance, large amounts of traffic or a large numb er of devices on the same bus can cause this to rise. If this number is always above 170, you may want to consider loading the driver with the TxThreshold parameter set to a higher value, such as 200. ⢠RstartCnt Re-Start Count, also known as RSC, is a support counter to see how often the restart effort occurs. Under some high traffic, or high utilization condit ions, the receiving unit does not start normally and needs a restart. RSC has no impact on the end user. It is intended for softwar e development use only. It is normal for this to be non-zero. ⢠REC Under some busy conditions, the receiving unit stalls and needs to be restarte d. This counter tracks the number of times it needs to be restart ed. RSC is a subset of REC. Most REC events restart normally, but the few that do not are REC events. Th is counter is intended for software developm ent use only and may be disregarded by the end user. It is normal for this to be non-zero. ⢠Checksum Good This counter indicates the number of received packe ts that the hardware has verified, the checksum, and if the checksum is valid. ⢠Checksum Bad This counter indicates the number of received packe ts that the hardware has verified, the checksum, and if the checksum is valid. ⢠Stalled Ints SW Resets This counter indicates the number of times the driver issued a complete reset (to the driver and to the adapter) as a result of a stalled in terrupt situation. A stalled interr upt situ ation is detected by the driver if there are no interrupts for a long time, and a SW interrupt issued by the driver fail s to arrive. ⢠Carrier Loss PHY Resets This counter indicates the number of times the driver reset the PHY a s a result of carrier sense loss. N1000 and N1000e custom counters ⢠Duplex Mode (1=Half, 2=Full) This counter indicates the current link duplex conf iguration. A value of 1 indicates the adapter is configured to half-duplex mode. A 2 indicates ful l-d uplex mode configuration. This counter exists in v1.12 or higher of the NetWare server driver. ⢠Line Speed (Mbps) This counter displays the cu rrent line speed in Mbps. ⢠Configured Duplex Mode (0=Auto, 1=Half, 2=Full) ⢠Configured Speed (Mbps; 0=Auto) ⢠PCI Mode (0=PCI, 1=PCI-x, 2=PCI-Expres s) ⢠Bus Clock Rate (Mhz) (0=Unavailable) ⢠Link Status (0=Up, 1=Down) This counter describes the current link status of the adapter. A value of 0 indicates the link is up and a value of 1 indicates that the link is down. This counter exists in v1.12 or higher of the NetWare server driver.
Installation 39 ⢠Link Status Change Count This counter indicates the number of ti mes the link status has been changed. ⢠Rx CRC Error Count This counter indicates the number of receive packe ts with CRC errors. Packets less than 64 bytes are not counted in this statistic. ⢠Alignment Errors This counter holds the number of receive packets wi th alignment errors. In order for a packet to be counted, it must pass address filtering and must be 64 bytes or greater in length. If receives are not enabled, then this counter does not increment. This counter is valid only in MII mode during 10/100 Mbps operation. ⢠Rx Symbol Error Count This counter indicates the number of symbol erro rs between reads. The coun t increases for every bad symbol received, whether or not a packet is currently being received and whether or not the link is up. This counter will increment along with the Rx Se quence Error Count when eith er the fiber cable is disconnected or the connector is not seated completely. Also, during driver load, this counter will increments minimally. ⢠Rx Error Count This counter indicates the number of packets receiv ed in which I_RX_ER was as serted by the PHY. In order for a packet to be counted, it must pass a ddress filtering and must be 64 bytes or greater (from <Destination Address> through <CRC>, inclusivel y) in length. If receives are not enabled, then this counter does not increment. In TBI mode, this counter increments on the reception of /V/ codes. ⢠Rx Missed Packets Count This counter indicates the number of missed packets. Packets are mi ssed when the hardware receive FIFO has insufficient space to stor e the incoming pa cket. This could be the result of having too few buffers allocated or because there is insuffici ent ban dwidth on the IO bus. These packets are also be counted in the Total Packets Received counte r as well as in Tota l Octets Received. ⢠Tx Single Collision Count This counter indicates the number of times that a transmitted packet encounte red a single collision. This counter increments only if th e device is in half-duplex mode. ⢠Tx Excessive Collisions Count This counter indicates when 16 or more collisi ons have occurred on a packet. Thi s counter increments only if the device is in half-d uplex mode. ⢠Tx Multiple Collision This counter indicates the number of times that a transmitted packet encountered more than one collision but less than 16. This counter incremen ts only if the device is in half-duplex mode. ⢠Tx Late Collisions Count This counter indicates collisions that occur after 512 bit (64 byte) tim es. This counter increments only if the device is in half-duplex mode. ⢠Tx TOTAL Collision This counter indicates the total number of collisions seen by the transmitter. This counter increm ents only if the device is in half-duplex mode. ⢠Tx DMA Underruns
Installation 40 This counter contains the number of frames that we re either not transmitted or re-transmitted due to a Tx DMA under-run. If the device is configured to retransmit on under-run, this counter may be updated multiple times for a single frame. Un der-runs occur because a lack of PCI bandwidth resulted in the device's internal transmit FIFO running dry during the tr ansmission of a frame. The adaptive algorithms within the soft ware keep this number under control. To see how the adaptive algorithm is doing, refer to the "Current Threshold" section. ⢠Tx Defer Count This counter indicates transmit defer events. A defer event occurs wh en the transmitter cannot immediately send a packet due to the medium being busy, either because another device is transmitting or the link is not up. ⢠Tx with No CRS This counter counts the number of su ccessful packe t transmissions in which the Carrier Sense input from the PHY was not asserted within one slot time of start of transmission. The PHY should as sert Carrier Sense during every transmissi on. Failure to do so may indicate that the link has failed, or the PHY has an incorrect link configuration. This counte r is not valid in TBI mode, and is only valid when the device is operating at full-duplex. ⢠Rx Sequence Error Count This counter indicates receive sequence error even ts. This counter increments along with the Rx Symbol Error Count when either the fiber cable is disconnected or the connector is not seated completely. ⢠Rx Length Error Count This counter indicates receive length error events. A length error occurs if an incoming packet is undersized (less than 64 bytes) or oversized (greater than 1522). ⢠Rx DMA Too Early This counter counts the total number of times that the device attempted to perform an Early DMA operation (by speculatively requesting the host bus prior to reception of th e entire packet) and ran out of data to DMA before the device was forced to relinquish owners hip of the bus. This condition is really not an "error" condition, bu t is nevertheless counte d for diagnostic purpos es. This statistic is valid only if Early Receives are enabled. ⢠XON Received Count This counter indicates the number of XON packe ts received. XON packets can occur when Flow Control is enabled and both the adapter and its link partner are configured to use it. ⢠XON Transmitted Count This counter indicates the number of XON packets tra nsmitted. XON packets can occur when Flow Control is enabled and both the adapter and it s link partner are configured to use it. ⢠XOFF Received Count This counter indicates the number of XOFF packe ts received. XOFF packets can occur when Flow Control is enabled and both the adapter and it s link partner are configured to use it. ⢠XOFF Transmitted Count This counter indicates the number of XOFF packets transmitted. XOFF packets can occur when Flow Control is enabled and both the adapter and it s link partner are configured to use it. ⢠Flow Control Received Unsupported Count
Installation 41 This counter indicates the number of unsupported fl ow control frames that are received. This counter increments when a flow control packet is received which does not match the adapter definition of a flow control packet. This counter is not expected to increment. ⢠64 Byte Packets Received This counter indicates the number of good packets received that are exactly 64 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠65â127 Byte Packets Received This counter indicates the number of good packe ts received that are 65â127 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠128â255 Byte Packets Re ceived This counter indicates the number of good pa ckets received that are 128â255 bytes in len gth, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠256â511 Byte Packets Re ceived This counter indicates the number of good pa ckets received that are 256â511 bytes in len gth, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠512â1023 Byte Packets Re ceived This counter indicates the number of good packe ts received that are 512â1023 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠>1024 Byte Pa ckets Received This counter indicates the number of good packe ts received that are 1024â1522 bytes in length, including CRC. Packets counted in the Missed Packe t Count are not counted here. This counter does not include received flow control packets. ⢠Good Packets Received Count (Any Length) This counter indicates the number of good packets received of any legal length. This cou nter does not include received flow control packets. ⢠Broadcast Packets Received Count This counter indicates the number of good broadcast p ackets received. ⢠Multicast Packets Received Count This counter indicates the number of good multicas t packets received. This counter does not include received flow control packets. ⢠Good Packets Transmitted Count (Any Length) This counter indicates the number of good packets transmitted of any legal le ngth. This cou nter does not include received flow control packets. ⢠Good Octets Received Count LO This counter, in conjunction with the Good Octets Received Count HI, makes up a 64-bit counter that indicates the number of good octets received. This statistic reflects the lower 32 bits of the 64-bit counter. This counter does not incl ude received flow control packets.
Installation 42 ⢠Good Octets Received Count HI This counter, in conjunction with the Good Octe ts Received Count LO, makes up a 64-bit counter that indicates the number of good octets received. Th is counter reflects the uppe r 32 bits of the 64-bit counter and is incremented each time the Good Oc tets Received Count LO has reached it maximum value and wraps to zero. This counter does not include received flow control packets. ⢠Good Octets Transmitted Count LO This counter, in conjunction with the Good Octe ts Transmitted Count HI, makes up a 64-bit counter that indicates the number of good octets transmitted. This counter reflects the lower 32 bits of the 64- bit counter. This counter does not in clude received flow control packets. ⢠Good Octets Transmitted Count HI This counter, in conjunction with the Good Octe ts Transmitted Count LO, makes up a 64-bit counter that indicates the number of good octets transmitte d. This counter reflects the upper 32 bits of the 64- bit counter and is incremented each time the G ood Octets Transmitted Count LO has reached it maximum value and wraps to zero. This counter do es not include received flow control packets. ⢠Receive No Buffers Count This counter indicates the number of times that pa ckets were received when there were no available buffers in host memory to store tho se packets. The packet is still received if there is space in the hardware FIFO. This counter does not include received flow control packets. Increasing Rx Descriptors on the command line duri ng load could result in this co unter incrementing less frequently. ⢠Receive Undersize Count This counter indicates the number of received packe ts that were less than minimum size (64 bytes) and had a valid CRC. ⢠Receive Fragment Count This counter indicates the number of received packe ts that were less than minimum size (64 bytes) but had a bad CRC. ⢠Receive Oversize Count This counter indicates the number of received pa ckets that were greater than maximum size (1522 bytes) and had a valid CRC. ⢠Receive Jabber Count This counter indicates the number of received pa ckets that were greater than maximum size (1522 bytes) but had a bad CRC. ⢠Total Octets Received LO This counter, in conjunction with the Total Octe ts Received HI, makes up a 64-bit counter that indicates the total number of octets received. This counter reflects the lower 32 bits of the 64-bit counter. ⢠Total Octets Received HI This counter, in conjunction with the Total Octe ts Received LO, makes up a 64-bit counter that indicates the total number of octets received. This counter reflects the upper 32 bits of the 64-bit counter and is incremented each time the Total Oc tets Received LO has reached it maximum value and wraps to zero. ⢠Total Octets Transmitted LO
Installation 43 This counter, in conjunction with the Total Octe ts Transmitted HI, makes up a 64-bit counter that indicates the total number of octets transmitted. This counter reflects the lower 32 bits of the 64-bit counter. ⢠Total Octets Transmitted HI This counter, in conjunction with the Total Octe ts Transmitted LO, makes up a 64-bit counter that indicates the total number of octets transmitted. This counter reflects the upper 32 bits of the 64-bit counter and is incremented each time the Total Oc tets Transmitted LO has reached it maximum value and wraps to zero. ⢠Total Packets Received This counter indicates the total nu mber of packets received. All pack ets are counted here, regardless of their length, whether they are erred, or whether they are flow contr ol packets. ⢠Total Packets Transmitted This counter indicates the total nu mber of packets transmitted. All packets transmitted will be counted here, regardless of their length or wh ether they are flow control packets. ⢠64 Byte Packets Transmitted This counter indicates the number of good packets tran smitted that are exactly 64 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠65â127 Byte Packets Transmitted This counter indicates the number of good packets tran smitted that are 65â127 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠128â255 Byte Packets Trans mitted This counter indicates the number of good packe ts tran smitted that are 128â255 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠256â511 Byte Packets Trans mitted This counter indicates the number of good packe ts tran smitted that are 256â511 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠512â1023 Byte Packets Transmitted This counter indicates the number of good packe ts transmitted that are 512â1023 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠>1024 Byte Packets Transmitted This counter indicates the number of good packets tran smitted that are 1024â1522 bytes in length, including CRC. This counter does not include transmitted flow control packets. ⢠Broadcast Packets Transmitted Count This counter indicates the number of good broadcast packets transmitted. ⢠Multicast Packets Transmitted Count This counter indicates the number of good multic ast packets transmitted. This cou nter does not include transmitted flow control packets. ⢠Rx Checksum Good Pkts This counter indicates the number of receive packets with good checks ums calculated. This counter increments only when RxTcpChecksum has been enab led and packets with good checksums are received. ⢠Rx Checksum Error Pkts
Installation 44 This counter indicates the number of receive packe ts with CRC errors. Packets less than 64 bytes are not counted in this counter. ⢠Tx Threshold This counter indicates the current value of the Early Transmit Threshold. The adapter starts transmitting the packet when the specified number of quad-words has been DMAed in from memory. ⢠Current IFS Value This counter indicates the current value of Inte r Frame Spacing between back-to-back transmits. ⢠Jumbo Frames Received This counter indicates the number of jumbo packets received (packets greater than 1514 in len gth). ⢠Jumbo Frames Transmitted This counter indicates the number of jumbo packets transmitted (pa ckets that are greater than 1514 in length). ⢠PCI-X mode (0=PCI, 1=PCI-X) This counter indicates whether the adapter is in a PCI. CPQANS custom counters ⢠Number of Tx ECBs Queued in Members This counter indicates the number of packets (Event Control Blocks) that are queued in team members for transmission. ⢠Number of Tx ECBs Queued in Unavailable ECBs Queue This counter indicates the number of transmit packets (Event Control Blocks) that are currently in the CPQANS transmit queue. ⢠Total Enqueued Tx Due to No Available ECBs This counter indicates the number of transmit packe ts that were queued du e to no available ECBs. ⢠Number of Probes Sent by All Team Members This counter indicates the number of pr obe packets sent by all team members. ⢠Number of Probes Received by All Team Members This counter indicates the number of prob e packets received by all team members. Advanced Network Services help - supported keywords ⢠AGG_SELECTION o Syntax: AGG_SELECTION=[BANDWID TH | COUNT] o Description: Sets active aggregator selection mode by bandwidth or count ⢠BALANCE_INTERVAL o Syntax: balance_interval=nnn o Description: Changes Balance interval, nnn counts 1/18 sec ⢠BALANCE_SET_DEFAULT o Syntax: balance_set_default o Description: Restores Balance interval to factory settin gs ⢠COMMIT o Syntax: commit [team=nnn]
Installation 45 o Description: Sets a certain mode. Use after binding to base drivers with Team=nnn. ⢠DELAY o Syntax: delay=nnn o Description: Delays the commit in nnn seconds ⢠FRAME o Syntax: frame=[ETHERNET_802.2 | ETHERNET _802. 3 | ETHERNET_II | ETHERNET_SNAP] ⢠HELP o Syntax: -H o Description: Displays Help o Example: CPQANS -H ⢠JOIN_INDIVIDUALS o Syntax: join_individuals=[yes | no] o Description: Joins all individual links to one aggregator ⢠LBN o Syntax: Reset LBN=nnn o Example: cpqans reset lbn=nnn o Description: Resets the logical board number (LBN) of a bound adapter. Supported only after commit command. Default value = 0 ⢠MAX_TX_QUEUE o Syntax: max_tx_queue=nnn o Description: Sets the Max number of TX ECBs queued for send ⢠MODE o Syntax: mode=[NFT | ALB | FEC | GEC | 802.3AD] ⢠NAME o Syntax: name=[any unique name] o Description: Sets a unique name ⢠PRIMARY o Syntax: primary | secondary o Description: Identifies the primary a dapter. Supported only in BIND command ⢠PROBE_BURST_SIZE o Syntax: probe_burst_size=nnn o Description: Changes number of probes to send in a retry ⢠PROBE_CHECK_INTERVAL o Syntax: probe_check_interval=nnn o Description: Changes probes check interval ⢠PROBE_RECHECK_INTERVAL o Syntax: probe_recheck_interval=nnn o Description: Changes probe retries check interval
Installation 46 ⢠PROBE_RETRY_COUNT o Syntax: probe_retry_count=nnn o Description: Changes probes retry count ⢠PROBE_SEND_INTERVAL o Syntax: probe_send_interval=nnn o Description: Changes probes send interval, nnn counts 1/18 sec ⢠PROBE_SET_DEFAULT o Syntax: probe_set_default o Description: Restores probes settings to f actory settings ⢠PROBES o Syntax: probes=[on | off] o Description: Enables /disables probes ⢠PROBES o Syntax: probes=[BROADCAST|MULTICAST] o Description: Changes probes addressing ⢠RECOMMIT o Syntax: recommit [team=nnn] o Description: Resets a certai n mode. Use after hot binding to base drivers with Team=nnn ⢠REMOVETEAM o Syntax: removeteam [team=nnn] o Description: Removes a team. Use after hot binding to base drivers with Team=nnn ⢠REMOVEVLANID o Syntax: RemoveVlanID=nnn o Description: Removes the selected VLAN ⢠RESET o Syntax: RESET LBN=nnn o Description: Supported on ly after Commit command ⢠SECONDARY o Syntax: primary | secondary o Description: Identifies the secondary a dapter. Supported only in BIND command. ⢠SMPMODE o Syntax: SMPMODE=[SMP | NONSMP] o Description: Enables /disables SMP aware ⢠STATUS o Syntax: status [team=nnn] o Description: Prints CPQANS status ⢠TEAM o Syntax: team=nnnn
Installation 47 o Description: Identifies the team. nnnn=DecimalNumber. ⢠TX_ECBS_TO _USE o Syntax: tx_ecbs_to_use=nnn o Description: Sets number of TX EC Bs to allocate per virtual adapter ⢠VLANID o Syntax: VlanID=nnn o Description: Sets team to VLAN mode. Creates MLID edge
Firmware upgrade utility 48 F ir m w ar e upgr ade utilit y In this section Windows ut ility ................................................................................................................ ...................... 48 c-Class BladeSys tem utility .................................................................................................... ................... 54 Linux utility .................................................................................................................. .......................... 54 DOS utility .................................................................................................................... ......................... 59 Windows utility These notes describe how to us e the Gigabit Server Adapter Firmware Upgrade Utility to flash server adapter firmware. Flashing consists of using software to replace the current server adapter firmware on a target server with a new firmware image. Server adapter firmware is typically flashed for the following reasons: ⢠To support new features ⢠To correct problems in a previous server adapter firmware version Customers should up grade their servers to the latest server adapter firmware version: ⢠When HP releases a new serv er adapter firmware version ⢠When installing a new server adapter ⢠After purchasing a new server ⢠When installing a SoftPaq that specifies th e need for the latest version of firmware The Gigabit Server Adapter Firmwa re Upgrade Utility automatically checks for driver, hardware, and operating system dependencies, and installs only the correct s erver adapter firmware upgrades required by each target server. Supported operating systems The upgrade utility support s the following operating system s: ⢠Microsoft® Windows® 2 000 Server ⢠Microsoft® Windows® 2000 Advanced Server ⢠Microsoft® Windows® 2000 Data Center ⢠Microsoft® Windows® 2 000 Terminal Server ⢠Microsoft® Windows 20 03 Server ⢠Microsoft® Windows 20 03 (32-bit) Enterprise Serv er ⢠Microsoft® Windows 20 03 (32-bit) Data Center ⢠Microsoft® Windows 20 03 (x64) Enterprise Server Supported Gigabit Server Adapters The upgrade utility provides support for t he following adapters:
Firmware upgrade utility 49 Standup adapters: ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC364T PCI Express Quad Port Gigabit Server Adapter ⢠HP NC370T PCI-X Multifunction Gigabit Server Adapter ⢠HP NC370F PCI-X Multifun ctio n Gigabit Server Adapter ⢠HP NC373F PCI Express Multifunc tion Gigabit Server Adapter ⢠HP NC373T PCI Express Multifunction Gigabit Serv er Adapter ⢠HP NC380T PCI Express Dual Port Multifunction Giga bit Server Adapter ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC320i Integrated PCI Express G igabit Server Adapter ⢠HP NC324i Integrated Dual Port PC I Express Gigabit Server Adapter ⢠HP NC325i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC326i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC370i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC371i Integrated PCI-X Mult ifunction Gigabit Server Adapter ⢠HP NC373i PCI Express Multifunction Gigabit Server Adapter ⢠HP NC7760 PCI-X Gigabit Server Adapter ⢠HP NC7761 PCI Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gigabit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter Mezzanine Adapters: ⢠HP NC320m PCI Express Dual Port 1Gb Server Adapter ⢠HP NC325m PCI Express Quad Port 1Gb Server Ad apter for c-Class BladeSystem ⢠HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem ⢠HP NC373m PCI Express Dual Port Multifunction Giga bit Server Adapter for c-Class BladeSyst em ⢠HP NC374m PCI Express Dual Port Multifunction Giga bit Server Adapter IMPORTANT: For each server adapter in the target server, the driver for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated.
Firmware upgrade utility 50 IMPORTANT: Before updating NC6770 or NC7770 server adapter s, move jumper J1 to the Enable position. After updating each server adapter, move jumper J1 to the Disable position. NOTE: NC6770, NC7770, and NC7780 server adapters require about 15 mi nutes each to update both the PXE and Boot Code. NOTE: To update the PXE code for embedded and mezzanine adapters, update the system BIOS. Drivers For each server adapter in the target server, the driv er for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. Installing the firmware utility The HP Network Adapter Online Firmware Upgrade Ut ility is located in the \apps\fwupgrade\windo ws directory obtained by extracting the HP SoftPaq fi le. Complete the following step s to download and extract the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. The upgrade utility installation is launched by clicking the Smart Component (CPXXXXXX.exe) file (interactive mode) or by using the command line pr ompt. The Smart Component for the upgrade utility is launched using the command line prompt to allow yo u to pass arguments to alter installation behavior. Command line syntax The general command line syntax for the Smart Component installation includes: CPXXXXXX [/S[ILENT]] [/F[ORCE]] [/R[EBOOT]] [/H[ELP]] [/?] CPXXXXXX is the command that launches the Smart Compon ent on the target server. The six X's represent the serialized number assigned to the component. Command line arguments The Smart Component recognizes the following comm and line arguments. Smart Components are not case-sensitive. NOTE: All arguments and information en closed in brackets are optional.
Firmware upgrade utility 51 Command line argument Description /S[ILENT] Supresses console messages. If this argument is omitted from the command line, console messages are displayed. /F[ORCE] Specifies whether an installation will be forced, regardless of the server adapter firmware version that is detected on the target server. The argument changes the Smart Component installation in the following ways: ⢠If the server adapter firmware on the target server is current, the firmware reinstalls itself, and the installed version number remains the same. ⢠If a newer version of the server adapter firmware is already installed on the target server, the component installs itself and downgrades the server adapter firmware to the older version number. If this argument is omitted from the command line, the installation is not forced. /R[EBOOT] Specifies whether the target server will reboot if the Smart Component requests a reboot. If this argument is omitted from the command line and the Smart Component requests a reboot, the server must be rebooted manually for th e server adapter firmware upgrade to take effect. The Windows server adapter upgrade component will NOT request a reboot. /H[ELP] Displays command line Help information. /? Displays command line Help information. Thi s argument is identical to the /H[ELP] argument. Command Line Examples The following are examples of how to use the CPXXXXXX.exe command. Command line input Result CPXXXXXX Deploys the CPXXXXXX.EX E Smart Component on the target server using the defaults of the component. CPXXXXXX /F Deploys the CPXXXXXX. E XE Smart Component on the target server, forcing the installation of the server adapter firmware. CPXXXXXX /S Deploys the CPXXXXXX. E XE Smart Component on the target server using the defaults of the component. All console messages are suppressed. CPXXXXXX /S /F /R Deploys the CPXXXXXX.EXE Smart Component on the target server, forcing the installation of the server adapter firmware and allowing the server to reboot automatically. All console messages are suppressed. Common installation log file
Firmware upgrade utility 52 The upgrade installation activity (including erro rs) is written by the Smart Component to a common installation log file called CPQSETUP.LOG in the \CPQSYSTEM\LOG subdirector y on the target server. Information regarding subsequent installation activity is appended to the same log file, which provides a chronological history of all component inst allation activity on the target server. Using the NicFwUpg.exe command line interface It is possible to launch the HP Online Firmware Up grade Utility by clicking the executable file d uring installation or by using the command line interface. The following in formation describes how to launch the utility using the command line interface. Command line syntax The utility supports the follo wing commands: Command Line Input Result nicfwupg.exe -c Displays the cu rrent firmware version of all HP NC-Series Broadcom adapters in the system nicfwupg.exe -l Displays the latest version of HP NC-Series Broadcom adapters in the system as determined by versions shipped with the component. nicfwupg.exe -a Retuns the saved image version of HP NC- Series Broadcom adapter s in the system nicfwupg.exe -s Suppresses console output nicfwupg.exe -f Forces downgr ades or restores as needed nicfwupg.exe -p Requests a restore on all adapters nicfwupg.exe -e [MAC Address] Requ ests a restore on one adapter. Supply the MAC address for the adapter. nicfwupg.exe -o [MAC Address] Performs an upgrade of boot code for one adapter. Supply the MAC address for the adapter. nicfwupg.exe -x [MAC Address] Performs an upgrade of PXE for one adapter. Supply the MAC address for the adapter. nicfwupg.exe -b [MAC Address] Performs an upgrade of CLP for one adapter. Supply the MAC address for the adapter. nicfwupg.exe -i [filename] Performs an upgrade and stores the image file to be used during a restore. Supply the filename. Using the nicfwup_mf.exe command line interface It is possible to launch the HP Online Firmware Up grade Utility for Multifunction adapters by clicking the executable file during installation or by using the command line interface. The fo llowing information describes how to launch the utility for Multifunction network adapters using the command line interface. Command line syntax The utility supports the follo wing commands:
Firmware upgrade utility 53 Command Line Input Result nicfwupg_mf.exe -c Displays the current firmware version of all HP NC-Series Broadcom adapters in the system nicfwupg_mf.exe -l Displays the latest version of HP NC-Series Broadcom adapters in the system as determined by versions shipped with the component. nicfwupg_mf.exe -a Retuns the saved image version of HP NC- Series Broadcom adapter s in the system nicfwupg_mf.exe -s Suppresses console output nicfwupg_mf.exe -f Forces down grades or restores as needed nicfwupg_mf.exe -p Requests a restore on all adapters nicfwupg_mf.exe -e [MAC Address] Requests a restore on one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -o [MAC Address] Performs an upgrade of boot code for one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -x [MAC Address] Performs an upgrade of PXE for one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -b [MAC Address] Performs an upgrade of CLP for one adapter. Supply the MAC address for the adapter. nicfwupg_mf.exe -i [filename] Perfor ms an upgrade and stores the image file to be used during a restore. Supply the filename. Return codes When the upgrade utility has fini shed running, the component reports a return code to the operating system or calling application. These return codes ar e processed by the %error level% operating system variable and are used to determine the stat us of a firmwa re upgrade proc edure. The following summarizes the ret urn codes for the Sm art Component installer and HP Online Firmware Upgrade Utility commands. Error level Meaning 0 Upgrade component failed to install 1 Upgrade component installed successfully 3 Upgrade component did not de tect any supported network adapters 4 Firmware upgrade was recommended, but the upgrade component was instructed to not update the firmware 5 All firmware versions are current and an upgrade is not recommended 6 At least one network adapter has firmware installed that is newer than the available firmware.
Firmware upgrade utility 54 c-Class BladeSystem utility The HP Gigabit Server Adapter Firmware Upgrade Ut ility for c-Class BladeSystem uses two separate utilities (ccfwupg1.bat and ccfwupg2.bat) to upgr ade firmware on the c-Cl ass BladeSystem adapters. The ccfwupg1.bat command upgrades the Boot Code an d Option ROM fi rmware image on the following c-Class Multifunction Gigabit Server Adapters: ⢠HP NC373i Multifunction Gigabit Server Adapter ⢠HP NC373m Multifunction Gigabit Server Adapter The ccfwupg2.bat command upgrades the Boot Code an d Option ROM fi rmware image on the following c-Class Gigabit Server Adapters: ⢠HP NC326i PCIe Dual Port Gigabit Server Adapter ⢠HP NC326m PCIe Dual Port Gigabit Server Adapter ⢠HP NC325m PCIe Quad Port Gigabit Server Adapter Upgrading firmware To upgrade the boot code and Option ROM im age on the adapter, complete the following: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. 8. Copy all files and folders from the \apps\ccfwupg\ccfwupg1 dir ectory or the \apps\ccfwupg\ccfwupg2 directory to a bootable diskette or USB drive. 9. Boot to DOS using the diskette for your a dapter. 10. Type ccfwupg1.bat or ccfwupg2.bat and press the Enter key. A confirmation message displays when the upgrade is complete. Linux utility The following information describes how to use th e Linux Gigabit Server A dapter Firmware Upgrade Utility to flash server adapter firmware. Supported operating systems The firmware upgrade utility support s the following Linux operating syst ems: Linux 32 ⢠Red Hat Enterprise Linux 4 ⢠Red Hat Enterprise Linux 5 ⢠SUSE Linux Enterprise Server 10 x86
Firmware upgrade utility 55 ⢠SUSE Linux Enterprise Server 9 x86 Linux 64 ⢠Red Hat Enterprise Linux 4 for AMD64 and Intel EM64T ⢠Red Hat Enterprise Linux 5 for AMD64 and Intel EM64T ⢠SUSE Linux Enterprise Server 10 for AMD64 and Intel EM64T ⢠SUSE Linux Enterprise Server 9 for AMD64 and Intel EM64T For a complete listing refer to the Linux Readme fil e located in the HP SoftPaq file.Complete the follo wing steps to download and extract the SoftPaq file. 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable with a file name based on the SoftPaq number. 7. The hpnicfwupg Linux online firmware upgrade utilit y is located in the apps \fwupgrade\linux folder. Supported Gigabit Server Adapters The upgrade utility provides support for t he following adapters: Standup adapters: ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC370T PCI-X Multifunction Gigabit Server Adapter ⢠HP NC370F PCI-X Multifun ctio n Gigabit Server Adapter ⢠HP NC373F PCI Express Multifunc tion Gigabit Server Adapter ⢠HP NC373T PCI Express Multifunction Gigabit Serv er Adapter ⢠HP NC380T PCI Express Dual Port Multifunction ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC320i PCI Express Gigabit Server Adapter ⢠HP NC324i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC326i PCI Express Dual Port Gigabit Server Adapter ⢠HP NC370i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC371i PCI-X Multifunction Gigabit S erver Adapter ⢠HP NC373i PCI Express Multifunction Gigabit Server Adapter
Firmware upgrade utility 56 ⢠HP NC7760 PCI-X Gigabit Server Adapter ⢠HP NC7761 PCI Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gig abit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter Mezzanine Adapters: ⢠HP NC320m PCI Express Dual Port Gigabit Server Ad apter ⢠HP NC325m PCI Express Quad Port 1Gb Server Ad apter for c-Class BladeSystem ⢠HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem ⢠HP NC373m PCI Express Dual Port Multifunction Giga bit Server Adapter for c-Class BladeSyst em ⢠HP NC374m PCI Express Dual Port Multifunction Giga bit Server Adapter IMPORTANT: For each server adapter in the target server, the driver for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. IMPORTANT: Before updating NC6770 or NC7770 server adapter s, move jumper J1 to the Enable position. After updating each server adapter, move jumper J1 to the Disable position. NOTE: NC6770, NC7770, and NC7780 server adapters require about 15 mi nutes each to update both the PXE and Boot Code. NOTE: To update the PXE code for embedded and mezzanine adapters, update the system BIOS. Drivers For each server adapter in the target server, the driv er for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. To download the latest Linux drivers: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Open the Linux folder to select drivers. Upgrading firmware This firmware upgrade utility updat es Boot code, PXE code, and CLP upgrades for c-Class BladeSystem (NC325m, NC326m and NC373m) serv er adapter images. The utility is released in RPM fo rmat. The file
Firmware upgrade utility 57 names for the packages are hpnicfwupg-<vers ion>.i38 6.rpm for Linux 32 and hpnicfwupg- <version>.x86_64.rpm for Linux 64. The inst ructions b elow describe the x32 version. 1. Before the RPM package is installed, check for th e existence of a previous version by issuing the rpm -q hpnicfwupg command. 2. If a previous version is found, remove it using the rpm -e hpnicfwupg command. 3. Install the RPM package using the hpnicfwupgâ<version>.rpm command. 4. Launch the adapter interface by entering ifup ethx or ifconfig ethx up . 5. Launch the upgrade uti lity in interactive mode b y entering hpnicfwupg . 6. Before the upgrade is attempted, a copy of the firmware image is saved. If the firmware upgrade fails due to an unrecoverable error, you can restor e your original fi rmware interactively by entering /usr/sbin/hpnicfwupg -p . If errors have occurred, use the hp nicfwupg utility to restore the server adapter fir mware from a backup copy. Command line syntax Use the following command line syntax. To Use command Verify previous version of hpnicfwupg package exists # rpm -q hpnicfwupg Remove an old version of hpnicfwupg package # rpm âe hpnicfwupg Install the RPM package # rpm -ivh hpnicfwupg-<version>.rpm Launch the firmware upgrade utility # user/sbin/hpnicfwupg Restore original firmware if errors occur # /usr/sbin/hpnicfwupg -p Uninstall the RPM package # rpm -e hpnicfwupg Command line arguments The upgrade utility recognizes the foll owing command line arguments. Command line argument Description ây Yes. Executes the utility with no user interaction. Default action is implied for all prompts. Messages are displayed to the screen and to the log. If ây is used with âs, the behavior is the same as if âs had been specified. âs Silent. Executes the utility silently with no user interaction. Default action is implied for all prompts. Messages are logged but are not displayed o n the screen. If the ây option is used with âs, ây is ignored. âf Force. Forces the firmwa re upgrade even if the firmware installed on the system is newer than the firmware in the RPM. This opti on is only valid with âs or ây. If neither are specified, the option is ignored.
Firmware upgrade utility 58 Command line argument Description âp Previous. Interactively restores previous firmware images that were automatically backed up prior to the firmware update. This option is useful for recovering from firmware upgrade failures. This option can be used with any other option. âh Help. Displays the usage mess age. If this option is used with any other option, the other options are ig nored. âc Displays server adapters and current firmware version without performing an updates. This option cannot be used with any other option. Restoring firmware image from backup Before the server adapter firmware image on the server is updated, the current image is saved in a file called <MAC_ADDRESS>.bc<Boot Code version>.p<PXE version>.bin . The backup file is saved in the /var/opt/hp/nicfw_backup directory of the target server. Use the hpnicfwupg command with the âp option to restore the original firmware image after an upgrade or after an upgrade failure. This option ca n only be used interactively. The firmware backup directory is examined for saved images and a me ssage is displayed showing the curr ent firmware versions and the versions from the saved image. You will be prompted be fore the restore is attempted. If more than one saved image is found for a particula r server adapter, you will be prompted for each image. The following is a sample screen sh ot of a typical restore. [root@rdple root]# /usr/sbin/hpnicfwupg -p Found HP NC7760 Gigabit Server Adapter MAC: 000BCD5035DE *** WARNING *** - Installed firmware is the same version as the selected firmware. Restore Boot Code 2.35 to 2.35 - y/n/q (n): q Firmware upgrade cancelled by User. Firmware upgrade log file The firmware upgrade installation stat us (including e rrors) is written by the Li nux server adapter upgrade package to a firmware upgrade log file called nicfwupg.log in the /var/opt/hp subdirectory on the target server. Information regarding subsequent inst allation activity is append ed to the same log file, providing a chronological history of all server adapte r firmware upgrade activity on the target server. Return codes When the Linux server adapter firmware upgrade packa ge has finished running, it report s the following return codes to the operating sy stem or calling application. Error level Meaning 0 Hpnicfwupg installed all selected firmware successfully. For options âc and âi , 0 means the information was displayed correctly.
Firmware upgrade utility 59 Error level Meaning 2 Hpnicfwupg did not attempt firmware upgrade because the installed firmware was up-to-date, the target hardware was not present, the command was invalid, the user responded no to all prompts, or the user selected qu it in response to a prompt. For options âc and âi, a 2 means that the information could not be displayed. 3 Hpnicfwupg attempted to upgr ade firmware but one or more upgrades failed due to an unrecoverable error. This return code is used even if some of the server adapters were upgraded successfully. DOS utility These notes describe how to us e the Gigabit Server Adapter Firmware Upgrade Utility to flash server adapter firmware. Flashing consists of using software to replace the current server adapter firmware on a target server with a new firmware image. Server adapter firmware is typically flashed for the following reasons: ⢠To support new features ⢠To correct problems in a previous server adapter firmware version Customers should up grade their servers to th e latest server adapter firmware ve rsion: ⢠When HP releases a new serv er adapter firmware version ⢠When installing a new server adapter ⢠After purchasing a new server ⢠When installing a SoftPaq that specifies th e need for the latest version of firmware The Gigabit Server Adapter Firmware Upgrade Utilit y automatically checks for driver, hardware, and operating system dependencies, and installs only the correct server adapter firmware upgrades required by each target server. Supported Gigabit Server Adapters The upgrade utility provides support for t he following adapters: Standup adapters: ⢠HP NC150T PCI 4-port Gigabit Combo Switch Adapter ⢠HP NC320T PCI Express Gigabit Server Adapter ⢠HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port ⢠HP NC6770 PCI-X Gigabit Server Adapter ⢠HP NC7770 PCI-X Gigabit Server Adapter ⢠HP NC7771 PCI-X Gigabit Server Adapter Embedded adapters: ⢠HP NC7760 PCI-X Gigabit Server Adapter
Firmware upgrade utility 60 ⢠HP NC7761 PCI Gigabit Server Adapter ⢠HP NC7780 Gigabit Server Adapter ⢠HP NC7781 PCI-X Gigabit Server Adapter ⢠HP NC7782 Dual Port PCI-X Gigabit Server Adapter IMPORTANT: For each server adapter in the target server, the driver for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. IMPORTANT: Before updating NC6770 or NC7770 server adapter s, move jumper J1 to the Enable position. After updating each server adapter, move jumper J1 to the Disable position. NOTE: NC6770, NC7770, and NC7780 server adapters require about 15 mi nutes each to update both the PXE and Boot Code. NOTE: To update the PXE code for embedde d adapters update the system BIOS. Drivers For each server adapter in the target server, the driv er for the adapter must be installed and enabled for the server adapter to be upgraded. If the driver is unable to communicate with the server adapter, the server adapter cannot be updated. Upgrading firmware 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. 8. Click the SoftPaq file (SP31728) located in the \apps\fwupgrade\dos directo ry. 9. Execute the downloaded file and follow the on-scr een instructions. After the files ha ve been unpacked, you may delete the Soft Paq file that you downloaded. 10. Copy these files to a bootable DOS diskette. 7781V235 BIN 150TV318 BIN 6770V219 BIN 7760V237 BIN 7761V318 BIN 7770V219 BIN 7771V235 BIN 7780V219 BIN 1020V318 BIN
Firmware upgrade utility 61 7782327B BIN NICFWUPG BAT PT5704C2 36 PXE_V322 BIN Q57UDIAG EXE SP29044 TXT PXE_V778 BIN PXE_V839 BIN 320TV355 BIN 11. Boot the system using the DOS diskette. 12. Run NICFWUPG.BAT from a command prompt.
Adapter configurations 62 A dapte r conf igur ations In this section HP NC-Series Br oadcom adapters ................................................................................................. ........... 62 HP NC-Series Multif unction adapters ............................................................................................ ............ 66 HP NC-Series In tel adapters.................................................................................................... ................. 69 Windows ........................................................................................................................ ...................... 91 Novell Ne tware ................................................................................................................. .................. 110 Solaris ........................................................................................................................ ........................ 122 HP NC-Series Broadcom adapters Keywords for Q57 NDIS2 driver This document describes the keywords for th e Q57 NDIS 2 driver version 7.65 (Q57.DOS). The following custom keywords are used in PROTOCOL.INI: ⢠BUSNUM ⢠DEVNUM ⢠FUNCNUM or PORTNUM ⢠LineSpeed ⢠Duplex ⢠NodeAddress ⢠FixCheckSumOff Descriptions ⢠BUSNUM and DEVNUMâThese two keywords are used concurrently and have been included for manufacturin g purposes. Do not use the m unless yo u are familiar with PCI device configuration. These two keywords are needed if multiple NC 1020, NC67xx, and/or NC77xx boards are in a system and a specific adapter (or ad apters) needs to be loaded in sp ecific order. BU SNUM specifies the PCI bus number, and DEVNUM specifies the devi ce number giv en to the board when configured by the PCI BIOS. The values are in decimal. ⢠LineSpeedâThis decimal parameter, 10 or 100, spec i fies the speed of the network connection. According to IEEE specifications, line speed of 100 0 cannot be forced and its only achievable by auto negotiation. ⢠DuplexâThis string parameter, HALF o r FULL, specif ies duplex mode on the Ethern et controller. The LineSpeed parameter must be set when this keyword is used. If neither the Duple x nor the LineSpeed parameters are specified the Ethernet controlle r will default to autonegotiate mode. ⢠NodeAddressâThis string parameter sp ecifies the network address us ed by the Ethernet controller. If Multicast Address or Broad cast Address was spec ified , the default MAC Address will be used.
Adapter configurations 63 ⢠FixCheckSumOffâThis strin g parameter turns off the driver work-around for the T CP/IP stack to recognize the 1's complimented version of the checksum. The following are examples of the use of these keyword s in the protocol.ini. [Q57] DriverName = "Q57$" BusNum = 3 DevNum = 14 PortNum = 2 LineSpeed = 100 Duplex = Full NodeAddress = "001020304050" To add more adapters (up to four), repeat the entry below, where n can be from 1 to 3. [Q57_n] DriverName = "Q57$" BusNum = DevNum = The following examples use additional entries to lo ad more than one adapter in the protocol.ini. ⢠[Q57] DriverName = "Q57$" BusNum = 3 DevNum = 10 ⢠[Q57_1] DriverName = "Q57$" BusNum = 3 DevNum = 11 ⢠[Q57_2] DriverName = "Q57$" BusNum = 3 DevNum = 12 ⢠[Q57_3] DriverName = "Q57$" BusNum = 3 DevNum = 13 Keywords for B06 NDIS2 driver This document describes the keywords fo r the B06 BNDIS 2 driver (bxnd20x.DOS). The following keywords are supported through the PROTOCOL.INI fil e. If a keyword is not present in the PROTOCOL.INI, the default value is chosen.
Adapter configurations 64 ⢠DRIVERNAME o Possible value: BXND20X$ o This field is case sensitive ⢠LINESPEED o Possible values: 10, 100, or 1000 o Default: AUTONEG ⢠BUSNUM o Possible values: the bus slot numb er the NC37xx/N C380x device is on o Default: the bus slot number occupied by one of NC37xx/NC380x devices in the system ⢠DEVNUM o Possible value: the de vice number the NC37xx/NC380x device is on o Default: the device number occupied by one of NC37xx/NC380x devices in the system ⢠FUNCNUM o Possible values: the fu nction number the NC37x/NC380x device is on o Default: the function number occupied by one of NC37x/NC380x devices in the system ⢠DUPLEX o Possible values: half or full (LineSpeed is required when this keywor d is used) o Default: full ⢠NODEADDRESS o Possible values: 6 hex byte s looking like "0A0B0C0D0E0F" o Default: the MAC address for the selected NC37xx/NC380x device Known limitations: For Windows Server 2003 or Window Server 2003 x 64 se lf-installation, a smartdrv driver needs to be loaded. Otherwise, bad performance or even incorrect I/O write errors might occur. PXE configuration: HP NC-series Broadcom and Multifunction adapters NOTE: PXE is not supported on the NC150T adapter. The Pre-boot Execution Environment (PXE) of HP adap ters has been upgraded to allow PXE functions to work with the HP ProLiant servers. The HP NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters have PXE support. PXE is a software protocol that allows your networked computer to boot with the images provided by remote servers across the network. PXE operates in a client/server environment. A network consists of one or more boot serv ers that provide boot images to multiple computers through the network. The PXE driver complies with the PXE-2.1 specification. The PXE driver is released with both mo nolithic and split binary images. This feature provides flexibi l ity to users in different environments in whic h the motherboard may or may not have built-in base-code.
Adapter configurations 65 HP PXE implementation is supported in the followin g environments: ⢠Linux Red Hat PXE Server. HP PXE clients are able to remote-boot and utilize network resources (for example, NFS mount) and to do Linux installation. In the case of a remote boot, the Linux u niversal driver binds seamlessly with the HP UNDI (Unive rsal Network Driver Interface) and provides a network interface in the Linux remote-booted client en vironment. ⢠Microsoft® Windows® 2000, Windo ws 2003, and Windows 2003x64 Remote Installation Services (RIS). HP PXE clients are able to install from the network. ⢠Intel APITEST. The HP PXE driver passes all API compliance test suites. ⢠DOS UNDI. Universal NDIS seamlessly binds with HP UNDI to provide NDIS2 interface to the upper- layer protocol stack. This allows computers to co nnect to network resources in a DOS environment. ⢠3Com® Boot Server. Client setup Setting up PXE client for NC32xx, NC37xx, NC 380x, NC1020, NC6770, and NC77xx Gig abit Server Adapters involves enabling/disablin g the PXE feature and setting up the BIOS for the boot order. Boot method (NC32xx, NC37xx , NC380x, NC1020, NC6770 , NC7770, and NC7771 only) By default, PXE detects if the BIOS supports BBS (BIOS Boot Specification). If the BIOS supports BBS, then PXE uses BBS as the boot method. If the BIOS does not support BBS, then PXE uses Int18h instea d. However, for some machines equipped with an old BI OS, this process may not work. In case of an old BIOS, users must force PXE to use Int18h, Int19h, or BBS as the boot method. To specify the boot method, press the following keys wh en you see the startup banner. Broadcom Gigabit Ethernet Boot Agent 2.2.6 Copyright (c) 2000 Broadcom Corporation All rights reserved ⢠For Int18h: Press the Ctrl F8 keys ⢠For Int19h: Press the Ctrl F9 keys ⢠For BBS: Press the Ctrl F10 keys Setup BIOS To boot from the network with PXE, make PXE the first bootable device under BIOS. This procedure depends on server BIOS implementation. Refer to the user manual for the server for additional information. Server setup Refer to the following instr uctions for your operating system. Windows® 2000 The current version of Windows® 2000 does not include a network driver for th e NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters. To perform remote installation with PXE, include a network driver for these adapters as part of the client's installation image on the server. Refer to Microsoft® Article ID Q246184, "How to Add Third-Party OEM Network Adap ters to RIS Installations." DOS UNDI/APITEST
Adapter configurations 66 To boot in DOS and connect to a network for the DO S environment, download the Intel® PXE PDK from the Intel® web site. This PXE PDK comes with a TFTP/ProxyDHCP/Boot server. The PXE PDK can be downloaded from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). Red Hat Linux Red Hat Linux 7.2 (or higher) distribution has PXE serv er support. It allows users to perform a complete Linux installation over the network. Distribution also comes with boot imagesâboo t kernel (vmlinux) and initial ram disk (initrd). These two images can be found in the HP SoftPaq file. ⢠/misc/src/trees/boot/vmlinux ⢠/misc/src/trees/initrd-netw ork.img Refer to Red Hat documentation for in structio ns on installing PXE ser ver on Linux. Initrd.img, distributed with Red Ha t 7.2, does not have a Linux netw ork driver for the NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC32 6m, NC10 20, NC67xx, and NC77xx Gigabit Server Adapters. ⢠/images/pxeboot/vmlinux ⢠/images/pxeboot/initrd.img Intel® has also created patches to the Linux PXE serv er code to allow clients to perform a remote boot. You can download these patches from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). A remote boot does not require a standard Linux network driver for t he HP NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC326m, NC102 0, NC67 xx, and NC77xx Gigabit Server Adapters. After the PXE client downloads the Linux kernel and in itial ram disk, the Linux universal driv er, that came with the Linux distribution, will bind with the UNDI code of the PXE to form a Linux network driver. HP NC-Series Multifunction adapters Accelerated iSCSI devices The iSCSI protocol transports block-level storage requests over TCP connection s. Using the iSCSI protocol, systems can connect to rem ote storage and use it as a physical disk (although the remote storage provider, or "target," may actually be providing "virtual" physical disks). HP Multifunction Gigabit Server Adapters can be en abled as virtual iSCSI host bus adapters (HBAs), which provide the ability to run iSCSI protocol over offloaded TCP connections . By offloading the TCP connection processing to a Multifunction Server Adapt er the CPU utilization is reduced, which provides additional processing capacity for applications. iSCSI is enabled and network settings are config ured for the iSCSI device through the Network Configuration Utility (NCU). Configuration is limited to IP configuration (including DHCP support), user- specified MAC address, and VL AN ID (only one allowed for the iSCSI device). Prerequisites To use the Multifunction Server Adapter iSCSI accelera ting feature, Microsoft® iS CSI Software Initiator version 2.0 or later must be installed on your sy stem . This software is required to configure the iS CSI protocol, including discovery portals, iSCSI targets, authentication, and so on. The command line utility iscsicli.exe , which is part of Microsoft iSCSI Software In itiator can also be used to configure iSCSI
Adapter configurations 67 protocol. See the Microsoft website ( http://www.microsoft.com ) to d ownload the Microsoft iSCSI Software Initiator installation wizard software an d to obtain installa tion and usage information. The accelerating iSCSI feature comes standard wi th standup HP Multifunction Server Adapters. Accelerating iSCSI capabili ties are available for embedded HP Mu ltifunction Gigabit Server Adapters through the purchase and installation of an HP ProLiant Essentials Accelerated iSCSI Pack license (" Accelerated iSCSI for embedded HP Multifunction server adapters " on page 153 ). Configuring advanced settings parameters If iSCSI is enabled on a Multifunction adapter port, you can adjust the SRB (iSC SI Request Block) Timeout Delta setting and create an iSCSI VLAN for the adap ter. Increasing this value may help prevent I/O timeouts if your network is highly congested and packets are being dropped frequently. iSCSI adapter information includes the following: o Current MAC Address. The current MAC address for the iSCSI device. o iSCSI Driver Name. The name of the driver for the iSCSI device o iSCSI Driver Version. The version of the driver for the iSCSI device. o Default Gateway. The default route for the iSCSI device. o Subnet Mask. The subnet mask (dotted decimal) for the iSCSI device. o iSCSI VLAN Id. The identification number (if any) for the virtual iSCSI VLAN device. o iSCSI Packet Priority. The packet priority for the iSCSI device. o SRB Timeout Delta. The increase to the timeout value for I/O operat ions. o Page File Created. Indicates whether or not the Windows page file was created on an iSCSI drive that is attached to the adapter. The values are Yes and No. o DHCP (Dynamic Host Configuration Protocol). Enabled indicates that DHCP is used to acquire IP configuration; Disabled indicates that the stat ically configured IP conf iguration is used. o IP Address. The IPv4 address (dotted decimal) for the iSCSI device. iSCSI boot configuration: Multifunction adapters HP iSCSI boot is currently supported on Linux operating syst ems only. The HP iSCSI boot feature allows you to b oot from a remote disk (known as the iSCSI target) on a Storage Area Network (SAN) without having to directly attach a boot disk. Booting from a remote disk on a SAN allows you to centralize the boot process and consolidate equipment resources. Unlike other implem entations, iS CSI boot does not require a separate DHCP server or a PXE server. For more information about HP iSCSI Boot refer to the HP iSCSI Boot for Linux User Guide . PXE configuration: HP NC-series Broadcom and Multifunction adapters NOTE: PXE is not supported on the NC150T adapter. The Pre-boot Execution Environment (PXE) of HP adap ters has been upgraded to allow PXE functions to work with the HP ProLiant servers.
Adapter configurations 68 The HP NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters have PXE support. PXE is a software protocol that allows your networked computer to boot with the images provided by remote servers across the network. PXE operates in a client/server environment. A network consists of one or more boot serv ers that provide boot images to multiple computers through the network. The PXE driver complies with the PXE-2.1 specification. The PXE driver is released with both mo nolithic and split binary images. This feature provides flexibi l ity to users in different environments in whic h the motherboard may or may not have built-in base-code. HP PXE implementation is supported in the followin g environments: ⢠Linux Red Hat PXE Server. HP PXE clients are able to remote-boot and utilize network resources (for example, NFS mount) and to do Linux installation. In the case of a remote boot, the Linux u niversal driver binds seamlessly with the HP UNDI (Unive rsal Network Driver Interface) and provides a network interface in the Linux remote-booted client en vironment. ⢠Microsoft® Windows® 2 000, Windows 2003, and Windows 2003x64 Remote Installation Services (RIS). HP PXE clients are able to install from the network. ⢠Intel APITEST. The HP PXE driver pas ses all API compliance test suites. ⢠DOS UNDI. Universal NDIS seamlessly binds with HP UNDI to provide NDIS2 interface to the upper- layer protocol stack. This allows computers to co nnect to network resources in a DOS environment. ⢠3Com® Boot Server. Client setup Setting up PXE client for NC32xx, NC37xx, NC 380x, NC1020, NC6770, and NC77xx Gig abit Server Adapters involves enabling/disablin g the PXE feature and setting up the BIOS for the boot order. Boot method (NC32xx, NC37xx , NC380x, NC1020, NC6770 , NC7770, and NC7771 only) By default, PXE detects if the BIOS supports BBS (BIO S Boot Specification). If the BIOS supports BBS, then PXE uses BBS as the boot method. If the BIOS does not support BBS, then PXE uses Int18h instea d. However, for some machines equipped with an old BI OS, this process may not work. In case of an old BIOS, users must force PXE to use Int18h, Int19h, or BBS as the boot method. To specify the boot method, press the following keys wh en you see the startup banner. Broadcom Gigabit Ethernet Boot Agent 2.2.6 Copyright (c) 2000 Broadcom Corporation All rights reserved ⢠For Int18h: Press the Ctrl F8 keys ⢠For Int19h: Press the Ctrl F9 keys ⢠For BBS: Press the Ctrl F10 keys Setup BIOS To boot from the network with PXE, make PXE the first bootable device under BIOS. This procedure depends on server BIOS implementation. Refer to the user manual for the serv er for additional information. Server setup Refer to the following instr uctions for your operating system. Windows® 2000
Adapter configurations 69 The current version of Windows® 2000 does not include a network driver for th e NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server Adapters. To perform remote installation with PXE, include a network driver for these adapters as part of the client's installation image on the server. Refer to Microsoft® Article ID Q246184, "How to Add Third-Party OEM Network Adap ters to RIS Installations." DOS UNDI/APITEST To boot in DOS and connect to a network for the DO S environment, download the Intel® PXE PDK from the Intel® web site. This PXE PDK comes with a TFTP/ProxyDHCP/Boot server. The PXE PDK can be downloaded from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). Red Hat Linux Red Hat Linux 7.2 (or higher) distribution has PXE serv er support. It allows users to perform a complete Linux installation over the network. Distribution al so comes with boot images âboot kernel (vmlin ux) and initial ram disk (initrd). These two images can be found in the HP SoftPaq file: ⢠/misc/src/trees/boot/vmlinux ⢠/misc/src/trees/initrd-netw ork.img Refer to Red Hat documentation for in structio ns on installing PXE ser ver on Linux. Initrd.img, distributed with Red Ha t 7.2, does not have a Linux netw ork driver for the NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC32 6m, NC10 20, NC67xx, and NC77xx Gigabit Server Adapters. ⢠/images/pxeboot/vmlinux ⢠/images/pxeboot/initrd.img Intel® has also created patches to the Linux PXE serv er code to allow clients to perform a remote boot. You can download these patches from the Intel website ( http://downloadfinder.intel.com/scripts-df- external/Support_Intel.aspx ). A remote boot does not require a standard Linux network driver for t he HP NC320T, NC320i, NC324i, NC325i, NC325m, NC326i, NC326m, NC102 0, NC67 xx, and NC77xx Gigabit Server Adapters. After the PXE client downloads the Linux kernel and in itial ram disk, the Linux universal driv er, that came with the Linux distribution, will bind with the UNDI code of the PXE to form a Linux network driver. HP NC-Series Intel adapters Boot Agent The Intel® Boot Agent is a software product that allo ws your networked client computer to boot using a program code image supplie d by a remote server. Implementation support for the Boot Agen t includes: ⢠Multi-protocol boot ROM code designed for re mote booting from Gigabit Ethernet Adapters ⢠Compliance with the Pre-boot eXec ution Env ironment (PXE) Version 2.1 Specification ⢠Remote Program Load (RPL) runtim e and loader software included wi th some versions of the Boot Agent Intel® Boot Agent offers:
Adapter configurations 70 ⢠Compliance with the Wired for Manage ment Baseline (WfM) 2.0 specification ⢠Compatibility with legacy boot agent environments that use BOOTP protocol ⢠Customization in pre-boot, Wi ndows®, and DOS environments Operating environment The Boot Agent operates in a client/server environment. Often, in this envir onment, one or more servers provide remote boot services to a large number of client computers thro ugh a common network. The computer system where the Intel® B oot Agent is loaded is considered to be a client with respect to the remote boot capability even if that system acts as a server after the syst em has finished booting. Configuration options The Intel® Boot Agent software provides configuration options that allow you to custom ize the behavior of the Boot Agent software. You can configure the Boot Agent in a pre-boot environment (before the operating system is loaded). Configuring the Boot Agent in a pre-boot PXE or RPL environment You can customize the behavior of the B oot Agent software through a pre-boot (operating system independent) configuration setup program contained within the adapter's flash ROM. A single user interface allows you to configure PXE and RPL protocols on HP Gigabit adapters. You can access this pre- boot configuration setup program each time the client computer cycles throug h the boot process. The boot process is triggered whenever any of the following boot events occur: ⢠Power on ⢠Hard reset ( Reset bu tton on system, if available) ⢠Soft reset ( Ctrl Alt Del ) ⢠Operating system or application-initiated system restart When the boot process begins, the screen clears and the computer begins it s Power On Self-Test (POST) sequence. Shortly after completion of the POST, the Boot Agent soft ware stored in flash ROM execut es. The Boot Agent then displays an initialization message, similar to the one below , indicating that it is active: Initializing Intel(R) Boot Agent Version X.X.XX PXE 2.0 Build 083 (WfM 2.0) Press Ctrl S to enter the Setup Menu. This display may be hidden by the manufacturer's splash screen. Refer to your manufactur er's documentation for details. To customize the behavior of the Boot A gent software in a pre-boot PXE or RPL environment complete the following steps: 1. Press the Ctrl S keys immediately after the initializati on message appears. A configuration setup menu appears allowing you to set configuration values for the Boot Agent. The configuration set up menu is the same for bo th kinds of adapters. If you do not press the Ctrl S keys, the Boot Agent software wil l proceed with the boot process eventually bringing up the operating syst em. If you miss your opportunity to p ress the Ctrl S keys within the allowed number of seconds, reboot the co mputer to try again. If you select a Setup Menu Wait Time setting of zero or a Show Setup Prompt setting of Disa bled, you will not be prompted to press the Ctrl S keys even though yo u can still enter the configuration setup menu using this key combination.
Adapter configurations 71 The configuration setup menu sho ws a list of configuration setting s on the left and their corresponding values on the right. Key descriptio ns near the bottom of the menu indicate how to change values for the configuration setti ngs. For each selected setting, a brief "mini-Help" description of its function appears just above the key descriptions. 2. Select the setting you need to change by using the arrow keys. 3. After you have accessed the settin g yo u want to change, press the Spacebar until the desired value appears. 4. If you want to change additional settings, repeat steps 2 and 3. 5. After you have completed your changes, press the F4 key to update the adapter with the new values. Any changed configuration values are appl ied as the boot process resumes. Boot Agent configuration settings Configuration settings Possible values Description Network Boot Protocol ⢠PXE (Preboot eXecution Environment) ⢠RPL (Remote Program Load) Controls whether the RPL or PXE boot protocol will be used. Select PXE for use with WfM-compatible network management programs, such as Intel® LANDesk® Manage ment Suite, Windows 2000 RIS, and Linux®. Select RPL for legacy-style remote booting, as well as for Novell® Netware® remote boot solutions. Depending on the configuration of the Boot Agent, this parameter may not be changeable. Boot Order ⢠Use BIOS Setup Boot Order ⢠Try network first, then local drives ⢠Try local drives first, then network ⢠Try network only ⢠Try local drives only Sets the boot order in which devic es are selected during boot up if the computer does not have its own control method. If your client computer's BIOS supports the BIOS Boot Specification (BBS), or allows PnP-compliant selection of the boot order in the BIOS setup program, then this setting will always be Us e BIOS Set up Boot Order and cannot be changed. In this ca se, refer to the BIOS setup manual specific to your client computer to set up boot options. If your client computer does not have a BBS- or PnP-compliant BIOS, you can select any one of the othe r possible values listed for this setting except for Use BIOS Setup Boot Order. Show Setup Prompt ⢠Enabled ⢠Disabled Controls whether or not the Boot Agent displays the Ctrl S prompt after POST. If you select Enabled, the Ctrl S prompt appears after POST so that you can press Ctrl S within the number of seconds allowed by the Setup Menu Wait Time setting to display the configuration setup menu. If you select Disabled and/or the Setup Menu Wait Time setting is zero, the Ctrl S setup prompt will not appear during the boot process. However, you can display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedly pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again.
Adapter configurations 72 Configuration settings Possible values Description Setup Menu Wait Time ⢠0 seconds ⢠2 seconds ⢠3 seconds ⢠5 seconds Controls the number of seconds the Boot Agent waits for you to press the Ctrl S keys, so as to suspend the boot process and then configure the way the Boot Agent operates. If you select a Setup Menu Wait Time setting of zero and/or a Show Setup Prompt setting of Disabled, the Ctrl S setup prompt will not appear during the boot process. Ho wever, you can still display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedl y pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again. If during PXE or RPL boot more than one adapter is in stalled in a computer and you want to boot from the boot ROM located on a specific adapter, you can do so by removing the adapter from the BIOS Boot Order or disabling the flash by running IBAUTIL -FlashDisable on the desired adapter. To display diagnostics information: Anytime the configuration setup men u is displayed, you may press the D key to display diagnostics information in the lower half of the screen . This in formation can be helpful during interaction with HP Customer Support personn el or your IT team members. After you press the D key, the informat ion displayed remains until you leav e the configuration setup screen (either by pressing the F4 key, pressing the Esc key, or by rebooting the computer). For more information about how to interpret the in formation displayed, refer to the "Diagnostics infor mation for pre-boot PXE or RPL environments" section. Auxiliary DOS utilities The IBAUtil allows you to install and configure the Bo ot Agent using the DOS environment. IBAU til is a utility program that provides an alternative means for modifying the adapter co nfiguration settings. Use IBAUtil to: ⢠Change the default settings of your HP NC 310F, NC3123, NC617 0, or NC7170 adapter. ⢠Enable/disable the Wake-on-LAN (WOL) an d Intel® Boot Agent capabilitie s. ⢠Allow in-the-field upgrades to the image stor ed in the flash component of the adapter. Wake-On-LAN and Intel® Boot Agent in a DOS environment Use IBAUtil to enable or disable WOL or Boot Agent fe atures. To obtain this utility: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. The IBAUtil file is located in the APPS\BOOTAGNT\N 100X directory. ⢠Wake-On-LAN. When enabled, the adapter can reac t to special "wake up" packets and power up the computer without end user intervention. However, there i s a higher power draw when the system
Adapter configurations 73 is in the suspended state when this i s enabled. It is strongly recommended that no more than two adapters should have this feature enabled per computer. ⢠Intel Boot Agent. When enabled, the computer can initiate PXE/RPL boot if a valid flash image is present on the adapter. Running IBAUtil in DOS IBAUtil (for DOS) must be run with the computer boot ed to DOS only. This utility cannot be run in a DOS window in any Microsoft® Windows® product or in DOS compatibility mode in IBM OS/2. The syntax for issuing IBAUtil command line parameters in DOS is: IBAUTIL [-option]... DOS command line parameters IBAUtil accepts up to 16 simultaneous parameters and does not perfor m both operations if orthogonal parameters (like the commands -FLASHENABLE -FLASHDI SABLE used together) are supplied. If opposing parameters are specified, IBAU til exits with an error. The parameters are grouped with their orthogonal pairs, where applicable. -H ELP, -EXITCODES and - IMAGEVERSION have special overridi ng characteristics. Unless noted, all other parameters may be used in conjunction with all other parameters. If you enter IBAUTIL without any command line options, the utility displays a listing of all of the Intel® network adapters found in your system. Valid parameters are listed below. All other parameters generate an error. Parameter Description -HELP or -? Displays command line help and exits. When -HELP is set, all other parameters are ignored. -EXITCODES Displays exit code help. When -EXITCODES are set, all other parameters except -HELP are ignored. -IMAGEVERSION or -IV Displays the PXE vers ions stored within the IBAUtil. When - IMAGEVERSION is set, all other parameters except -EXITCODES and -HELP are ignored. Adapter Selection Options -ALL Selects all adapters found in system, works for both EEPROM and Flash activity. -NIC=XX Selects a specific adapter (1â16). -BLINK Blinks the LED on the selected adapter for 10 sec onds to provide a method for identifying an adapter. Flash Programming Options -AUTO or -QUIET Runs IBAUtil without asking for user intervention. It does NOT create a backup image. This paramete r must be used with -NIC or - ALL if there is more than one adapter in the system. It runs without - ALL or -NIC=xx if there is only one adapter in the system. -UPGRADE or -UP Downloads the Boot Agent stored within IBAUtil to the adapters in the system. -FLASHENABLE or -FE Enables Boot ROM. Saves -FLASHDISABLE or -FD Disables Boot ROM.
Adapter configurations 74 Parameter Description -SAVE Saves existing Intel Boot Agen t and EEPROM settings to a file. The file is named based on the PCI Vendor and Device ID of the adapter. -RESTORE Restores previously saved Inte l Boot Agent image. IBAUtil looks for a file name based on the PCI Vendor and Device ID of the adapter. If -RESTORE is used with -UPGRADE, an error is generated. Power Management Options -WOLDISABLE or -WOLD Disables WOL bit on Fast Ethernet and disables APMPME and APME EEPROM bits on gigabit. -WOLENABLE or -WOLE Enables WOL bit on Fast Ethernet and enables APMPME and APME EEPROM bits on gigabit. -LWSENABLE OR -LWSE Enables changing the legacy OS Wakeup Support Option on 10/100 adapters. -LWSDISABLE OR -LWSD Disables changing the legacy OS Wakeup Support Option on 10/100 adapters. PXE/RPL Configuration Options -SETUPENABLE or -STE Enables Setup Menu. -SETUPDISABLE OR -STD Disables Setup Menu. -DEFAULTCONFIG or -DEFCFG Resets configurat ion and control words in the EEPROM back to default settings. -TITLEENABLE OR -TLE Enables initial title message. -TITLEDISABLE OR -TLD Disables initial title message. -PROTOCOLENABLE OR -PRO E Enables changing the boot protocol option. -PROTOCOLDISABLE OR -PROD Disables changing the boot protocol option. -ORDERENABLE OR -ORDE Enables changing the boot order option. -ORDERDISABLE OR -ORDD Disables ch anging the boot order option. -SETWAITTIME=X or -SWT=X Sets the setup delay time for the ada pter to X seconds. Valid delay times are 0, 2, 3, and 5. Only the first instance of this parameter is valid. All other instances of the parameter generate an error. -RPL BOOT Sets the network boot protocol to RPL if an RPL image is present in the adapter's flash. -PXE BOOT Sets the network boot protocol to PXE if a PXE image is present in the adapter's flash. DOS error codes IBAUtil returns codes to the DOS co mmand line when an error occurs. ⢠0âSuccess ⢠1âOperator termination ⢠2âInvalid adapter ⢠3âBad command line parameter ⢠4âEEPROM checksum failure
Adapter configurations 75 ⢠5âEEPROM read failure ⢠6âEEPROM write fail ure ⢠7âEEPROM dump failure ⢠8â(Not used) ⢠9âNo memory ⢠10âNo adapters found Examples of IBAUtil command lines The following are examples of some typical IBAUtil command lines. ⢠Disables WOL on all HP network adapters: IBAUTIL -ALL -WOLDISABLE ⢠Enables WOL on the second HP netw ork adapter found in your system: IBAUTIL -NIC=2 -WOLENABLE ⢠Updates the Boot Agent image on all HP network adapters: IBAUTIL -ALL -UPGRADE ⢠Enables the Boot ROM on the fir st HP Network Adapter found in your system: IBAUTIL -NIC=1 -FE Boot Agent messages The following error and information messages may be di splayed during initializati on. Possible causes and how to avoid an error condition are provided where applicable. ⢠Flash device wrong size This error message can occur when using IBAUtil. The full version of Boot Agent so ftware cannot be loaded into a flash ROM device smaller than 64 KB. If a 32 KB (or smaller) flash ROM device is detected in the adapter, the flash update software returns this message. Assuming the device is socketed, remove the "too small" device and install a blank 64 KB (or larger) flash ROM device. ⢠Invalid PMM function number PMM is not installed or is not work ing correctly. Try updating the BIOS. ⢠PMM allocation error PMM could not or did not allocate the reques ted amount of memory for driver usage. ⢠Press Ctrl S to enter the Setup Menu. Instructs you how to enter the configuration setup menu. ⢠PXE-E00: This system does not have enough free conventional memory. The Boot Agent cannot continue. System does not have enough free memory to run PXE image. The Boot Ag ent was unable to find enough free base memory (below 6 40K) to install the PXE client software. The system cannot boot via PXE in its current configuration. The error retu rns control to the BIOS and the system does not attempt to remote boot. If this error persists, try updating your system's BIOS to the most-recent version. Contact your system adm inistrator or HP customer support to resolve the problem. ⢠PXE-E01: PCI Vendor and Device IDs do not match! Image vendor and device ID do not match those loca ted on the card. Be sure the correct flash image is installed on the adapter. ⢠PXE-E04: Error reading PCI configuration space. The Boot Agent cannot continue.
Adapter configurations 76 PCI configuration space could not be read. Machin e is probably not PCI compliant. The Boot Agent was unable to read one or more of the adapter's PCI configuration registers. The adapter may be misconfigured, or the wrong Boot A gent image may be installed on the adapter. The Boot Agent returns control to the BIOS and not attempt to remote boot. Try to update the flash image. If this does not solve the problem, contact your system administrat or or HP Customer Support. ⢠PXE-E05: The LAN adapter's configuration is corrupted or has not been initialized. The Boot Agent cannot continue. The adapter's EEPROM is corrupted. The Boot Agent determ ined that the adapter EEPROM checksum is incorrect. The agent will return control to the BIOS and not attempt to remote boot. Try to update the flash image. If this does not solve the problem, contact your system administrat or or HP Customer Support. ⢠PXE-E06: Option ROM requires DDIM support. The system BIOS does not support DDIM. The BIOS does not support the mapping of the PCI expansion ROMs into upper memory as required by the PCI specification. The Boot Ag ent cannot function in this system. The Boot Agent returns cont rol to the BIOS and does not attempt to remote boot. You may be able to resolve the problem by updating the BIOS on your system. If updating your system's BIOS does not solve the pro blem, contact your system administrator o r HP customer support to resolve the problem. ⢠PXE-E07: PCI BIOS calls not supported. BIOS-level PCI services not available. Machine is prob ably not PCI compliant. ⢠PXE-E09: Unexpected UNDI loader error. Status == xx The UNDI loader returned an unknown e rror status. xx is the status returned. ⢠PXE-E20: BIOS extended memory copy error. BIOS could not move the image into extended memory. ⢠PXE-E20: BIOS extended memo ry copy error. AH == xx Error occurred while trying to copy the image in to extended memory. xx is the BIOS failur e code. ⢠PXE-E51: No DHCP or BOOTP offers received. The Boot Agent did not receive any DHCP or BOOTP re sponses to it s initial request. Be sure that your DHCP server (and/or proxyDHCP server, if one is in use) i s properly configured and has sufficient IP addresses available for lease. If you are using BOOTP, be sure that the BOOTP service is running and is properly configured. ⢠PXE-E53: No boot filename received. The Boot Agent received a DHCP or BOOTP offer, but has not received a valid filename to download. If you are using PXE, please check your PXE and BINL configu rat ion. If you are using BOOTP, be sure that the TFTP service is running an d that the specific path and filename are correct. ⢠PXE-E61: Media test failure. The adapter does not detect link. Be sure that the cable is good and is attached to a working hub or switch. The link light visible from th e back of the adapter should be lit. ⢠PXE-EC1: Base-code ROM ID structure was not found. No base code can be located. An incorrect fl ash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC3: BC ROM ID structure is invalid.
Adapter configurations 77 Base code cannot be installed. An incorrect fl ash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC4: UNDI ROM ID structure was not found. UNDI ROM ID structure signature is incorrect. An incorrect flash ima ge is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC5: UNDI ROM ID structure is invalid. The structure length is incorrect. An incorrect flash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC6: UNDI driver image is invalid. The UNDI driver image signature is invalid. An incorrect flash image is install ed or the image has become corrupted. Try to update the flash image. ⢠PXE-EC8: !PXE structure was not found in UNDI driver code segment. The Boot Agent cannot locate the needed !PXE stru cture reso urce. An incorrect flash image is installed or the image has become corr upted. Try to update the flash image. ⢠PXE-EC9: PXENV structure was not found in UNDI driver code segment. The Boot Agent cannot locate the needed PXENV structure. An inco rrect flash image is installed or the image has become corrupted. Try to update the flash image. ⢠PXE-M0F: Exiting Intel Boot Agent. Ending execution of the RO M image. ⢠This option has been locked and cannot be changed. You have attempted to change a configuration setting that has been locked by your system administrator with IBAUtil. This message can ap pear from the configuration setup menu when operating in a stand-alone environment. If you thin k you should be able to change the configuration setting, consult your system administrator. ⢠PXE-M0E: Retrying network boot; press ESC to cancel. The Boot Agent did not successfully complete a netw ork boot due to a network error (such as not receiving a DHCP offer). The Boot Agent continues to a ttempt to boot from the network until successful or until canceled by the user. This feat ure is disabled by default. For information on how to enable this feature, contact HP Customer Support. Troubleshooting procedures The following list of problems and associated sol ution s covers a representative set of problems that you might encounter while using the Boot Ag ent. If you are experiencing a problem that is not listed in this section, contact HP Customer Support. After booting, my computer experiences problems After the Boot Agent product has finished its sole task (remote booti ng), it no longer has any effect on the client computer operation. Thus, any issues that arise after the boot process is complete are most likely not related to the Boot Agent product. If you are having problems with the local (client) or network operating syst em, contact the operating system manufacturer for as sistance. If you are having problems with some application program, contact the application manufacturer for assistance. If you are having problems with any of your computer's hardware or with the BIOS, contact HP Customer Support for assist ance. Cannot change boot order
Adapter configurations 78 If you are accustomed to redefini ng your computer's boot order using the motherboard BIOS setup program, the default settings of the Boot Agent set u p program can override that setup. To change the boot sequence, you must first override the Boot Ag ent setup program defaults. To start the Boot Agent configuration setup progra m, press the Ctrl S keys during the time the initialization message appears. A configuration setup menu appears allow ing you to set configuration values for the Boot Agent. If you do not press the Ctrl S keys, the Boot Agent softwar e proceeds with th e boot process, eventually bringing up the operating system. To change your computer's boot order sett ing, refer to the "Configuring t he Boot Agent in a pre-boot PXE or RPL environm ent" section. My computer does not complete POST If your computer fails to boot with an adapter instal led, but does boot when you remove the adapter, try moving the adapter to another computer an d using IBAUtil to disable the flash chip. If this does not work, the proble m may be occurring before the Boot Agent softwa re even begins operating. In this case, there may be a BIOS proble m with your computer. Contact HP Customer Support for help in correcting your problem. Not getting a prompt to go to Boot Agent setup program This is actually a feature of the Boot Agent product. As part of the Boot Agent's Configurati on Setup Program, two of the options, Show Setup Prompt and Setup Wait Time, allow the boot process to proceed after POST without interruption by the Ctrl S prompt (Press the Ctrl S keys to enter the Setup Menu) If the Show Setup Prompt has previously been set to Disabled and/or if the Setup Menu Wait Time option is set to zero seconds, the prompt does not disp lay after POST. However, even though you are not prompted to press the Ctrl S keys, you can still enter the config uration setup menu to customiz e configuration settings as f ollows: Repeatedly press the Ctrl S keys immediately after POS T until th e Configuration Setup Menu appears. If the Configuration Setup Menu does not appear after repeatedly pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again. While in the Configuration Setup Men u, you can restore the prompt message as follows: ⢠If Show Setup Prompt has been set to Disabled, change it to Enabled. ⢠If Setup Menu Wait Time is set to zero seconds, change it to an y of the wait time values that exceed zero seconds. Refer to the list of possibl e Set Menu Wait Time values shown in the "Boot Agent configuration settings" section. There are configuration/operation problems with the boot process If your PXE client receives a DHCP address, but then fails to boot, you know the PXE client is working correctly. Check your network or PXE server config uration to trou bleshoot the problem. Contact HP Customer Support if you need further assistance. Diagnostics information for pre-boot PXE or RPL environments Anytime the Configuration Setup Men u is displayed (refer to the "Config uring the Boot Agent in a pre- boot PXE or RPL environment" section), you may press the D key to display diagnostics information on the screen. This information can be helpful during interact ion with HP Customer Support personnel or your IT team members. After you press the D key , the information remains displayed until you reboot your computer. Actual diagnostics information may vary, depending upon the adapter(s) installed in your computer. Diagnostics information may include the f ollowing items: ⢠PWA NumberâThe Printed Wire Assembly number of the device as stored in the EEPROM
Adapter configurations 79 ⢠MAC AddressâThe individual address of the device as stored in the EEPROM ⢠I/OâThe I/O address for PCI access determined by the software. On cards without an I/O address, all zeros are displayed ⢠MemoryâThe memory map PCI access address det ermined by the software ⢠SlotâThe slot number reported by the BIOS. The number displayed is the BIOS v ersion of the PCI slot number. Therefore, actual positions of adapte rs within slot s may not be displayed as expected. Slots are not always enumerated in an obvious manne r, and the diagnostic s will only report what is indicated by the BIOS. Boot agent glossary This glossary defines terms, abbrevia tions, and acronyms that apply directly to the Intel® Boot Agent. Term / Abbreviation / Acronym Definition ACPI Advanced Configuration and Power Interface AOL Alert on LAN API Application Program Interface. The spec ific method presc ribed by an OS or an application program by which a programmer writing an application program can make requests of the operating system or the application. BAID BIOS Aware IPL Device. The BIOS contains all code required to IPL from the device. BAR Base Address Register BBS BIOS Boot Specification BC BaseCode. The PXE BaseCode, included as a core prot ocol in EFI, is comprised of a simple network stack (UDP/IP) and a few common network protocols (DHCP, Bootserver Discovery, TFTP) that are useful for remote bootin g machines. BCV Boot Connectio n Vector. A field in the Plug and Play (PnP) header for a device with an associated option ROM. BEV Boot Entry Vector. A field in the Plug and Play (PnP) Header of a device with an associated option ROM. PXE is implemented as a BEV option ROM. BINL Binary Image Negotiation Layer BIOS Basic Input/Output System. The program a personal computer's microprocessor uses to get the computer system started after yo u turn it on. It also manages data flow between the computer's OS and attached devices. BIS Boot Integrity S ervices BOOTP Bootstrap Protocol. A legacy remote b ooting protocol developed originally for use with UNIX. Used as Linux's server side PXE host software. Runs a Daemon once installed. DDIM Device Driver Initialization Model DHCP Dynamic Host Configuration Protocol. An industry standard internet protocol defined by the IETF. DHCP was defined to dyna mically provide communications-related configuration values such as network address es to network client computers at boot time. DHCP is specified by IETF RFC s 1534, 2131, and 21 32. EEPROM Electrically Erasable Pr ogrammable Read-Only Memory ESD Electro-Static Discharge Flash A high-density, truly non-volatile, hig h-performance, read-write memory solution, also characterized by low power consumption, extreme ruggedness, and high reliability.
Adapter configurations 80 Term / Abbreviation / Acronym Definition Flb Flash Library IA32 Shorthand for computer platforms ba sed on Intel 32-bit architecture CPUs. IA64 Shorthand for computer platforms ba sed on Intel 64-bit architecture CPUs. IAL Intel Architecture Lab IBA Intel Boot Agent IETF Internet Engineering Task Force. The IETF is a large open int ernational community of network designers, operators, vendors, and researchers concerned with the evolution of the Internet architecture and the smooth operation of t he internet. It is open to any interested individual. For more details, refe r to the Internet Engineering Task Force website. IPL Initial Program Load. Also known as the bootstrap or boot process. LCM LANDesk Configuration Manager LOM LAN On Motherboard. This is a network de vice that is built onto the motherboard (or baseboard) of the machine. MTFTP Adds multicast support to a TFTP server. NBP Network Bootst rap Program. The first pr ogram that is downloaded into a machine that has selected a PXE capable devi ce fo r remote boot services. A typical NBP examines the machine it is running on to try to determine if t he machine is capable of running the next layer (OS or application). If the machine is not capable of running the next layer, control is returned to th e EFI boot manager and the next boot device is selected. If the machine is capable, t he next layer is downloaded and control can then be passed to the downlo aded program. Though most NBP s are OS loaders, NBPs can be written to be stan d-alone applications such as diagnostics, backup/restore, remote management agents, browsers, and so on. NIC Network Interface Controller. Also referred to as adapter. Technically, a network device that is inserted into a bus on the motherboard or into an expansion board. For the purposes of this document, the term adapter is used in a generic sense, meaning any device that enables a netwo rk connection (including LOMs and network devices on external buses, such as USB 1394, and so on). OS Operating System PMM POST Memory Mana ger. A mechanism used by option ROMs to allocate RAM memory for use during system startup. PnP Plug and Play POST Power On Self-Test proxyDHCP proxyDHCP is used to ease the transition of PXE clients and servers into an existing network infrastructure. proxyDHCP provid es additional DHCP information that is needed by PXE clients and boot servers wi thout making changes to existing DHCP servers. PXE Preboot Executi on Environment. PXE pr ovides a way for a system to initiate a network connection to various servers prior to loading an OS. This netw ork connection supports a number of standard IP protocols such as DHCP and TFTP, and can be used for purposes such as softwa re installation and system inventory maintenance. PXE 16 PXE Specification v2.1 or ea rlier. Also known as 16-bit PXE. RFC Request For Comments. Used to denote that a new standard or modification to standard is be introduced by IE TF or other standards bodies.
Adapter configurations 81 Term / Abbreviation / Acronym Definition RIS Remote Installation Services. A Micr osoft® service that uses PXE to deliver the Network Bootstrap Program (NBP) to start the boot process. Sh ips with Windows® 2000 Advanced Server. ROM Read-Only Memory. When used in this guide, ROM refers to a non-volatile memory storage device on an adapter RPL Remote Program Load. A bootable network device protocol, developed and used mostly by IBM. TCO Total Cost of Ownership. Business models often calculate the approximate total cost of ownership for technology initiative \s through formulas that incorporate both expenditure on hardware and software an d maintenance, replacement, training, and all aspects of business that are impacted by the core system. Remote management tools like PXE help lower the overall cost of t he ownership. TFTP Trivial File Tran sfer Protocol. An indust ry standard internet protocol defined by the IETF to enable the transmission of files across the internet. Trivial File Transfer Protocol (TFTP, Revision 2) t o support NBP downloa d is specified by IETF RFC 1350 . UNDI Universal Network Driver Interface. UNDI is an archit ectural interface to adapters. Traditionally adapters have had custom inte rfaces and custom drivers (each ada pter had a driver for each OS on each pl atform architecture). UNDI provides a hardware-dependent me chanism for the PXE base code to use a adapter for network access without controlling the adapter hardware directly. H/W UNDI is an architectural hardware interface to an adapter; S/W UNDI is a software implementation of the H/W UNDI. URL Uniform Resource Locator. The address of a file (resource) accessible on the internet. USB Universal Serial Bus. A P lug and Play (PnP) interface between a c omputer and add- on devices. WfM Wired for Management WOL Wake on LAN Keywords for N100 NDIS2 driver The following how to configure the parameter s for the NDIS 2 driver version 4.56 (N100 DOS) . General configuration parameters The following configuration parame ters are supported through the PR OTOCOL.INI file. The "Possible Errors" section describes the driver's behavior when erro r conditions are detected. If an error is fatal, the driver will not load. In all other cases the driver will load and use the default value. ⢠DRIVERNAME o Syntax: DRIVERNAME=[N100$ | N1002$ | etc] o Example: DRIVERNAME=N100$ o Default: None. This is a required parameter. o Normal Behavior: The driver finds its section in PR OTOCOL.INI by matching its instance ID to the value for this parameter. o Possible Errors: The device driv er uses a DOS and OS/2 function to display the name of the driver it is expecting. This function cannot disp lay a "$" character. For this reason, the user may
Adapter configurations 82 see a message referring to this value with the "$." The user must rem ember to enter the "$" character as part of the parameter's value. ⢠SPEED o Syntax: SPEED=[Auto | 10 | 100] o Example: SPEED=10 o Default: Auto-Speed-Detect as a function of PHY detection o Normal Behavior: If either 10 or 100 is specified, a message indicates that Auto-Speed-Detect has been disabled and the driver sets the speed us ing the value specified. If the parameter i s not provided at all, or if Auto is prov ided, Auto-Speed-Sensing is enabled. o Possibl e Errors: If the SPEED parameter is set to an invalid valu e, the parameter is ignored and the default value is used, and a message indicates a "Par ameter value out of range" error. If the FORCEDUPLEX parameter is assigned a valid non-zero value, this parameter is required. If the duplex is forced, Auto-Speed-Sensing is disabled and the speed and full-duplex modes are set as specified by the user. An N-Way PHY will not be able to detect speed changes in this configuration until the driver is reloaded. ⢠FORCEDUPLEX o Syntax: FORCEDUPLEX=[Auto | 1 | 2] o Example: FORCEDUPLEX=1 o Default: Duplex mode configuration detected during PHY initialization. o Normal Behavior: The parameter value of 1 wi ll force the duplex mode to half-dup lex. A parameter value of 2 will force th e duplex mode to full-duplex. The SPEED parameter must be specified and must be valid if either half or full-duplex is forced. Auto-detection by the PHY is disabled and the sp eed and full-duplex modes are set as specified. The N-Way PHY will not be able to auto-detec t the mode until the driver is reloaded. o Possibl e Errors: If the FORCEDUPLEX parameter is set to an invalid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. If the FORCEDUPLEX mode does not match the actu al mode of the port in use, the driver will function for certain levels of traffic and then begin to lose frames. If the port in use is FULL duplex but it is not capable of performing an N-way auto-negotiation, the user MUST use FOR CEDUPLEX set to FULL. ⢠PHYADDRESS o Syntax: PHYADDRESS=[0..32] o Example: PHYADDRESS=1 o Default: 1, implying logical PHY 1 at address 1. o An address of 0 implies logical PHY 0. PHY 0 can only be at address 0. PHY 1 can have addresses in the range 1 - 31. PHY address 32 can be used to indicate the presence of an 82503 interface.
Adapter configurations 83 o Normal Behavior: The value will be used by the PH Y detection rout ine to locate and initialize the PHY. If a PHY is not found at the address specifie d, the driver will load but not function properly, because PHY detection occurs afte r the initial load has occurred. o Possible Errors: If the PHYADDRESS parameter is set to an invalid value, the parameter is ignored and the default value is used, and a message indi cates a "Parameter value out of range" error. ⢠SLOT o Syntax: SLOT=[0x0..0x1FFF] Optional unless you wa nt to specify a particular a d apter or module that is not loaded by default. (By default, the driver loads on the first adapter it finds.) o Example: SLOT=0x1C o Default: The driver will Auto-Configur e o Normal Behavior: The driver uses the value of th e parameter to decide which adapter to control. o Possibl e Errors: If only one adapter is installed and the value does not correctly indicate the adapter slot, a message indicates that the value does not match th e actual configuration, and the driver find s the adapter and uses it. If more than one adapter is installed and the va lue does not indicate the adapter slot, a message displays indicating possible slots to use and the driver loads on the next available slot. ⢠NODE o Syntax: NODE="12 hexadecimal digits" The value must be exactly 12 hexadecima l digits, enclosed in double quot es The value cannot be all zeros. The value cannot have the Multicast bit set (LSB of 2nd digit = 1). o Example: NODE="00AA00123456" o Default: Value from EEPROM installed on adapter o Normal Behavior: The Current Station Address in th e NDIS MAC Service-Specific Characterist ics (MSSC) table is assigned the value of this parameter. The adapter hardware is programmed to receive frames with the destinat ion address equal to the Current Station Address in the MSSC table. The Permanent Station Address in the MSSC table will be set to reflect the node address read from the adapter's EEPROM. o Possible Errors: If any of the rules described above is violated , an error message indicates the rules for forming a proper address, and the driver treats this as a fatal error. ⢠CONGENB o Syntax: CONGENB=[0 | 1] o Example: CONGENB=1 o Default: Congestion Control is disabled. o Normal Behavior: The absence of th is parameter or the value 0 resu lts in no Congestion Control. The value of 1 enables the Congestion Contro l feature. o Possible Errors: If the CONGENB parameter is set to an invalid value, the parameter is ignor ed and the default value is used, and a message indi cates a "Parameter value out of range" error. ⢠RXFIFO o Syntax: RXFIFO=[0..15 | 0x0..0xF] o Example: RXFIFO=8
Adapter configurations 84 o Default: 8, corresponding to 32 byte s or 8 dwords in the receive FIFO. o Normal Behavior: No message wi ll be displayed. The driver will use the specified value to configure. o Possible Errors: If the RXFIFO pa rameter is set to an invalid valu e, the parameter is ignored and the default value is used, and a message indica tes a "Parameter value out of range" error. ⢠TXFIFO o Syntax: TXFIFO=[0..15 | 0x0..0xF] o Example: TXFIFO=8 o Default: 8, corresponding to 32 bytes or 8 dwords in the transmit FIFO. o Normal Behavior: No message wi ll be displayed. The driver will use the specified value to configure. o Possible Errors: If the TXFIFO pa rameter is set to an invalid valu e, the parameter is ignored and the default value is used, and a message indica tes a "Parameter value out of range" error. ⢠TXTHRESHOLD o Syntax: TXTHRESHOLD=[0..254 | 0x0..0xFE] o Example: TXTHRESHOLD=32 o Default: 32, corresponding to 32 * 8=256 bytes o Normal Behavior: The value will be set in every transmit command. The initial value can only be changed at load time, although the driver will increase the value if underruns are d etected. Setting a low value (less than 16) in high bus latency sy stems could cause excessive TX DMA underruns, reducing throughput until th e driv er reacts. Setting a value of more than 200 completely eliminates the possibility o f TX DMA underruns. o Possibl e Errors: If the TXTHRESHOLD parameter is set to an in valid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. ⢠IFS o Syntax: IFS=[0..255 | 0x0..0xFF] o Example: IFS=16 o Default: 0 (minimum spacing between frames) o Normal Behavior: The value will be used to config ure the 82557 at initiali zation time. A value of 0 implies minimum IFS between 2 transmit frames. Increasing the value reduces the probability of collisions under heavy load due to multiple client s, but will have a negative impact on transmit throughput in a point-to-point single-client packet burst. o Possible Errors: If the IFS paramet er is set to an invalid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. ⢠RXDMACOUNT o Syntax: RXDMACOUNT=[0..127 | 0x0..0x7F] o Example: RXDMACOUNT=16 o Default: 0 implying that the Rx DMA will contin ue until it is done. No other DMA request can preempt it.
Adapter configurations 85 o Normal Behavior: The value will be used to config ure the 82557 at initialization ti me. Setting it to a low value will force frequent internal arbitr ation, which will affect overall performance. The value can be tuned to obtain optimum balance between the TX and Rx DMA machines. o Possibl e Errors: If the RXDMACOUNT parameter is set to an in valid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. ⢠TXDMACOUNT o Syntax: TXDMACOUNT=[0..127 | 0x0..0x7F] o Example: TXDMACOUNT=16 o Default: 0 implying that the Tx DMA will contin ue until it is done. No other DMA request can preempt it. o Normal Behavior: The value will be used to config ure the 82557 at initialization ti me. Setting it to a low value will force frequent internal arbitr ation, which will affect overall performance. The value can be tuned to obtain optimum balance between the Tx and Rx DMA machines. o Possibl e Errors: If the TXDMACOUNT parameter is set to an in valid value, the parameter is ignored and the default value is used, and a message is indica tes a "Parameter value out of range" error. ⢠TRANSMITBUFFERS o Syntax: TRANSMITBUFFERS=[4 .. 30] o Example: TRANSMITBUFFERS=14 o Default: 4 for DOS driver, 10 for OS/2 driver o Normal Behavior: The driver allocates the number of buffer s indicated by the parameter value. o Possibl e Errors: If the value provided is out of range, the paramet er is ignored and the default value is used, and a message indicates a "Parameter value out of range" error. CAUTION: If the number of buffers allocated caus es the driverâs size to exceed MAXIMUM, the driver adjusts the number of buffers d own ward, by reducing each of the counts of transmit/receive buffers by one unt il the total allocation will fit in the available memory. A message indicates the actual number of buff ers allocated. Under DOS without Window for Workgroups (WFW), MAXI MUM is define d as 64Kb for all code and data. Under DOS and WFW, the driver must not claim more space than its load size, but since the driver cannot detect that it is running under WFW, a message warns the user when the size allocated exceeds the size of the driver when loaded. Under OS/2, MAXIMUM is defin ed as 64Kb for all data. ⢠RECEIVEBUFFERS o Syntax: RECEIVEBUFFERS=[4 .. 30] o Example: RECEIVEBUFFERS=14 o Default: 8 for DOS driver, 10 for OS/2 driver o Normal Behavior: The driver allocates the number of buffers indicated by the parameter value. o Possibl e Errors: If the value provided is out of range, the paramet er is ignored, the d efault value is used, and a message indicates a "Parameter value out of range" error.
Adapter configurations 86 CAUTION: If the number of buffers allocated caus es the driverâs size to exceed MAXIMUM, the driver adjusts the number of buffers d own ward, by reducing each of the counts of transmit/receive buffers by one unt il the total allocation will fit in the available memory. A message indicates the actual number of buff ers allocated. Under DOS without Window for Workgroups (WFW), MAXI MUM is define d as 64Kb for all code and data. Under DOS and WFW, the driver must not claim more space than its load size, but since the driver cannot detect that it is running under WFW, a message warns the user when the size allocated exceeds the size of the driver when loaded. Under OS/2, MAXIMUM is defined as 64Kb for all data. ⢠MCWA o Syntax: MCWA=[0..255 | 0x0..0xFF] o Example: MCWA=8 o Default: 6 o Normal Behavior: This function is based on the system timer. When the counter reaches 2 seconds (in both OS/2 and DOS), the driver checks to see if the driver has received no frames since the last time the counter ran, and the Re ceive Ha ng workaround will be executed. When the workaround condition exists, the serv ice routin e that runs the workaround will attempt to run the workaround every timer tick (18.3/sec in DOS, ~32/sec in OS/2). The EEPROM contains information that will control the running of the workaround. If in the event that the EEPROM is configured so that the workaround is not need ed, the counter will not be added to the timer chain, saving system resourc es. o Possible Errors: The value is not checked in any wa y, only 8 bits of value are used. The counter's effectiveness could be compromised if anot her execut able disrupts the timer chain. ⢠FRAMEGAP o Syntax: FRAMEGAP=[6..64 | 0x06..0x40] o Example: FRAMEGAP=12 o Default: 6 o Normal Behavior: In normal circumstances, this will not be needed. If packets are being dropped while running a T4 card, the FRAMEGAP must be increased. A value of 12 or 15 will fix the problem in most cases. Do not use this p arame ter unless you are exp eri encing dropped packets. o Possible Errors: If an invalid ra nge or syntax for this parameter occurs, the default value is used. The driver does not give any outward indica tion as to the state of this parameter. ⢠CACHEFLUSH o Syntax: CACHEFLUSH=[0|1] o Example: CACHEFLUSH=1 o Default: 0 o Normal Behavior: Use this parame ter during a remote installation or "push-install" of Windows NT®. o Possible Errors: Any non-zero valu e sets this parameter to 1. The driver does not give any outward indication as to th e state of this parameter.
Adapter configurations 87 Keywords for N1000 NDIS2 driver The following describes how to config ure the parameters for the NDIS 2 driver vers ion 3.62 (N1000.DOS). Configuration The following configuration parame ters are supported through the PR OTOCOL.INI file. The "Possible Errors" section describes the driver's behavior when erro r conditions are detected. If an error is fatal, the driver will not load. In all other cases the driver will load and use the default value. ⢠DRIVERNAME o Syntax: DRIVERNAME=[N1000$ | N10002$ | etc] o Example: DRIVERNAME=N1000$ o Default: None. This is a required parameter. o Normal Behavior: The driver finds its section in PR OTOCOL.INI by matching its instance ID to the value for this parameter. o Possible Errors: The device driver uses a DOS fu nction to display the name of the driver it is expecting. This function cannot display a "$" ch aracter. For this reason, the user may see a message re ferring to this valu e without the "$." The user must remember to enter the "$" character as part of the parameter's value. ⢠SPEEDDUPLEX o Syntax: SPEEDDUPLEX = [0 | 1 | 2 | 3] o Example: SPEEDDUPLEX = 1 o Default: Speed and Duplex mo de configuration detected during PHY initialization. o Normal Behavior: The parameter value of: 0 will force the speed to 10 Mbps and the duplex mode to half duplex. 1 will force the speed to 10 Mbps and the duplex mode to full duplex. 2 will force the speed to 100 Mbps an d the duplex mode to half duplex. 3 will force the speed to 100 Mbps an d the duplex mode to full duplex. o Auto-detection by the PHY is disabled and the sp eed and full duplex modes are set as specified. The PHY will not be able to auto-detect the mode until the driver is reloaded. o Possibl e Errors: If the SPEEDDUPLEX parameter is set to an inva lid value, the parameter is ignored and the default value is used, and a message indicate s a "Parameter value out of range" error. If the SPEEDDUPLEX mode does not match the actual mode of the port in use, the driver may or may not function for certain levels of traffic and then begin to lose frames. The SPEEDDUPLEX parameter is valid for HP NC71 xx Server Adapters or Modules only and is ignored if used with HP NC61 xx Server Adapters or Modules. The HP NC71xx Server Adapters and Modules cannot be forced to 1000 Mbps. ⢠SLOT o Syntax: SLOT=[0x0..0x1FFF] Optional unless you wa nt to specify a particular a d apter or module that is not loaded by default. o Example: SLOT=0x1C o Default: The driver will Auto-Configur e if possible.
Adapter configurations 88 o Normal Behavior: The driver uses the value of th e parameter to decide which adapter to control. o Possibl e Errors: If only one adapter is installed and the value do es not correctly indicate adapter slot, a message indicates that the value does not match the actual configuration, and the driver finds the adapter and uses it. If more than one adapter is installed and the va lue does not indicate the adapter slot, a message indicates that the specified slot is invalid for the driver instance. The parameter is ignored and the default value is used. ⢠NODE o Syntax: NODE = "12 hexadecimal digits" The value must be exactly 12 hexadecima l digits, enclosed in double quotes. The value cannot be all zeros. The value cannot have the Multicast bit set (LSB of 2nd digit == 1). o Example: NODE = "02AA00123456" o Default: Value from EEPROM installed on adapter o Normal Behavior: This value sets the adapter's Individual Address, overriding the value read from the adapter EEPROM. The address must cons ist of 12 hexadecimal digits, enclosed in double quotes. The value cannot be all zeros and cannot have the Multicast bit set. Use only the hexadecimal digits 2, 6, A, or E as the seco nd digit of the twelve required in the NODE ADDRESS parameter. Use any hexadecimal digits in all other positions. o Possible Errors: If any of the rules described above is violated , an error message indicates the rules for forming a proper address, and the driver treats this as a fatal error. ⢠CACHEFLUSH o Syntax: CACHEFLUSH= [0|1] o Example: CACHEFLUSH=1 o Default: 0 o Normal Behavior: Use this parame ter during a remote installation or "push-install" of Windows NT®. o Possible Errors: Any non-zero valu e sets this parameter to 1. The driver does not give any outwards indication as to th e state of this parameter. Example PROTOCOL.INI file PROTOCOL.I NI sectio n for the HP Gi ga bit Server Adapters and Modules: ⢠DRIVERNAME parameter o Use this parameter to set 1st adapter = N1000$, 2nd adapter = N10002$, and so on o DRIVERNAME = N1000$ ⢠NODE parameter o Use this to change the MAC address us ed by the adapter o NODE="02AA00123456" ⢠SLOT parameter o Use this to configure a particular instance of th e driver designated by DRIVERN AME, to use the adapter in a particular PCI slot. Slots should be specified in hex notation.
Adapter configurations 89 o SLOT=0x7 ⢠CACHEFLUSH parameter o Set this to 1 to hook the system Disk Buffer Flush interrupt. This causes the driver to properly tur n off the adapter when the system is shut down by Windows NT® during remo te installation (push- install). o CACHEFLUSH=1 o Known Limitations: This driver doe s not suppor t any application that uses protected mode, for example, EMM386. In addition, testing has demons trated that the GHOST software application fails when used with this driver under Netx (VLMs are OK). PXE configuration for HP NC-Series Intel adapters The Pre-boot Execution Environment (PXE) of HP adap ters has been upgraded to allow PXE functions to work with the HP ProLiant servers. The Boot Agent software provides the ability to custom ize it s behavior through its configurati on options. Changing the Boot Agent configuration is required only the first time you add a new network adapter to a client computer within your network. You can configure the Boot Agent within any of the followin g environments: ⢠A pre-boot environment (before an operating syst em is loaded) ⢠A Windows® environment ⢠A DOS environment Configuring the Boot Agent in a pre-boot environment You can customize the behavior of the B oot Agent software through a pre-boot (operating system- independent) configuration program contained within the adapter's fl ash ROM. You can access this pre- boot configuration setup program each time the cl ient computer cycles through the boot process. When you power up or perform a hard or soft system reset, the following message is displayed during the initialization process of the Boot Agent firmware . It indicates that the boot agent is active. Initializing Intel(R) Boot Agent Version 3.0.05 PXE 2.0 Build 078 (WFM 2.0), RPL V2.73 The following message is then displayed duri ng the boot process to Boot A gent Services: Intel(R) Boot Agent 3.0.05 Copyright(c) 1997-2000, Intel Corporation Client MAC Address: 00 50 8B xx xx xx GUID:00000000-0000-0000-000000000000 Client IP: 123.xxx.xxx.xxx MASK: 123.xxx.xxx.xxx DHCP IP: 123.xxx.xxx.xxx Gateway IP: 123.xxx.xxx.xxx Changing the Boot Agent configuration settings To access the Boot Agent Setup Menu, immediately after the initialization messages, and befor e any subsequent message is displayed, quickly press Ctrl S . This menu provides the ability to customize the "Boot Agent Configuration Settings." The Boot Agent setup menu displays a list of configur ation setting s on the left, and their corresponding values on the right. Key functions are described near the bottom of th e menu. The menu indicates how to change the values of settings. To access a brief Help descript ion of a setting, select it.
Adapter configurations 90 To customize the Boot Agent configuration settings: 1. Highlight the setting you want to chan ge using the arrow keys. Then press Enter. 2. Press the spacebar until the desired value appears. 3. If you want to change additional settings, repeat steps 1 and 2. 4. After you have completed your changes, press F4 to update the adapter with the new values. Boot Agent configuration settings The following table describes the Boot Agent config uration settings. Configuration settings Possible values Description Network Boot Protocol ⢠PXE (Preboot eXecution Environment) ⢠RPL (Remote Program Load) Controls whether the RPL or PXE boot protocol will be used. Select PXE for use with WfM-compatible network management programs, such as Intel® LANDesk® Manage ment Suite, Windows 2000 RIS, and Linux®. Select RPL for legacy-style remote booting, as well as for Novell® Netware® remote boot solutions. Depending on the configuration of the Boot Agent, this parameter may not be changeable. Boot Order ⢠Use BIOS Setup Boot Order ⢠Try network first, then local drives ⢠Try local drives first, then network ⢠Try network only ⢠Try local drives only Sets the boot order in which devic es are selected during boot up if the computer does not have its own control method. If your client computer's BIOS supports the BIOS Boot Specification (BBS), or allows PnP-compliant selection of the boot order in the BIOS setup program, then this setting will always be Us e BIOS Set up Boot Order and cannot be changed. In this ca se, refer to the BIOS setup manual specific to your client computer to set up boot options. If your client computer does not have a BBS- or PnP-compliant BIOS, you can select any one of the othe r possible values listed for this setting except for Use BIOS Setup Boot Order. Show Setup Prompt ⢠Enabled ⢠Disabled Controls whether or not the Boot Agent displays the Ctrl S prompt after POST. If you select Enabled, the Ctrl S prompt appears after POST so that you can press Ctrl S within the number of seconds allowed by the Setup Menu Wait Time setting to display the configuration setup menu. If you select Disabled and/or the Setup Menu Wait Time setting is zero, the Ctrl S setup prompt will not appear during the boot process. However, you can display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedly pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again.
Adapter configurations 91 Configuration settings Possible values Description Setup Menu Wait Time ⢠0 seconds ⢠2 seconds ⢠3 seconds ⢠5 seconds Controls the number of seconds the Boot Agent waits for you to press the Ctrl S keys, so as to suspend the boot process and then configure the way the Boot Agent operates. If you select a Setup Menu Wait Time setting of zero and/or a Show Setup Prompt setting of Disabled, the Ctrl S setup prompt will not appear during the boot process. Ho wever, you can still display the configuration setup menu by repeatedly pressing the Ctrl S keys immediately after POST until the menu appears. If the configuration setup menu does not appear after repeatedl y pressing the Ctrl S keys, you were likely not fast enough. In this case, reboot and try again. Windows Modifying adapter properties using the NCU This following describes the adapter properties available in the HP Ne twork Configuration Utility (NCU). IMPORTANT: NC31xx, NC6132, NC6133, and NC7132 adapters are not supported in Windows Server 2003 x64 . HP 10/100 server adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default = Auto/Auto o Range = Auto/Aut o; 10/Half; 10/Full; 100/Half; 100/Full NOTE: The NC3133 is a 100BASE-FX fiber optic upg rade module that supports 100 Mbps only. ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Adaptive Performance Tuning: NIC Bandwidth/CPU Utilization (NC31xx adapters only). Specifies the number of frames the adapter receives before trig gering an interrupt. Under normal operation, the adapter generates an interrupt each time a frame is received. Reducing the number of interrupts improves CPU use. Setting a low value for this param eter causes a high rate of receive interrupts (that is, a setting of the lowest value will result in an interrupt for every frame recei ved). A low value increases adapter bandwidth, but may reduce CPU effi ciency, slowing your computer. Setting a high value for this parameter reduces the interrupt rate (that is, increases the number of frames the
Adapter configurations 92 adapter receives before generating an interrupt). A high valu e improves CPU efficiency, but may reduce adapter bandwidth. o Default = 1536 o Range = 0â4096 (increments of 32) ⢠Number of Coalesce Buffers (NC31xx adapters only). Specifies the number of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map registers are available, the driver is forced to queue the packe t for later transmission. The preferred method of transmitting data is to use map registers, because it is the most efficient method. o Default = 8 o Range = 1â 32 ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 48 o Range = 8â1024 (in increments of 8) ⢠Transmit Control Blocks. Specifies how many Transmit Control Blocks (TCBs) the driver allocates for adapter use. This directly corresponds to the number of outstanding packets the driver can have in its send queue. If too few TCBs are used, performance suffers. If too ma ny TCBs are used, the driver unnecessarily consumes memory resources. o Default = 32 o Range = 1â64 ⢠Adaptive Transmit Threshold. Determines how many bytes are in the FIFO before transmission begins. o Default = 200 o Range = 0â200 ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The recommended selection is Off and it is t he default se tting. To disable Flow Control, select Off from the Value list on the Advanced tab. o Off. (Default) PAUSE frame receipt and transmission is disabled o Generate. PAUSE frame transmission is enabled o Respond. PAUSE frame receipt is enabled o Both on. PAUSE frame receipt and transmission is enabled HP NC110x, NC310F, NC340x, NC360x, NC36 4x, NC61xx, and NC71xx server adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default 1000BASE-SX or -LX Adapters: Auto/Auto 10/100/1000 Adapters: Aut o/Auto
Adapter configurations 93 o Range 1000BASE-SX or -LX Adapters: Auto/Auto 10/100/1000 Adapters: Aut o/Auto; 10/Hal f; 10/Full; 100/Half; 100/Full; 1000/Full NOTE: This property is disabled fo r the NC310F and NC61xx adapters. ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. o Default = 1514 Bytes o Range NC310F and NC61xx: 1514 Bytes; 40 88 Bytes; 9014 Bytes; 16128 Byt es NC110x, NC340x, NC360x, NC364x, and NC71xx: 1514 Bytes; 4088 Byt es; 9014 Bytes ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default Windows® 2000: 256 (NC110x and NC36 4x not supported) Windows Server 2003 and Windows Server 2003 x6 4: 128 o Range = 16â768 (increments of 8) ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default Windows® 2000: 256 (NC110x and NC36 4x not supported) Windows Server 2003 Wi ndows Server 2003 x64 NC110x, NC310F, NC340x, NC360x, NC364x, N C6170, and NC7170: 256 All other adapters: 160 o Range Windows® 2000
Adapter configurations 94 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) Windows Server 2003 Wi ndows Server 2003 x64 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) NC110x, NC360x, and NC364x: 80-2048 (increments of 8) ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 256 o Range Windows® 2000 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) Windows Server 2003 NC6132, NC6133, and NC6134: 80â76 8 (increments of 8) NC6136, NC7131, and NC7132: 80â25 6 (increments of 8) NC310F, NC340x, NC6170, and NC7170: 80â40 96 (increments of 32) NC110x, NC360x, and NC364x: 80-2048 (increments of 8) ⢠Rx TCP Checksum Offload. Offloads the task of computing the checksum for incoming TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx TCP Checksum Offload. Offloads the task of computing the checksum for outgoing TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx IP Checksum Offload. Offloads the task of computing th e checksum for outgoing IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Rx IP Checksum Offload. (NC110x, NC310F, NC340 x, NC360x, NC364x, NC6170, and NC7170 only). Offloads the task of computing the checksum for incoming IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled
Adapter configurations 95 ⢠Enable PME. (NC110T, NC310F, NC340x, NC360x, NC364T, NC61 70, and NC7170 only) Allows the use of power management and wake-up functions. o Default = No Action o Range = Disable; Enable; No Action; H/W Default ⢠Wake On. (NC110T, NC310F, NC340T, NC360x, NC364T, NC6170, and NC71 70 only) Available only when Enable PME is enabled. Allows you to choose what types of packets will cause the adapter to wake. o Default = Disable (NC110, NC360x, NC364x Defa ult=OS Controlled) o Range = Disable; Magic Packet; Direct Packe t Magic and Direct Packets; OS Controlled ⢠Large Send Offload. Enables offloading of large TCP packets (NC11 0T, NC310F, NC340x, NC360x, NC364T, NC6170, and NC7170 only). Windows 2003 and Windows 2003 x64 only. o Default = Enabled o Range = Disabled; Enabled ⢠Gigabit Master Slave Mode. (NC110x, NC360x, NC3 64x only) Determines IEEE 802.3ab Master- Slave resolution during gi gabit auto negotiation. o Default = Auto Detect o Range = Auto Detect, Force Master Mode, Force Slave Mode ⢠Smart Power Down. (NC110T, NC360x, and NC364T only) Minimizes power consumpti on by enabling the adapter to enter a deep sleep mode under certain conditions. o Default = H/W Default o Range = Disable, H/W Default, Enable ⢠Wake on Link. (NC110T, NC360x, and NC364T only) Allows wake-up from Advanced Configuration and Power Interface (ACPI) when link is reconnected. Windows 2003 and Windows 2003x64 o Default = Disable o Range = Disable, Force ⢠Flow Control. (Windows Server 2003 only) Determines the type of fl ow control that is needed. (NC110x, NC360x, NC364x only): o Default = Both On o Range = Off, Respond, Generate, Both On HP NC150T, NC32x, NC37xx, NC380T, NC1 020, NC67xx, and NC77xx Server Adapters ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 200 o Range = 100â600 NOTE: This property is not supported on HP Multifunction Server Adapters. ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of
Adapter configurations 96 system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 200 o Range = 100â512 NOTE: This property is not supported on HP Multifunction Server Adapters. ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o NC150T and NC67xx Range = Auto/Aut o Default = Auto/Auto o NC32x, NC37xx, NC380T, NC1020, and NC77xx Range = Auto/Aut o; 10/Half; 10/Fu ll; 100/Half; 100/Full; 1000/Full Default = Auto/Auto ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. When TCP/IP Offload Engine (TOE) is enabled, th is property cannot be set to greater than 1514 bytes. If this property is set to grea ter than 1514 bytes, TOE cannot be enabled. o Default = 1514 Bytes o Range = 1514 Bytes; 4088 Bytes; 9014 B ytes NOTE: This property is not supporte d by NC150T, NC32x, NC1020, and NC7761 adapters. ⢠Wake Up Capabilities. Determines the type of packets that will cause the adapter to wake up the system. o Default = Both o Range = None; Magic Packet; Wake Up Frame; Both NOTE: This property is not supported on HP Multifunction Server Adapters. ⢠WOL Speed. (Not supported on NC150T, NC37x, NC380T, NC67xx, and NC7782 adapters) Specifies the speed at which the adapter connects to the network during Wake-on- LAN mode.
Adapter configurations 97 o Default = Auto o Range = Auto, 10Mb, 100Mb ⢠Flow Control. Specifies the type of flow control that is needed. o Default = Auto o Range = Disable; Rx Pause; Tx Pause; Rx/Tx Pause; Auto ⢠Rx TCP Checksum Offload. Offloads the task of computing the checksum for incoming TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx TCP Checksum Offload. Offloads the task of computing the checksum for outgoing TCP or UDP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Tx IP Checksum Offload. Offloads the task of computing th e checksum for outgoing IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Rx IP Checksum Offload. Offloads the task of computing th e checksum for incoming IP packets, thereby improving performance. o Default = Enabled o Range = Disabled; Enabled ⢠Ethernet@WireSpeed (NC37xx, NC380x and NC77xx only). Enables linking at 100 Mbps when using a cable that does not support Gigabit speed. o Default = Enabled o Range = Disabled; Enabled NOTE: This property is not su pported by NC7761 adapters. ⢠Large Send Offload. Normally, the TCP segmentat ion is performed by the protocol stack. By enabling the Large Send Offload parameter, the TCP segmentation can be performed by the Gigabit Ethernet Adapter. To enable the Large Send Offload parameter, select Enable fr om the Value list on the Advanced tab. o Default = Enabled o Range = Disabled; Enabled ⢠IPMI . Enables Intelligent Platform Management Inte rface, which provides a means for monitoring, control, and automatic recovery of servers. IPMI can only enabled on one port at a time. IPMI cannot be enabled on a port that has Large Send Offload enabled. (This property is only supported on the NC320i, NC7781 and NC7782 adapters.) Multifunction Gigabit Server Adapters only ⢠iSCSI properties
Adapter configurations 98 o SRB Timeout Delta. Increases or decreases the tim eout value for I/O operations through the iSCSI HBA. Increasing this value may help prevent prem ature I/O timeouts if yo ur network is highly congested and packets are being dropped frequently. The default setting is 15 seconds. This property only displays if iSCS I is enabled on the adapter. o iSCSI VLAN. Allows you to create a VLAN for an iSCSI-enabled device. ⢠TCP/IP Offload Engine (TOE). Offloads TCP connections to th e Multifunction Server Adapter. (Microsoft Scalable Networking Pack (SNP) fo r Windows 2 003 SP1 or later and Windows 2003 x64 SP1 or later environments only.) See the Microsoft ( http://www.microsoft.com ) website to download the latest version of the Microsoft Wind ows Server 2003 Scalable Networking Pack (SNP) and the required Microsoft hotfixes in arti cle numbers 921136, 919948, and 923187 o f the Microsoft Knowledge Base (KB). o Default = Enabled o Range = Disabled; Enabled ⢠Receive-Side Scaling (RSS). Allows the network load from a ne twork adapter to be balanced across multiple CPUs. (Microsoft Scalable Networking Pack (SNP) for Windows 2003 SP1 or later and Windows 2003 x64 SP1 or later) o Default = Enabled o Range = Disabled; Enabled ⢠Optimized Interrupt Moderation . When this configuration is set to optimize, the driver automatically and continuously determines the best Target DPC (d eferred procedure call) Rate to use based on the current network conditions. When th is configuration is set to manual , the user can select a Target DPC Rate and the driver adjusts the adapter's inte rrupt setting to achieve that DPC rate. Changing this setting may improve performance for some conf igurations. ⢠Target DPC Rate . Determines the desired DPC rate. The driver adjusts the adapter's interrupt moderation setting dynamically bas ed on network conditions to achieve the desired DPC rate. This configuration is valid only when "Optimized Interrupt Moderation" is set to manual. Modifying HP NC-Series Broadcom and Multifunction adapter properties Although the default values should be appropriate in most cases, you may change any of the available options to meet the requirements of your specific system. Use the followin g procedure to verify or change the properties after the adapter driver is installed. To change the adapter properties, display the Driver Properties Advanced tab as follows: 1. On the desktop, click Start then select Control Panel > System to display the System Properties window. 2. Select the Hardware tab and click Device Manager . The Device Manager window appears. 3. Scroll down the list of hardware devices to Network Adapters and click the plus ( ) . This displays the list of adapters currently configured. 4. Right-click the adapter to configure and select Properties. The Properti es window appears. 5. Select the Advanced tab. The list of properties and cu rrent values for the adapter appears. 6. Select the property to modify and adjust the Value as needed. 7. Save the settings. a. When the adapter property conf iguration is complete, click OK to accept the settings.
Adapter configurations 99 b. If prompted to restart your computer, click Yes . Rebooting the system for new adapter properties to take effect is not required; however, rebooting is recommended to re-initialize all regi sters. Windows Server 2003 and Windows Server 2003 x64 adapter properties ⢠802.1p QoS Packet Tagging . The 802.1p QoS parameter is a standard that enables Quality of Service. This property is disabled by de fault. To change this property, select Enable or Disable. ⢠Checksum Offload. Normally the Checksum Offload function is computed by the protocol stack. By selecting one of the Checksum Offload properti es , the checksum can be computed by the Gigabit Ethernet Adapter. To enable one of the Checksum Offload properties s elect the property from the Value list on the Advanced tab. o None. Disables checksum offloading o Rx TCP/IP Checksum. Enables receive TC P, IP, and UD P checksum offloading o Tx TCP/IP Checksum. Enables transmit TCP, IP, and UDP checksu m offloading o Tx/Rx TCP/IP Checksum. (Default) Enables transmit and receive TCP, IP, and UDP checksum offloading ⢠Ethernet@WireSpeed. This property is only available throug h the HP Network Configuration Utility. The Ethernet@WireSpeed property enables a 1000Bas e-T Ethernet adapter to establish a link at a lower speed when only two pairs of wires are available in the cable plant. By default, this property is enabled. To disable the Ethernet @ WireSpeed prop erty, select Disable from the Value list. o Disable. Disables Ethernet@Wir eSpeed o Enable. (Default) Enables Ethernet@WireSpeed ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The recommended selection is Auto and it is the default setting. To dis able Flow Control, select Disable from the Value list on the Advanced tab. o Auto. (Default) Auto detect flow control o Disable. PAUSE frame receipt an d transmission is disabled o Tx PAUSE. PAUSE frame transmission is enabled o Rx PAUSE. PAUSE frame receipt is enabled o Rx/Tx PAUSE. PAUSE frame receip t and transmission is enabled ⢠Large Send Offload. Normally, the TCP segmentat ion is performed by the protocol stack. By enabling the Large Send Offload property, the TC P segmentation can be performed by the Gigabit Ethernet Adapter. To enable the Large Send Offl oad property, sele ct Enable from the Value drop- down menu on the Advanced tab. o Disable. Disables large send offloading o Enable. Enables large send offloading ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = Not Present o Range = Value; Not Present
Adapter configurations 100 ⢠Speed & Duplex. The Speed & Duplex property allows the user to s et the connection speed to the network and mode. The Duplex Mode allows the adapter to transmit and receive network data simultaneously. The adapter is set to Auto (optimum connection) by default. Set the speed and mode as described below: o 1 Gb Full Auto. Sets the speed to 1 Gbps and mode to Full-Duplex o 10 Mb Full. Sets the speed to 10 Mbps and mode to Full-Duplex o 10 Mb Half. Sets the speed to 10 Mbps and mode to Half-Duplex o 100 Mb Full. Sets the speed to 100 Mbps and mode to Full -Duplex o 100 Mb Half. Sets the speed to 100 Mbps and mode to Half-Duplex o Auto. (Default) Sets the speed and mode fo r optimum network connection (recommended) NOTE: Auto is the recommended selection. It allo ws the adapter to dynamically detect the line speed and duplex mode of the network. Wh enever the network capability changes, the adapter will automatically detect and adjust to the new line sp eed and duplex mode. The Half-Duplex selection forces the adapter to connect to the network in Half-Duplex mode. The adapter may not function if the network is not configured to operate at the same mode. The Full-Duplex selection forc es the adapter to connect to the network in Full-Duplex mode. The adapter may not function if the network is no t configured to operate at the same mode. ⢠Wake Up Capabilities. This property is only available throug h the HP Network Configuration Utility. The Wake Up Capabilities property allows the user to set the adapte r to wake up from a low power mode when it receives a networ k wake up frame. Two wake up frames are possible: Magic Packet and Wake Up Frame. By default, the adapter is set to Both. To choose the type of frame that will cause the adapter to wake up, select the appropriate option from the Value list on the Advanced tab. o Both. (Default) Selects both Magic Packet and Wake up Frame as wake up frames o Magic Packet. Selects Magic Pa cket as the wake up frame o None. Selects no wake up frames o Wake Up Frame. Selects Wake up Frame as the wake up frame ⢠WOL Speed. This property is only available through the HP Network Confi guration Utility. The WOL Speed property allows the user to select t he sp eed at which the adapter connects to the network while in Wake-on-LAN mode. By default, the adapter is set to Auto. o 10 Mb. Sets the speed to 10 Mb o 100 Mb. Sets the speed to 100 Mb o Auto. (Default) Sets the speed for optimum network connection ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers are alloca ted).This property is only available through the HP Network Configuration Utility. o Default = 200 o Range = 100â600 ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too
Adapter configurations 101 many receive descriptors are used, the driver un necessarily consumes me mory resources. This property is only available through the HP Network Configuration Utility. o Default = 200 o Range = 100â512 Multifunction Gigabit Server adapter properties ⢠SRB Timeout Delta. Increases or decreases the tim eout value for I/O operations through the iSCSI HBA. Increasing this value may help prevent prem ature I/O timeouts if yo ur network is highly congested and packets are being dropped frequently. The default setting is 15 seconds. This property only displays if iSCSI is enabled on the adapter. iSCSI capabilities are only available through HP Multifunction Gigabi t Server Adapters and Multifunction Server Adapters for HP BladeSystem running Windows 2000, Window s Se rver 2003, and Windows Server 2003 x 64. ⢠Receive-Side Scaling (RSS). Allows the network load from a ne twork adapter to be balanced across multiple CPUs. (Microsoft Scalable Networking Pack (SNP) for Windows 2003 SP1 or later and Windows 2003 x64 SP1 or later) ⢠TCP Offload Engine (TOE). Offloads TCP connections to the Mult ifunction Server Adapter. (Microsoft Scalable Networking Pack (SNP) for Windows 2003 SP1 or later an d Windows 2003 x64 SP1 or later environments only.) See the Microsoft ( http://www.microsoft.com ) website to download the latest version of the Microsoft Windows Serv er 2003 Scalable Networking Pack (SNP) and the required Microsoft hotfixes in article num bers 921136, 919948, an d 923187 of the Microsoft Knowledge Base (KB). Windows 2000 adapter properties ⢠802.1p QoS. The 802.1p QoS parameter is a standard th at enables Quality of Service. This parameter is disabled by default. To change this parameter, select Enable or Disable . ⢠Checksum Offload. Normally the Checksum Offload function is computed by the protocol stack. By selecting one of the Checksum Offload parameter s , the checksum can be computed by the Gigabit Ethernet Adapter. To enable one of the Checksum Offload parameters, select the parameter. o None. Disables checksum offloading o Rx TCP/IP Checksum. Enables receive TC P, IP, and UD P checksum offloading o Tx TCP/IP Checksum. Enables transmit TCP, IP, and UDP checksu m offloading o Tx/Rx TCP/IP Checksum. (Default) Enables transmit and receive TCP, IP, and UDP checksu m offloading ⢠Ethernet@WireSpeed. The Ethernet@WireSpeed parameter enables a 1000Base-T Ethernet adapter to establish a link at a lower speed when only two pairs of wires are available in the cable plant. By default, this parameter is en abled. To disable the Ethernet@WireSpeed parameter, select Disable from the Value list on the Advanced tab. o Disable. Disables Ethernet@Wir eSpeed o Enable. (Default) Enables Ethernet@WireSpeed NOTE: This property is not supported by NC1020 and NC7761 adapters. ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The
Adapter configurations 102 recommended selection is Auto and it is the default setting. To dis able Flow Control, select Disa ble from the Value list on the Advanced tab. o Auto. (Default) Auto detect flow control o Disable. PAUSE frame receipt an d transmission is disabled o Tx PAUSE. PAUSE frame transmission is enabled o Rx PAUSE. PAUSE frame receipt is enabled o Rx/Tx PAUSE. PAUSE frame receip t and transmission is enabled ⢠Speed & Duplex. The Speed & Duplex parameter allows the user to set the connection speed to the network and mode. The Duplex Mode allows the adapter to transmit and receive network data simultaneously. The adapter is set to Auto (optimum connection) by default. Set the speed and mode as described below: o 10 Mb Full. Sets the speed to 10 Mbps and mode to Full-Duplex o 10 Mb Half. Sets the speed to 10 Mbps and mode to Half-Duplex o 100 Mb Full. Sets the speed to 100 Mbps and mode to Full -Duplex o 100 Mb Half. Sets the speed to 100 Mbps and mode to Half-Duplex o Auto. (Default) Sets the speed and mode fo r optimum network connection (recommended) NOTE: 1 Gbps speed is enabled by selecting Auto. NOTE: Auto is the recommended selection. It allo ws the adapter to dynamically detect the line speed and duplex mode of the network. Wh enever the network capability changes, the adapter will automatically detect and adjust to the new line sp eed and duplex mode. The Half-Duplex selection forces the adapter to connect to the network in Half-Duplex mode. The adapter may not function if the network is not configured to operate at the same mode. The Full-Duplex selection forc es the adapter to connect to the network in Full-Duplex mode. The adapter may not function if the network is no t configured to operate at the same mode. ⢠Wake Up Capabilities. The Wake Up Capabilities parameter a llows the user to set the adapter to wake up from a low power mode when it receives a network wake up frame. Two wake up frames are possible: Magic Packet and Wake Up Frame. By default, the adapter is set to Both. To choose the type of frame that will cause the adapter to wake up, select the appropriate option from the Value drop-down list on the Advanced tab. o Both. (Default) Selects both Magic Packet and Wake up Frame as wake up frames o Magic Packet. Selects Magic Pa cket as the wake up frame o None. Selects no wake up frames o Wake Up Frame. Selects Wake up Frame as the wake up frame NOTE: This property is not su pported by NC150T adapters. ⢠WOL Speed. The WOL Speed parameter allows the user to select the spe ed at which the adapter connects to the network while in Wake-on-LAN mode . By default, the adapter is set to Auto. o 10 Mb. Sets the speed to 10 Mb o 100 Mb. Sets the speed to 100 Mb
Adapter configurations 103 o Auto. (Default) Sets the speed fo r optimum network connection NOTE: This property is not supported on the NC150T and NC7782 adapters. Multifunction Gigabit Server adapter iSCSI properties ⢠SRB Timeout Delta. Increases or decreases the tim eout value for I/O operations through the iSCSI HBA. Increasing this value may help prevent prem ature I/O timeouts if yo ur network is highly congested and packets are being dropped frequently. The default setting is 15 seconds. This property only displays if iSCSI is enabled. Modifying HP NC-Series Intel adapter properties Although the default values should be appropriate in most case s, you may change any of the available options to meet the requirements of your specific system. Use the followin g procedure to verify or change the properties after the adapter driver is installed. To change the adapter properties, display the Driver Properties Advanced tab as follows: 1. On the desktop, click Start then select Control Panel > System to display the System Properties window. 2. Select the Hardware tab and click Device Manager . The Device Manager window appears. 3. Scroll down the list of hardware device s to Network Adapters and click the plus ( ). This displays the list of adapters currently configured. 4. Right-click the adapter to configure and select Properties. The Properti es window appears. 5. Select the Advanced tab. The list of properties and cu rrent values for the adapter appears. 6. Select the property to modify and adjust the Value as needed. 7. Save the settings. a. When the adapter property conf iguration is complete, click OK to accept the settings. b. If prompted to restart your computer, click Yes . Rebooting the system for new adapter properties to take effect is not required; however, rebooting is recommended to re-initialize all regi sters. Adapter properties ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p ta gging to send network traffic with different priority levels. o Default = Disabled o Range = Disabled; Enabled ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 128 (for NC110x, NC360x, NC364x = 25 6) o Range = 16â768 (increments of 8) ⢠Flow Control. The Flow Control property allows the use r to enable or disable the receipt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The
Adapter configurations 104 recommended selection is Auto and it is the default setting. To dis able Flow Control, select Disa ble from the Value list on the Advanced tab. o Auto. (Default) Auto detect flow control o Disable. PAUSE frame receipt an d transmission is disabled o Tx PAUSE. PAUSE frame transmission is enabled o Rx PAUSE. PAUSE frame receipt is enabled Rx/Tx PAUSE. PAUSE frame receip t and transmission is enabled NOTE: User control of Flow Control parameter is not available for NC71xx adapters. ⢠Link Speed & Duplex. Allows the user to set the connectio n speed to the network and mode. Full- Duplex Mode allows the adapter to transmit and re ceive network data simultaneously. This property is not available for NC310F and NC61xx adapters through the Microsoft® User Interface. o Default = Auto Detect o Range = NC340x and NC71xx Adapters: Auto Detect; 10Mbps/Half-Duplex; 10Mbp s/Full- Duplex; 100Mbps/Half-Duplex; 100Mbps/ Full-D uplex; 1000 Mbps/Full-Duplex ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = Not Present o Range = Value; Not Present ⢠Number of Receive Descriptors. Specifies the number of buffers used by the driver when copying data to the protocol memory. In high network lo ad situations, increasing receive desc riptors can increase performance. The tradeoff is that this also increases the amount of syste m memory used by the driver. If too few receive buff ers are used, perf ormance suf fers. If too many receive buffers are used, the driver unnecessarily consum es memory resources. o Default NC110x, NC310F, NC340x, NC360x, NC364x, N C6170, and NC7170 = 256 All others = 160 o Range NC6132, NC6133, and NC6134 = 80â768 (incr ements of 8) NC6136, NC7131, and NC7132 = 80â256 (incr ements of 8) NC310F, NC340x, NC6170, and NC7170 = 80â4 096 (increments of 8) NC110x, NC360x, and NC364x = 80-2048 (increments of 8 ) ⢠Number of Transmit Descriptors. Specifies the number of resources allocated to transmit packets. o Default = 256 o Range NC6132, NC6133, and NC6134 = 80â768 (incr ements of 8) NC6136, NC7131, and NC7132 = 80â256 (incr ements of 8) NC310F, NC340x, NC6170, and NC7170 = 80â4 096 (increments of 8) NC110x, NC360x, and NC364x = 80-2048 (increments of 8 )
Adapter configurations 105 ⢠Offload Receive TCP Checksum. Offloads the task of computing the checksum for incoming TCP or UDP packets, thereby improving performance. o Default = On o Range = Off; On ⢠Offload Transmit TCP Checksum. Offloads the task of computing the checks um for outgoi ng TCP or UDP packets, thereby improving performance. o Default = On o Range = Off; On ⢠Offload Transmit IP Checksum. This property does not apply to the NC6132, NC6133 or NC6134 adapters. Offloads the task of computing the checksum for outgoing IP packets, thereby improving performance. o Default = On o Range = Off; On ⢠Offload Receive IP Checksum. (NC110x, NC310 F, NC340x, NC360x, NC364x, NC6170, and NC7170 only). Offloads the task of computing the checksum for incoming IP packets, thereby improving performance. o Default = On (NC110, NC360x, NC364x = Off) o Range = Off; On ⢠Enable PME (NC110x, NC310F, NC340x, NC360x, NC364x, N C6170, and NC7170 only). Allows the use of power management and wake-up functions. o Default = No Action (NC110x, NC360x, NC364x = OS Controlled) o Range = Disabled; Enabled; Hardware Default; No Action (NC110x, NC360x, NC36 4x = OS Controlled, Disabled, Enabled) ⢠Wake On Settings (NC110x, NC310F, NC34 0x, NC360x, NC364x, NC6170, and NC7170 only). Available only when Enable PME is enabled. Al lows you to choose what types of packets will cause the adapter to wake. o Default = Disabled (NC110x, NC360x, NC364x = OS Controlled) o Range = Disabled; Wake on Directed Packet; Wake on Magic & Directed; Wake on Magic Packet; OS Controlled ⢠Large Send Offload (Windows® Server 2003 only; NC 110x, NC310F, NC340T, NC360x, NC364x, NC6170, and NC7170 only). Enables offloading of larg e TCP packets. o Default = On o Range = Off; On ⢠Smart Power Down. (NC110T, NC360x, and NC364T only) Minimizes power consumpti on by enabling the adapter to enter a deep sleep mode under certain conditions. o Default = Hardware Default o Range = Disable, Enable, Hardware Default ⢠Wake on Link. (NC110T, NC360x, and NC364T only) Allows wake-up from Advanced Configuration and Power Interface (ACPI) when link is reconnected. o Default = Disable o Range = Disable, Force
Adapter configurations 106 Modifying NC31xx Fast Ethernet adapter properties Although the default values should be appropriate in most cases, you may change any of the available options to meet the requirements of your specific system. Use the followin g procedure to verify or change the properties after the adapter driver is installed. To change the adapter properties, display the Driver Properties Advanced tab as follows: 1. On the desktop, click Start then select Control Panel > System to display the System Properties window. 2. Select the Hardware tab and click Device Manager . The Device Manager window appears. 3. Scroll down the list of hardware devices to Network Adapters and click the plus ( ) . This displays the list of adapters currently configured. 4. Right-click the adapter to configure and select Properties. The Properti es window appears. 5. Select the Advanced tab. The list of properties and cu rrent values for the adapter appears. 6. Select the property to modify and adjust the Value as needed. 7. Save the settings. a. When the adapter property conf iguration is complete, click OK to accept the settings. b. If prompted to restart your computer, click Yes . Rebooting the system for new adapter properties to take effect is not required; however, rebooting is recommended to re-initialize all regi sters. Adapter Properties ⢠802.1p QoS Packet Tagging . Enables or disables IEEE 802.1p tagging to send network traffic with different priority levels. o Default = Off o Range = Off; On ⢠Coalesce Buffers. Specifies the number of memory buffers avai lable to the driver in case the driver runs out of available map registers. This buffer ar ea is also used when a packet consists of many fragments. If no coalesce buffer s or map registers are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 8 o Range = 1â 32 ⢠Flow Control Settings. The Flow Control property allows the user to enable or disable the recei pt or transmission of PAUSE frames. PAUSE frames enab le the adapter and the switch to control the transmit rate. The side that is receiving the PA USE frame will moment arily stop transmitting. The recommended selection is Off and it is t he default se tting. To disable Flow Control, select Off from the Value list on the Advanced tab. o Default = Off o Range = Off, Generate, Re spond, Respond and Generate Off. PAUSE frame receipt and transmission is disabled Generate. PAUSE frame transmission is enabled Respond. PAUSE frame receipt is enabled Respond and Generate. PAUSE frame re ceipt and transmission is enabled
Adapter configurations 107 ⢠Link Speed & Duplex. Allows the user to set the connectio n speed to the network and mode. Full- Duplex Mode allows the adapter to transmit and receive network data simultaneously. o Default = Auto Detect o Range 100 Mbps/Full-Duplex: Sets the speed at 100 Mbps and mode to Full-D uplex 100 Mbps/Half-Duplex: Sets the speed at 100 Mbps and mode to Half-Dupl ex 10 Mbps/Full-Duplex: Sets the speed at 10 Mbps and mode to Full-Duplex 10 Mbps/Half-Duplex: Sets the speed at 10 Mbps and mode to Half-Duplex Auto Detect: Sets the speed and mo de for optimum network connection ⢠Locally Administered Address. Specifies the user-define d MAC Address of the adapter; overrides the burned-in MAC Address. This is a read-o nly field if the adapter is teamed. o Default = Not Present o Range = Value; Not Present ⢠Receive Buffers. Specifies the number o f buffers used by the driver when copying data to the protocol memory. In high network load situ ations, increasing receive buffers can increase p erformance. The tradeoff is that this also increases the amount of sy stem memory used by the driver. If too few receive buffers are used, performance suf fers. If too many receive buffers are used, the driver unnecess arily consumes memory resources. o Default = 48 o Range = 8â1024 (increments of 1) ⢠Transmit Control Blocks. Specifies the number of resources that are allocated to transmit packets. o Default = 16 o Range = 1â64 Modifying adapter properties in Windows NT This document describes the Windows NT® 4.0 adapter properties available for fiber and copper server adapters. Properties for 10/100 Co pper Server Adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default = Auto/Auto o Range = 10/Half; 10/Full; 100/Half; 100/Full; Auto /Auto NOTE: The NC3133 is a 100BASE-FX fiber optic upg rade module that supports 100 Mbps only. ⢠Locally Administered Address. Specifies the user-defin ed MAC Address of the adapter, which overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed. o Default = NULL o Range = Valid unic ast MAC Address
Adapter configurations 108 ⢠Adaptive Interframe Spacing. Compensates for excessive Ethernet p acket collisions on the network. The default setting works best for most c omputer s and networks by dynamically adapting to the network traffic conditions. However, in some rare cases, you ma y obtain better performance by manually setting the spacing value. Setting a valu e forces a static gap between packets. Increasing the value increases the delay between frames bein g transmitted. o Default = 1 o Range = 1â255 ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also used when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 8 o Range = 1â32 ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 32 o Range = 8â1024 (in increments of 8) ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 12 o Range = 1â16 ⢠Transmit Control Blocks. Specifies how many Transmit Control Blocks (TCBs) the driver allocates for adapter use. This directly corresponds to the number of outstanding packets the driver can have in its send queue. If too few TCBs are used, performanc e suffers. If too ma ny TCBs are used, the driver unnecessarily consumes memory resources. o Default = 16 o Range = 1â80 ⢠Transmit Threshold. Specifies the number of bytes before the PCI adapter emptie s its internal transmit FIFO onto the wire. The value is multiplied by 8 to produce the number of bytes. For example, if Transmit Threshold = 200, the number of bytes is 1600. This is greater than the maximum packet size for Ethernet. Consequently, the adapter will no t attempt early transmits. Although this is the safest setting, the best performance is ac hieved wh en the Threshold parameter is as low as possible without producing underruns. o Default = 200 o Range = 0â200 Properties for NC61 xx and NC71xx Server Adapters Exception: HP NC6170 and NC7170 Server Adapte rs are not supported under Windows NT® 4.
Adapter configurations 109 ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default 1000BASE-SX or -LX Adapters: 1000/Fu ll 10/100/100 Adapters: Auto/Auto o Range 1000BASE-SX or -LX Adapters: 1000/Fu ll 10/100/100 Adapters: Auto/Auto; 10/ Half ; 10/Full; 100/Half; 100/Full; 100/Full ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. o Default = 1514 Bytes o Range = 1514 Bytes; 4088 Bytes; 9014 B ytes; 16128 Bytes NOTE: All Gigabit team members must support maximum frame size for this parameter to be enabled. If any one team member is not supported, then it is disabled for the team. Size of the frame is determined by the lowest value for the team members. The 16128 sett ing applies for Fiber Gigabit Ethernet server adapters only. ⢠Number of Coalesce Buffers. Specifies the numbe r of memory buffer s available to the driver in case the driver runs out of available map registers. This buffer area is also us ed when a packet consists of many fragments. If no coalesce buffers or map regi sters are available, the driver is forced to queue the packet for later transmission. The preferred method of transmitting data is to use map registers, since it is the most efficient method. o Default = 128 o Range = 16â768 (increments of 8) ⢠Number of Receive Descriptors: Minimum/Maximum. Specifies the number of desc riptors used by the driver when copying data to the protocol memory . In high network load situations, increasing receive descriptors can inc rease performance. The tradeoff is that this also incr eases the amount of system memory used by the driver. If too few rece iv e descriptors are used, performance suffers. If too many receive descriptors are used, the driver unn ecessarily consumes memory r esources. o Default = 160 o Range NC6132, NC6133, and NC6134: 64â76 8 (increments of 8) NC6136, NC7131, and NC7132: 64â25 6 (increments of 8) ⢠Number of Transmit Descriptors: Minimum/Maximum. Specifies the number o f descriptors to allocate per Transmit Control Block (TCB). This value directly affects the number of map registers allocated for the adapter (the higher the number, the more map registers ar e allocated). o Default = 256 o Range NC6132, NC6133, and NC6134: 64â76 8 (increments of 8) NC6136, NC7131, and NC7132: 64â25 6 (increments of 8)
Adapter configurations 110 Properties for NC67 xx and NC77xx Server Adapters ⢠Speed/Duplex Setting. Changes the current speed and duplex settings. Be sure this setting is compatible with the link partner (examples: switch, hub). o Default = Auto/Auto o Range = Auto/Aut o; 10/Half; 10/Fu ll; 100/Half; 100/Full; 1000/Full ⢠Maximum Frame Size. Specifies the maximum number of bytes in a single packet. Larger frames can increase throughput and decrease CPU use by putti ng more data in each packet, thereby sending out fewer packets. This is equivale nt to Jumbo Frames and must comply with switch limitations. o Default = 1514 Bytes o Range = 1514 Bytes; 4088 Bytes; 9014 B ytes NOTE: All Gigabit team members must support maximum frame size for this parameter to be enabled. If any one team member is not supported, then it is disabled for the team. Size of the frame is determined by the lowest value for the team members. The 16128 sett ing applies for Fiber Gigabit Ethernet server adapters only. ⢠Flow Control. Specifies the type of flow control that is needed. o Default = Disable o Range = Disable; Rx Pause; Tx Pause; Rx/Tx Pause; Auto Novell Netware Advanced Server Program for NetWare The following describes how to install and co nfigure th e Advanced Server Program (ASP). Installing ASP IMPORTANT: Teaming NC150, NC32x, NC37x, NC1020, NC6770, or NC77xx adapters with NC110x, NC310, NC31xx, NC340x , NC360x, NC364x, NC61xx, or NC71xx adapters is not supported for NetWare. Load QASP.LAN just like a standard LAN driver with all necessary frame types for the team. QASP.LAN requires a special VSLOT parameter to specify the virt ual slot. The virtual slot can be viewed as a team number. NOTE: Before loading the network drivers, the recommended sequence is to loa d QASP.LAN. This allows QASP.LAN to determine the initial li nk state of the bound adapters without delay. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 If you are using the NetWare NIC agent ( for example, Insight Manager) and teaming, you must load BMAPI.NLM. Loading network drivers To load the network drivers:
Adapter configurations 111 1. Load the network drivers for the network adapters th at will be part of the team. The frame types loaded should be the same for all adapters in the team and the same as those loaded for QASP.LAN. Do not bind protocols directly to these adapters. 2. Load ODINEB.NLM (a Novell supplied NLM) before all network drivers. Example: LOAD ODINEB.NLM LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 3. Load QASP.LAN. a new screen is displayed with all virtual adapter settings and statistics. Press the Alt Esc keys to switch back to the console and continue with step 3. 4. Bind QASP.LAN to the network adapters in the team by using a custom QASP BIND command at the console. Example: BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII If there are multiple frame types loaded on the virt ual and the physical adapters, it is only necessary to bind one frame type on the virtual adapter to the same frame type on the physic al adapter. The other frame types will be automatically bound. 5. Bind protocols to QASP.LAN. Example: BIND IP QASP_1_EII ADDR=x.x.x.x MASK=x.x.x.x NOTE: Configuration of QASP.L AN should be performed manually by editing the AUTOEXEC.NCF file. NWCONFIG.NLM (or INSTALL.NLM) cannot completely configure QASP.LAN. Uninstalling ASP To uninstall the adapter and the QASP drivers: 1. For the adapter driver, at the Command Line Interface (CLI) enter the following command: UNLOAD Q57 The response is: Q57 "Driver Name" unload Module Q57.LAN unloaded 2. For the QASP driver, at the Command Line Interface (CLI) ent er the following command: UNLOAD QASP The response is: Module QASP.LAN unloaded NOTE: QASP cannot be unloaded if one or more adapters are bound to QASP. SLB and TRUNK MODE selection Use "MODE=SLB" for Smart Load-B alance mode or "MODE=TRUNK" for Generic Trunking mode. The default is SLB. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 MODE=TRUNK
Adapter configurations 112 NOTE: In SLB mode, IPX traffic is load-balanced on the send side but not on the receive side. Loading frame types After one or more network adapters are bound to a vi rtual adapter, additional frame types can only be loaded in the virtual adapter if the corresponding fram e types are also loaded in the bound adapters. For example, ETHERNET_802.2 can be loaded in QASP VS LOT 1 only if ETHERNET_802.2 is loaded for the Q57 driver in SLOT 1 and 2. Similarly, a virtual adapter can only be bound to a physical adapter if the ph ysical adapter has all the frame types loaded in the virtual adapter. Example: LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_802.2 NAME=QASP_E82 VSLOT=1 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_1_E82 SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_2_E82 SLOT=2 Hot STANDBY In Smart Load-Balance (SLB) mode, one or more adapters can be designated as hot standbys. Use the keyword STANDBY in the QASP BIND command to indicate binding an adapter as a hot standby. Standby is only valid for Smart Load-Balance mode. Example: BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII STANDBY In the above example, Q57_1_EII and Q57_2_EII are bound as primary and hot standby adapters respectively. Configuring VLANs NOTE: VLANs are not supported on non-HP adapters. To add VLANs to a team: 1. Load QASP.LAN with the all necessary frame types and specify the VLAN ID for ea ch frame type. You can specify a maximum of 64 VLAN IDs and each VLAN ID can be loaded up to four times with four different frame types. Example: (VLAN ID 2 for Ethernet II) LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 2. Load the network drivers for the adapters in the team with all the frame type s specified in step 1. One or more VLAN IDs specified in step 1 do no t have to be specified when loading the network drivers. Each frame type loaded in step 1 only needs to be loaded once for each network driver even if in step 1 it is loaded multiple tim es with different VLAN IDs. Example: LOAD ODINEB.NLM LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 3. Bind QASP.LAN to the adapters in the team for each protocol. Example: BIND QASP1 Q57_1_EII
Adapter configurations 113 BIND QASP1 Q57_2_EII If there are multiple VLANs (each with one or more frame types) loaded on the virtual adapter, it is only necessary to bind one frame type on one VLAN on the virtual adapter to the same frame type on the physical adapter. The other VLANs will be automatically bound. 4. Bind protocols to QASP.LAN. Example: BIND IP QASP_1_V2_EII ADDR=x.x.x.x MASK=x.x.x.x This example creates a team with two adapters us ing VLAN ID 2. Outbound packets will be tagged with VLAN ID 2 and only similarly tagged packets will be received by the adapters in the team. Additional VLANs with different VLAN IDs can be created in the same team. The maximum number of VLANs per virtual slot is 64. The valid range of VLAN IDs is from 1 to 4094. VLAN=0 indicates the VLAN is untagged and is the default. Us e decimal numbers to specify the VLAN ID. The following are examples of multiple VLAN configura tions: LOAD QASP FRAME=ETHERNET_II NAME=QASP_1_V100_EII VSLOT=1 VLAN=100 LOAD QASP FRAME=ETHERNET_II NAME=QASP_1_V200_EII VSLOT=1 VLAN=200 LOAD QASP FRAME=ETHERNET_II NAME=QASP_1_V300_EII VSLOT=1 VLAN=300 LOAD ODINEB.NLM LOAD Q57 FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57 FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=4 VLAN=1 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_V100_EII ADDR=172.16.210.1 MASK=255.255.0.0 BIND IP QASP_1_V200_EII ADDR=172.17.210.1 MASK=255.255.0.0 BIND IP QASP_1_V300_EII ADDR=172.18.210.1 MASK=255.255.0.0 NOTE: When BIND QASP1 Q57_1_EII is ex ecuted, the adapter Q57_1_EII is b ound to all three VLANs. If you are unable to log in to th e server after configurin g ASP, add the following command lines before loading ASP: UNLOAD SLPTCP LOAD SLPTCP Jumbo Frames Jumbo Frames are supported in all balance modes. Th e maximum frame size will be automatically set to the smallest maximum frame size of all adapters in the team. Use appropriate keywords to ena ble jumbo frames when loading the adapter drivers. NOTE: This property is not supported on the NC150T and NC1020. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 JUMBO=9000 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 JUMBO=9000 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII Additional command line keywords Additional command line keywords include: ⢠CHECKSUM=ON. This keyword enables QASP .LAN to offload TCP/UDP and IP checksums to the bound adapters if supported by the OS. This improv es performance if some or all adapters in the
Adapter configurations 114 team support hardware checksums. Be sure to load the adapter drivers with hardware checksums enabled. Default is ON. Example: A team of two adapters with hardware checksums enabled. LOAD QASP.LAN FRAME=ETHERNET_II NAME=Q57_1_EII CHECKSUM=ON VSLOT=1 LOAD ODINEB.NLM LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII CHECKSUM=ON SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII CHECKSUM=ON SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII ⢠NOSCREEN. This keyword disables the menu-driven screen when QASP.LAN is loaded for the first time. By default, this screen/menu int erface will be loaded when QASP.LAN is loaded and cannot be turned off without unloading QASP.LAN. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 NOSCREEN ⢠GVRP. This keyword enables GVRP (GARP VLAN Registra tion Protocol) for the VLAN that is loaded. An untagged 802.2 frame type must be loaded in the virtual adapter and all bound physical adapters for GVRP to take effect. This is necessary because GVRP uses untagged 802.2 frames to advertise VLAN memberships. Use VLAN =0 FRAM E=ETHERNET_802.2 in the LOAD command to specify untagged 802.2 frame type. Example: LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 GVRP LOAD QASP.LAN FRAME=ETHERNET_802.2 NAME=QASP_1_E82 VSLOT=1 VLAN=0 Editing the AUTOEXEC.NCF file When an adapter configuration is saved, the NetWar e install program adds load and bind statements to the AUTOEXEC.NCF file. By accessing this file, you can verify the parameters configured for e ach adapter, add or delete parame ters, or modify parameters. A valid AUTOEXEC.NCF file is shown below wi th various VLAN and teaming examples. ⢠Example 1: Team of 2 network adapters with frame type Ethernet_II and one VLAN, number 2. Load QASP.LAN with the frame types and VLAN ID(s) specified. LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 Load the network drivers for the adapters in the team with the same frames types. LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 Bind QASP.LAN to the adapters in the team for each protocol. BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII Bind protocols to QASP.LAN. BIND IP QASP_1_V2_EII ADDR=192.168.2.200 MASK=255.255.255.0 ⢠Example 2: Team of two network adapters with fram e type Ethernet_II and three VLANs, number 2, 3, and 4.
Adapter configurations 115 Load QASP.LAN with the frame types and VLAN ID(s) specified. LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V2_EII VSLOT=1 VLAN=2 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V3_EII VSLOT=1 VLAN=3 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V4_EII VSLOT=1 VLAN=4 Load the network drivers for the network adapters in the team with the same frame types and VLANs specified. LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 Bind QASP.LAN to the network adapters in the team for each protocol. QASP BIND is only used for the first VLAN all other VLANs are automatica lly # bound to the virtual adapter (VSLOT=1). BIND QASP1 Q57_1_V_EII BIND QASP1 Q57_2_V_EII Bind protocols to QASP.LAN. BIND IP QASP_1_V2_EII ADDR=192.168.2.200 MASK=255.255.255.0 BIND IP QASP_1_V3_EII ADDR=192.168.3.200 MASK=255.255.255.0 BIND IP QASP_1_V4_EII ADDR=192.168.4.200 MASK=255.255.255.0 mount all NOTE: If you modify any adapter parameters, you must reboot the system for the changes to take effect. If you make changes and do no t reboot, you may experience configuration problems. ⢠Example 3: Multiple SLB Teams LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_II NAME=QASP_1_EII LOAD Q57.LAN SLOT=2 FRAME=Ethernet_II speed=100fd NAME=Q57_3_EII LOAD Q57.LAN SLOT=3 FRAME=Ethernet_II speed=100fd NAME=Q57_4_EII BIND QASP1 Q57_3_EII BIND QASP1 Q57_4_EII BIND IP QASP_1_EII ADDR=192.222.20.1 MASK=255.255.255.0 LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_II NAME=QASP_2_EII LOAD Q57.LAN SLOT=10016 FRAME=ETHERNET_ii 100hd NAME=Q57_1_Eii LOAD Q57.LAN SLOT=10017 FRAME=Ethernet_ii 100hd NAME=Q57_2_Eii BIND QASP2 Q57_1_EII BIND QASP2 Q57_2_EII BIND IP QASP_2_EII ADDR=192.222.21.1 MASK=255.255.255.O NOTE: When binding Q57_ 1_EII to QASP_ 1_EII, Q57_1_ 82 is also bo und to QASP7_1_ E82. NOTE: The VLSOT range is from 1 to 4 (only four teams can be configured). When binding, add the VSLOT number to the QASP driver. ⢠Example 4: Multiple SLB Teams with Multiple Frame Types
Adapter configurations 116 LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_II NAME=QASP_1_EII LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_802.2 NAME=QASP_1_E82 LOAD Q57.LAN SLOT=1 speed=100fd FRAME=Ethernet_II NAME=Q57_1_EII LOAD Q57.LAN SLOT=2 speed=100fd FRAME=Ethernet_II NAME=Q57_2_EII LOAD Q57.LAN SLOT=1 speed=100fd FRAME=Ethernet_802.2 NAME=Q57_1_E82 LOAD Q57.LAN SLOT=2 speed=100fd FRAME=Ethernet_802.2 NAME=Q57_2_E82 bind qasp1 q57_1_EII bind qasp1 q57_2_EII BIND IP QASP_1_EII ADDR=192.222.20.1 MASK=255.255.255.0 BIND IPX QASP_1_E82 NET=100 LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_II NAME=QASP_2_EII LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_802.2 NAME=QASP_2_E82 LOAD Q57.LAN SLOT=3 FRAME=Ethernet_II NAME=Q57_3_EII LOAD Q57.LAN SLOT=4 FRAME=Ethernet_II NAME=Q57_4_EII LOAD Q57.LAN SLOT=3 FRAME=Ethernet_802.2 NAME=Q57_3_E82 LOAD Q57.LAN SLOT=4 FRAME=Ethernet_802.2 NAME=Q57_4_E82 bind qasp2 q57_3_EII bind qasp2 q57_4_EII BIND IP QASP_2_EII ADDR=192.222.21.1 MASK=255.255.255.0 BIND IPX QASP_2_E82 NET=100 ⢠Example 5: Multiple SLB Teams with Maximum VSLOTS LOAD QASP.LAN VSLOT=1 FRAME=Ethernet_II NAME=QASP_1_EII LOAD Q57.LAN SLOT=1 speed=100fd FRAME=Ethernet_II NAME=Q57_1_EII LOAD Q57.LAN SLOT=2 speed=100fd FRAME=Ethernet_II NAME=Q57_2_EII bind qasp1 q57_1_eii bind qasp1 q57_2_eii BIND IP QASP_1_EII ADDR=192.222.20.1 MASK=255.255.255.0 LOAD QASP.LAN VSLOT=2 FRAME=Ethernet_II NAME=QASP_2_EII LOAD Q57.LAN SLOT=3 speed=100fd FRAME=Ethernet_II NAME=Q57_3_EII LOAD Q57.LAN SLOT=4 speed=100fd FRAME=Ethernet_II NAME=Q57_4_EII bind qasp2 q57_3_eii bind qasp2 q57_4_eii BIND IP QASP_2_EII ADDR=192.222.21.1 MASK=255.255.255.0 LOAD QASP.LAN VSLOT=3 FRAME=Ethernet_II NAME=QASP_3_EII LOAD Q57.LAN SLOT=5 speed=100fd FRAME=Ethernet_II NAME=Q57_5_EII LOAD Q57.LAN SLOT=6 speed=100fd FRAME=Ethernet_II NAME=Q57_6_EII bind qasp3 q57_5_eii bind qasp3 q57_6_eii BIND IP QASP_3_EII ADDR=192.222.22.1 MASK=255.255.255.0
Adapter configurations 117 LOAD QASP.LAN VSLOT=4 FRAME=Ethernet_II NAME=QASP_4_EII LOAD Q57.LAN SLOT=7 speed=100fd FRAME=Ethernet_II NAME=Q57_7_EII LOAD Q57.LAN SLOT=8 speed=100fd FRAME=Ethernet_II NAME=Q57_8_EII bind qasp4 q57_7_eii bind qasp4 q57_8_eii BIND IP QASP_4_EII ADDR=192.222.23.1 MASK=255.255.255.0 ⢠Example 6: VLAN with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V100_EII VSLOT=1 VLAN=100 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V200_EII VSLOT=1 VLAN=200 LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_V300_EII VSLOT=1 VLAN=300 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 VLAN=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 VLAN=1 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_V100_EII ADDR=172.16.210.1 MASK=255.255.0.0 BIND IP QASP_1_V200_EII ADDR=172.17.220.1 MASK=255.255.0.0 BIND IP QASP_1_V300_EII ADDR=172.18.230.1 MASK=255.255.0.0 NOTE: When binding Q57_ 1_V100_EII to Q57_1_EII , Q57_1_EII is al so bound to the other VLANs on the same VSLOT. VLAN refers to VLAN ID, and a vali d VLAN ID ranges from 1 to 4094. ⢠Example 7: Trunking with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 MODE=TRUNK LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_EII ADDR=172.16.210.1 MASK=255.255.0.0 NOTE: Switch ports must be configured for Trun king (for example, FEC or GEC for Cisco switches). ⢠Example 8: GVRP with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 VLAN=100 GVRP LOAD QASP.LAN FRAME=ETHERNET_802.2 NAME=QASP_1_E82 VSLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_1_E82 SLOT=1 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_2_EII SLOT=2 LOAD Q57.LAN FRAME=ETHERNET_802.2 NAME=Q57_2_E82 SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_EII ADDR=172.16.210.1 MASK=255.255.0.0 NOTE: In this example, VLAN 100 will be advertis ed to the switch using GVRP. Notice that untagged 802.2 frames must be loaded to allow send ing and receiving GVRP frames. Switch ports must be configured with GVRP enabled. ⢠Example 9: 802.3ad with QASP LOAD ODINEB.NLM LOAD QASP.LAN FRAME=ETHERNET_II NAME=QASP_1_EII VSLOT=1 MODE=802.3ad LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=1
Adapter configurations 118 LOAD Q57.LAN FRAME=ETHERNET_II NAME=Q57_1_EII SLOT=2 BIND QASP1 Q57_1_EII BIND QASP1 Q57_2_EII BIND IP QASP_1_EII ADDR=172.16.210.1 MASK=255.255.0.0 NOTE: 802.3ad mode requires the Ethernet sw itch to support 802.3ad with LACP (Link Aggregation Control Protocol). Configuration parameters for Q57.LAN driver Configuration parameters for Q57.L AN driver include: ⢠TxDescriptors= o This keyword is to initialize Descripto r resources on the adapter for transmits. o Recommended Min = 100 o Max = 512 o Default = 200 NOTE: Min = 1 when used with DOS Client32 and when Keyw ord P3=1 or 2. ⢠RxBuffers= o This keyword is to pre-allocate receive ECBs & Receive adapter re sources. This setting may be affected by the NetWare server maximum/ minimum packet receive buffer settings. o Recommended Min = 32 o Max = 512 o Default = 200 NOTE: Min = 1 when used with DOS Client32 and when Keyw ord P3=1 or 2. ⢠Speed= o This keyword is to force the line speed of the adapter. When this keyword is used, the switch to which the adapter cable is connected must be forc ed. If not, the adapter and switch may run at different duplex modes. It is important to have the switch and adapter set properly. The best way to ensure good connectivity is to not set the speed keyword and allow fo r automatic link setup (auto-negotiation). o Choices are: AUTO, 10FD,10HD, 100FD, 100HD o Default is Auto. (Note that 1000 FD is only auto-detec ted at this time.) ⢠Link= o This keyword is only used to allow the adapter to negotiate a specific or forced line speed with a switch that is not forced, but instead setup for auto-negotiation. It is best to allow for auto- negotiation of the card and switch by not setting this keyword or the speed keyword. Only use this keyword if the speed keyword is set to something other than AUTO. o Choices are: AUTO, FORCE o Default is FORCE because the speed keyword is usually used when a switch and the adapter speeds are both forced to a specific value. ⢠RxTicks=
Adapter configurations 119 o This keyword is to enable the use of batching receives within a speci fic time period. o Min = 0, disabled o Max = 5000000, 5 seconds o Units are in micro seconds o Default = 360 ⢠TxPacketsPer= o This keyword is to enable the use of batching transmits to a specific amount of packets. o Min = 0, disabled o Max = 100 o Default = 64 ⢠CheckSum= o This keyword is to enable or disable the transmit & receive checksum off loading feature. The checksum off loading support is only for TCP/IP packets. For that reason, the default OFF. o Choices are: ON, OFF, TX, RX o Default is ON ⢠TxFlow= o This keyword allows enabling/di sabling of TxFlow control. o Choices are: ON, OFF o Default is O FF ⢠RxFlow= o This keyword allows enabling/di sabling of RxFlow control. o Choices are: ON, OFF o Default value is OFF ⢠VLAN= o This keyword allows enabling/dis abling of VLAN tagging support. This keyword is only used when the QASP program is loaded for VLAN support on the Q57 driver. Do not set the VLAN keyword to 1 if the QASP program is not being used. o Choices are: 0 (Disabled), 1 (Enabled) o Default value is 0 ⢠Jumbo= o This keyword enables Jumbo frame support. When enabled, jumbo packets of up to 9000 bytes are supported. o Jumbo=1514-9000 o Default is jumbo packets not supported. ⢠PDriver= o This keyword allows for the driver to operate in persistent driver mode. Only use it if the adapter is placed in a Hot Plug PCI slot and only if required to swap with an exact same board. o Choices are: ON, OFF o Default value is OFF
Adapter configurations 120 NOTE: If you modify any adapter parameters, you must reboot the system for the changes to take effect. If you make changes and do no t reboot, you may experience configuration problems. NetWare Client 32 driver installation for HP adapters The NC10032.LAN driver is a high-p erformance server driver that meets the Op en Data Link Interface (ODI) v1.11 C specification. It can be used on Cl ient 32 workstat ions with NetWare v3.X servers and greater. Installing Client 32 on DOS 1. Install the driver and its support files for DOS. 2. From the root directory, run MAKEDISK .BAT to create driver diskettes. 3. Locate the Client 32 Software. NOTE: For NetWare 5, the software is located on the Novell Client Software CD. 4. For DOS installation, run INSTALL.EXE. NOTE: INSTALL.EXE is located in the follow ing CD directory: \PRODUCTS\DOSWIN32. 5. Follow the Novell prompts. 6. For TCP/IP, select TCP/IP protocol stack at the Options/Product screen. 7. At the 32-bit LAN Driver screen, scroll to the end of li sted drivers and select User Specified 32-bit driver. 8. Insert the HP driver diskette. 9. Select HP Ethernet or Fast Ethernet NIC. 10. Select needed setting and press the F10 key to continue. 11. Follow the remaining prompts. NET.CFG parameters This section describes NET.CFG parameters for the HP driver. Each statement must appear on a separate line under LINK DRIVER N100C32 and each must be indented with at least a single space or tab character. Example NET.CFG Statements DRIVER N100C32 SLOT 16 FRAME ETHERNET_802.2 NODE ADDRESS 220023456789 ⢠FORCEDUPLEX. If no duplex setting is specified, the a dapter auto-negotiates the duplex with the hub/switch. If the adapter and hub/switch are unable to auto-negotiate, HP adap ters default to half- duplex. Full-duplex allows for simultaneous transm it and receive data transfer. Full-dupl ex can be used only when the adapter is connected to a hub th at supports and is enabled both for full-duplex mode and for the specified speed.
Adapter configurations 121 o Syntax: FORCEDUPLEX=n Where n= 0âauto-negotiate 1âhalf-duplex 2âfull-duplex NOTE: To set the adapter to half- or full-duplex, yo u must set the SPEED parameter to either 10 or 100. o Default: auto-negotiation o Examples: 100 Mbps full-duplex: FORCEDUPLEX=2 SPEED=100 10 Mbps full-duplex: FORC EDUPLEX=2 SPEED=10 ⢠SPEED. If no speed setting is specified, the adapter auto-negotiates the speed with the hub/swit ch, either 10 or 100 Mbps. If the adapter and hub/sw itch are unable to auto-negotiate, the adapter defaults to 10 Mbps. o Syntax: SPEED=n (n = 10 or 100) o Default: The adapter automatically senses speed. Examples of system file settings ⢠Example NET.CFG Link Driver N100C32 NetWare DOS Requester FIRST NETWORK DRIVE F NETWARE PROTOCOL NDS BIND Protocol IPX IPX SOCKETS 40 Protocol TCPIP IP_ADDRESS 201.201.202.25 IP_ROUTER 201.201.202.1 IP_NETMASK 255.255.255.0 PATH TCP_CFG C:\NW5C32\TCP BIND N100C32 ⢠Example STARTNET.BAT for Client 32 SET NWLANGUAGE=ENGLISH C:\NW5C32\NIOS.EXE LOAD C:\NW5C32\NBI32.NLM LOAD C:\NW532\LSLC32.NLM LOAD C:\NW5C32\CMSM.NLM LOAD C:\NW5C32\ETHERTSM.NLM LOAD C:\NW5C32\N100C32.LAN FRAME=ETHERNET_II LOAD C:\NW5C322\TCPIP.NLM LOAD C:\NW5C322\TRANNTA.NLM LOAD C:\NW5C322\IPX.NLM LOAD C:\NW5C322\SPX_SKTS.NLM LOAD C:\NW5C322\CLIENT32.NLM ⢠Example CONFIG.SYS DEVICE=C:\DOS\HIMEM.SYS device=c:\dos\emm386.exe noems DOS=HIGH,umb
Adapter configurations 122 FILES=100 buffers=25 LASTDRIVE=m fcbs=4,4 device=c:\6150\amsida.sys device=c:\6150\amsicd.sys /d:cpqcd001 ⢠Example AUTOEXEC.BAT PROMPT $p$g @ECHO OFF C:\dos\mscdex.exe /d:cpqcd001 @CALL C:\nw5c32\STARTNET Solaris To customize the driver edit "/kernel/drv/bcme.conf" and update the respective p arameters in this file. The parameters include the following. ⢠ForceSpeedDuplex. Configures link (or instance) to a ce rtain Speed and Duplex. By default, AutoNegotiate (0) is set. The set up is based on the following values: 0âAutoNegotiate 1â10 Mbps speed and Half-Duplex mode 2â10 Mbps speed and Full-Duplex mode 3â100 Mbps speed and Half-Duplex mode 4â100 Mbps speed and Full-Duplex mode 5âForce 1000 Mbps Full-Duplex mod e (Fiber adapter only) 6âAutoNegotiate only 1000 Mbps F ull-Duplex mode 7âAutoNegotiate only 1000 Mbps Half-Duplex mod e 8âAutoNegotiate only 100 Mbps Ful l-Duplex mode 9âAutoNegotiate only 100 Mbps Half-D uplex mode 10âAutoNegotiate only 10 Mbps Full-Duplex mode 11âAutoNegotiate only 10 Mbps Half-Duplex mod e Example: Configure adapters of instance#0 and in stance#3 to 100 Mbps Full-Duplex, and 10 Mbps Half-Duplex. ForceSpeedDuplex=2,0,0,1,0, 0,0,0,0,0,0,0,0,0,0,0; ForceSpeedDuplex=0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0; ⢠FlowControl. Configures the flow control pa rameters of a link. The setu p is based on the following values: 0âBoth Tx and Rx flow control are disabled. 1âTx flow control is enabled. Pause frames will be sent if resource is low, but the device will not process Rx Pause Frame. 2âRx flow control is enabled. If the device receives Pause Frame, it will stop send ing, but the device will not send Pause Frame if resource is low. 3âBoth Rx and TX flow control are enabled. Pause fr ames will be sent if resourc e is low. If the device receives Pause Frame, it will stop sending.
Adapter configurations 123 4âAdvertise both Rx and TX flow control being en able and negotiate with link partner. If link AutoNegotiate is not enabled, then both Tx & Rx Flow Contro l are disabled. Example: FlowControl=0,0,0, 0,0,0,0,0,0,0, 0,0,0,0,0,0; ⢠MaxJumboFrameSize. Configure s the Jumbo Frame feature of a link. The valid range for this parameter is 0 to 9000. If the value configured is less than 1500, then the Jumbo Frame feat ure is disabled. BCM5705 adapters do not support jumbo fr ame and this parameter will be ignored by the driver Example: MaxJumboFrameSize=0,0, 0,0,0,0,0,0,0,0, 0,0,0,0,0,0; ⢠TxPacketDescCnt. Configures the number of Tx packet de scriptors. The valid range is 32 to 512. More system memory reso urces are used for larger number of Tx Pack et Descriptors. The default value is 200. Example: TxPacketDescCnt=200; ⢠RxStdDescCnt. Configures the number of Rx packet descriptors. The valid range for this parameter is 32 to 512. More system memory resourc es are used for larger number of Rx Packet Descriptors. The default value is 500. Example: RxStdDescCnt=500; ⢠RxJumboDescCnt. Configures the number of Rx Jumbo packe t descriptors. The valid value is 32 to 256. More system memory resourc es are used for la rger number of Rx Jumbo packet descriptors. This parameter is only used if the jumbo frame feature is enabled. The default value is 50. Example: RxJumboDescCnt=50; ⢠RxCoalescingTicks. Configures the number of Rx Host Co alescing Ticks in microseconds. This determines the upper-boun dary of the time interval th at the device will generate interrupts if one or more frames are received. The default va lue is 150. Example: RxCoalescingTicks=150; ⢠RxMaxCoalescedFrames. Configures the number of Rx Maximum Coalesced Fram es parameters. This determines the upper-boundary of maximum number of Rx buffer descriptors that the device processes before it will generate an interrupt. The default value is 10. Example: RxMaxCoalescedFrames=10; ⢠TxCoalescingTicks. Configures the number of Tx Host Co alescing Ticks in microseconds. This determines the upper-boun dary of the time interval th at the device will generate interrupts if one or more frames are sent. The default value is 500. Example: TxCoalescingTicks=500; ⢠TxMaxCoalescedFrames. Configures the number of Tx Maximu m Coalesced Frames parameters. This determines the upper-boundary of maximum number of Tx buffer descriptors that the device processes before it will generate an interrupt. The default value is 80. Example: TxMaxCoalescedFrames=80; ⢠RxCoalescingTicksDuringInt. Configures the number o f Rx Host Coalescing Ticks in microseconds during interrupt. This determines the upper-boundary of the time interval that the device will gen erate interrupts if one or more fr ames are received du ring interrupt handling. The default value is 75. Example: RxCoalesci ngTicksDuringInt=75; ⢠TxCoalescingTicksDuringInt. Configures the number of Tx Host Coalescing Ticks in microseconds during interrupt. This determines the upper-boundary of the time interval that the device will gen erate interrupts if one or more fr ames are received du ring interrupt handling. The default value is 75.
Adapter configurations 124 Example: TxCoalesci ngTicksDuringInt=75; ⢠RxMaxCoalescedFramesDuringInt. Configur es the number of Rx Maximum Coale sced Frames parameters during interrupt handling. This determin es the upper-boundary of maximum number of Rx buffer descriptors that the device processes before it will generate an interrupt during interrupt handling. The default value is 10. Example: RxMaxCoalescedFramesDuringInt=10; ⢠TxMaxCoalescedFramesDuringInt. Configures the n umber of Tx Maximum Coalesced Frames parameters during interrupt handling. This determin es the upper-boundary of maximum number of Tx buffer descriptors that device proc esses before it will generate an interrupt during interrupt handling. The default value is 10. Example: TxMaxCoalescedFramesDuringInt=10; ⢠StatsCoalescingTicks. Configures how often adapter statis tics are DMAed to host memory in microseconds. Default is 1000000. Example: StatsCoalescingTicks=1000000; ⢠DoubleCopyTxBufferSize. Configures double copy Tx buffer size. If a packet to be transmitted is less than this parameter and spanned more than one fragment, these fragments of this packet will be combined into one fragment. Example: DoubleCopyTxBufferSize=64; Customize driver configuration via ndd command Driver configurations can also be temporarily changed with the Solaris ndd command. Any changes made with the ndd command are temporary and will be lost when you reboot the system. To make configuration changes survive after reboo t, modify bcme.conf instead. ⢠To display parameters that are configurable via ndd: ndd /dev/bcme '?' The system should ret urn the following: ? (read only) Instance (read and write) ForceSpeedDuplex (read and write) FlowControl (read and write) TxPacketDescCnt (read and write) RxStdDescCnt (read and write) RxCoalescingTicks (read and write) RxMaxCoalescedFrames (read and write) TxCoalescingTicks (read and write) TxMaxCoalescedFrames (read and write) RxCoalescingTicksDuringInt (read and write) RxMaxCoalescedFramesDuringInt (read and write) TxCoalescingTicksDuringInt (read and write) TxMaxCoalescedFramesDuringInt (read and write) StatsCoalescingTicks (read and write) DoubleCopyTxBufferSize (read and write) DoubleCopyTxBufferSize (read and write) BlinkLeds (write only) ⢠To configure a particular adapter, the Instance paramete r must be set to the instance associated with a particular adapter. For example, to force adapte r of instance 1 to 100 Mbps Full-Duplex: ndd -set /dev/bcme Instance 1 ndd -set /dev/bcme ForceSpeedDuplex 3
Adapter configurations 125 ⢠To query current configuration of Flow Control of instance 3: ndd -set /dev/bcme Instance 3 ndd -get /dev/bcme FlowControl ⢠To blink all LEDs for 10 seconds of adapter of instance 5: ndd -set /dev/bcme Instance 5 ndd -set /dev/bcme BlinkLeds 10
Team properties 126 T eam pr oper tie s In this section Adapter te aming ................................................................................................................ .................. 126 Multifunction gigabit se rver adapter teaming ................................................................................... ........ 129 Windows 2000 and Wind ows NT4 te aming .......................................................................................... 1 30 NetWare te aming ................................................................................................................ ................ 139 Adapter teaming ⢠Adapters must all be on the same Laye r 2 (Broadcast) network to be teamed. ⢠The maximum configuration is eight teams with ei ght adapters in ea ch team for a maximum of 64 ports. ⢠For information about teaming Multifunction Gigabit Server adapters, refer to Multifunction gigabit server adapter teaming (on page 129 ) in this guide. Teaming types for HP Fast Ethern et and Gigabit Server Adapters NOTE: Dual Channel Network Fault Tolerance te aming capabilities are not supported on systems running Windows® 2000. ⢠Automatic (Recommended). Default setting. The system automati cally selects the best team type o r combination of team features based on the config uration of the adapters in the team. 802.3ad will be used to place team members into port trunks/channels if possible. All transmit packets are load balanced among all team members based on a Lo ad Balancing algorithm in the teaming device driver. The receive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are lo ad balanced among the remaining adapters. Currently, Automatic is supported in Window s® 2000, Windows Server 2 003, and Windows Server 2003 x64. ⢠802.3ad Dynamic Dual Channel Load Balancing (INP). ProLiant Essentials Intelligent Ne tworking Pack (INP) only. There must be a minimum of two adapters in a dual channel team. All team members within each group are dynamically placed into a port- trunk/channel by dynamic Link Aggregation Contro l Protocol (LACP) agreement with the switch. A failure of a team member to synchronize joining a po rt-trunk or channel with the switch is treated as an error and that team member is considered fail ed. All transmit packets are load balanced among all team members based on a Load Balancing algo rithm in the teaming device driver. The receive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are load balanced among the remaining adapters. ⢠802.3ad Dynamic Dual Channel Network Fault Tolerance (INP). Available in the ProLiant Essentials INP only. Th ere must be a minimum of two adapters in a dual channel team and a minimum of one adapter per gr oup. All team members within each group are dynamically placed into a port-trunk/channel by dy namic LACP agreement with the switch. A failure of a team member to synchronize joining a port-tru nk or channel with the switch is treated as an error and that team member is considered fail ed. Dynamic Dual Channel Network Fault Tolerance
Team properties 127 (NFT) provides the safety of additional backup li nks between the server and hub/switch. Dynamic Dual Channel NFT is implemented with one channe l team providing a primary adapter and an other channel team providing a secondary, backup adapte r(s). During normal operations, if the adapters in the primary team fail, a link to one of the ad apters in the secondary, backup team automatically takes over. ⢠802.3ad Dynamic Dual Channel NFT and Preference Order (INP). Available in the ProLiant Essentials INP only. Pr ovides the same optio ns as 802.3ad Dynamic Dual Channel NFT except it offers the additional option of setting the priority order of the adapters that are used in the primary and secondary channels. ⢠Switch-assisted Dual Channel Load Balancing (INP). Available in the ProLiant Essentials INP only. There must be a minimum of two adapters in a dual channel team and a minimum of one adapter per group. One group of team members is treated as a Switch-assisted Load Balancing (SLB) group to one switch, and a second group of team members is treated as an SLB gr oup to a second switch. All transmit packets are load balanced among all team members ba sed on a Load Balancing algorithm in the teaming device driver. The rece ive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are load balanced among the remaining adapters. Only Gigabit adap ters can be used for Switch-assisted Dual Channel Load Balancing. When you select Switch-a ssisted Dual Channel Load Balancing, the Team Members window bec omes Team Members Grouping and displays two windows l abeled Group 0 and Group 1. The up and down arrows are used to assign the members to be in Group 0 and Group 1. If a team contains an iSCSI-enabled adapter, you cannot use the Switch-assisted Dual Channel Load Balancing (INP) setting. ⢠Switch-assisted Dual Channel Network Fault Tolerance (INP). Available in the ProLiant Essentials I NP only. There must be a minimum of two adapters in a dual channel team and a minimum of one adapter per group. Switch-assisted dual channel Network Fault Tolerance (NFT ) provides the safety of additional backup links betwee n the server and hub/switch. It is implemented with one channel team providing the primary adapter and a second channel team providing the secondary, backup adapter(s). During normal operations , if the adapters in the primary team fail, a link to an adapter on the Secondary team automatically takes over . Team members can be moved to Group 0 or Group 1 on the Team Properties Teaming Controls tab. ⢠Switch-assisted Dual Channel NFT and Preference Order (INP) . Available in the Pr oLiant Essentials INP only. Provides the same options as Switch -assisted Dual Channel Network Fault Tolerance except it offers the additional opti on to select the priority order of the group (Group 0 or Group 1) used for the primary and secondary channels. ⢠802.3ad Dynamic with Fault Tolerance. All team mem bers are dynamically placed into a port- trunk/channel by dynamic protocol agreement (LACP) with the switch. A failure of a team member to synchronize joining a port-trunk or channel with th e switch is treated as an error and that team member is considered failed. All transmit pa ckets are load balanced among all team members based on a Load Balancing algorithm in the teamin g device driver. The receive packets are load balanced among all team members by the switch. If a failure of any team member occurs, the packets are load balanced among the remaining adap ters. If a team contai ns an iSCSI-enabled adapter, you cannot use the 802.3ad Dyna mic with Fault Tolerance setting. ⢠Switch-assisted Load Balancing with Fault Tolerance (SLB). SLB increases a server's throughput using Fast Ethernet or Gigabit Ethernet ports. SLB lo ad balances all traffic. Unlike TLB, SLB can be configured to increase both transmitting and rece iving channels between your serv er and switch. For example, an SLB team containing four HP Fast Ethernet adapters configured for full-duplex operation would provide an aggregate maximum transmit rate of 400 Mbps a nd an aggregate maximum receive rate of 400 Mbps, resulting in a total bandwidth of 800 Mb ps.
Team properties 128 SLB works only with SLB-en abled switches, such as the Catalyst 600 0 series. The SLB software continuously analyzes loading on each adapter an d balances network traffic across the adapters as needed. Adapter teams configured f or SLB also provide the benefits of NFT. To use SLB, you must have at least two HP Fast Et hernet or Gigabit adapters installed in your server and linked to the same SLB-enabled switch. For SLB, you can only team 100 Mbps adapters together (no 10 Mbps or Gigabit), with the exception of the NC71xx and NC77xx server adapters, which are Gigabit and can run at 10/100 Mbps. ⢠Transmit Load Balancing with Fault Tolerance (TLB). TLB creates a team of adapters to increase a server's throughput. If you configure TLB, NFT is automatically configured. TLB wor ks with any industry-standard 10BASE-T, 100BASE-TX, or Gigabi t switch. All transmit IP frames are load balanced among all team members based on a Load Balancing algorithm in the teaming device driver. The Current Primary adapter transmits all non- IP frames, and receives all frames for the team. If a failover event occurs, one of the non-Primar y adapters assumes the role of Current Primary adapter, and transmit IP packets are load balanced among all remaining team members. If a failure occurs in any of the non-Primary adapters, the pa ckets are load balanced among all remaining team members. To use TLB, you must have at least two NCxxxx a dapters (ports) capable of the same speed and linked to the same Layer 2 (Broadcast) sw itch insta lled in your server. They cannot be connected to the same switch and/or be members of different VLANs. With TLB, as you add adapters to your server, you can group them in teams to provide a si ngle virtual adapter with increased transmission bandwidth. For example, a TLB team containing four HP Fast Ethernet adapters configured for full- duplex operation would provide an aggregate maximum transmit rate of 400 Mb ps and a 100 Mbps receive rate resulting in a total bandwidth of 500 Mbps. TLB load balances IP traffic under Window s NT® and IP and IPX traffic under NetWare. ⢠Transmit Load Balancing with Fault Tolerance and Preference Order. All transmitted IP packets are load balanced among all team members based on a Load Balancing algorithm in the teaming device driver. The Current Primary adapter transmit s all non-IP packets, and receives all frames for the team. If a failover event occurs, one of the non-Primary adapters assumes the role of Current Primary adapter, and transmitted IP packets are load balanced among all remaining team members. If a failure occurs in any of the non-Primary a dapters, the packets are load balanced among all remaining team members. The user s ets the priority of the adapters relative to each other within the team. The priority of the adapters is an additi onal consideration when determining the Primary adapter. ⢠Network Fault Tole rance Only (NFT). NFT provides the safety of an additional backup link between the server and hub/switch. NFT is implemented wi th a Primary adapter and a Secondary (backup) adapter(s). During normal operations, if the Primary adapter fails, the link to the Secondary adapter automatically takes over. o NFT can occur between two TOE-enabled adapters ⢠Network Fault Tolerance only with Preference Order. NFT provides the safety of an additional backup link between the server and hub/switch. NF T is implemented with a Primary adapter and a Secondary (backup) adapter(s). During normal operatio ns, if the Primary adapter fails, the link to the Secondary adapter automatically takes over. The user sets the priority of the adapters relative to each other within the team. The priority of the adapters is an additional consideration when determining the Primary adapter. Currently, NFT is supported in Windows® 200 0, Windows Server 2003, and Windows Server 2003 x64, Windows NT® 4.0, NetWare 6.x, and UnixWare 7.x. Transmit Balancing Method
Team properties 129 Select or change the Transmit Balancing Method for th e team. If the team type is Network Fault Tolerance Only with Preference Order or Network Fault Toleranc e Only (NFT), the Transmit Balancing Algorithm list disabled and Automatic is selected. o Automatic . (Recommended) Default settin g. Teaming driver selects the load balancing mechanism based on the packet type. o TCP Connection . Load balances transmit TCP packets using the TCP connection. o Destination IP Address . Load balances transmit IP packets using the last four bits of the destination IP Address. o Destination MAC Address . Load balances transmit IP packets using the last four bits of the destination MAC Address. o Round Robin (Packet order not guaranteed) . Load balances transmit IP packets among all team members. A packet is sent on one team member, the next packet is sent out on the next team member, and so on. When the last team member is utilized, the rota tion begins again. UnixWare 7.x and OpenServer 6.x teami ng notes UnixWare 7.x and OpenServer 6.x operating systems support Network Fault Tolerance (NFT) but not Transmit Load Balancing (TLB) or Switch-A ssisted Load Balancing (SLB). Multifunction gigabit server adapter teaming If a TOE- or iSCSI-enabled adapter is included as one of the following team types, the team type is changed to TransmitLoadBalancing. ⢠Automatic ⢠802.3ad Dynamic Dual Channel Load Balancing (INP) ⢠Switch-assisted Dual Channel Load Balancing (INP) ⢠802.2ad Dynamic with Fault Tolerance ⢠Switch-assisted Load Balancing with Fault Tolerance (SLB) The following are limitations for a TOE-en abled team: ⢠TOE teaming performance is reduced if a switch or network device is configured with the round robin aggregation load balancing method. Configure the aggregation load balancing method on the switch or network device to somethin g else ot her than round robin and then re-start the team (disable the team and enable the team for all VLANs). ⢠The 802.3ad Dynamic Dual Channel Load Balancing (INP) and Switch-assisted Dual Channel Load Balancing (INP) team types are not supported fo r TOE-enabled adapters. Select a TOE-supported team type and apply the change, if necessary, before enabling TOE. It is necessary to apply the new team type, if you are unable to enable TOE. When creating a team with all TOE-capable adapters: ⢠If at least one adapter is TOE enabled, the team is created with TOE enabled and all adapters on the team are set to TOE enabled. ⢠If none of the adapters are TOE enabled, th e team is created with TOE disabled. ⢠If TOE is disabled on a team of TOE-capable adapters, enabling TO E for the team changes all of the adapters to TOE enabled. When creating a team with TOE-capable and non-TOE capable adapters:
Team properties 130 ⢠If the new team includes at least one TOE-en abled adapter and at least one non-T OE capable adapter, the team is created with TOE disabled and TO E is disabled on all of the individual teamed adapters. ⢠If a non-TOE capable adapter is added to a TO E-enabled team, the team is changed to TOE disabled. TOE functionality for ea ch teamed adapter is disabled. Disabling or dissolving a TOE-enabled team: ⢠If TOE teaming is disabled on an existing TOE-en abled team, TOE is disabled on all of the tea med adapters. ⢠If a team is dissolved, the TOE functionality for ea ch adapter that was on the team does not change. RSS teaming characteristics When creating a team with all RSS-capable adapters: ⢠If at least one adapter is RSS-enabled, the team is created with RSS-enabled and all adapters on the team are set to RSS-enabled. ⢠If none of the adapters are RSS-enabled, th e team is created with RSS-disabled. ⢠If RSS is disabled on a team of RSS-capable adapters, enabling RSS for the team changes all of the adapters to RSS-enabled. When creating a team with RSS-ca pable and non-RSS capable adapters: ⢠If the new team includes at least one RSS- en abled adapter and at least one non-RSS capable adapter, the team is created with R SS-disabled and RSS functionality fo r each adapter that is in the team is disabled. ⢠If a non-RSS capable adapter is added to an R SS-enabled team, the team is changed to RSS- disabled and RSS functionality for each a dapter that is in the team is disabled. Disabling or dissolving an RSS-enabled team: ⢠If RSS teaming is disabled on an existing RSS-enabled team, the RSS functionality for each adapter that is in the team is disabled. ⢠If a team is dissolved, the RSS fu nctionality for each adapter that was in the team does not change. Windows 2000 and Windows NT4 teaming Windows 2000 teaming NOTE: The installation of adapter drivers from ot her vendors is not supported by HP. To avoid unpredictable results when using the HP Teaming function or other features supported by the HP Network Configuration Utility, install only HP adapter driver s. Upgrading teams to Windows® 2000 The Windows® 2000 Primer Utility from HP lets you re move unsupported Windo ws® 2000 utilities before or after upgrad ing your OS. These uns upporte d utilities w ere designed to run on Windows NT® 4.0 and may have been installe d using the following media: ⢠Any version of the HP SSD for Windows NT® Setup program ⢠Any version of HP Integration Maintenance Utility and HP System Partition Upgrad e Utility installed from HP SmartStart
Team properties 131 ⢠Any HP Software and Documentation CD Version 5.0 or earlier shipped with an HP 10/100 or Gigabit adapter The HP Primer Utility is supported on the following oper ating systems: ⢠Microsoft® Windows NT® 4.0 ⢠Microsoft® Windows® 2000 For PRIMER.EXE, refer to the HP Softpaq #10652 at the HP support website ( http://h18000.www1.hp.com/s upport/files/index.html ) Before running HP Primer Utility Before running the HP Primer Utility, you should be aware of the following: ⢠Retaining Teaming Config uration Settings in Windows NT® 4.0 HP recommends that you record all Windows NT® 4. 0 teaming configuration s ettings. After you run Primer and upgrade the OS to Windows® 2000, yo u can install the new HP Network Configuration Utility for Windows® 2000 and reconfig ure th e adapter teams using the saved settings. ⢠Windows NT® 4.0 components removed by Primer The Primer utility removes the install ed files, th e program links, and the registry entries for the previously installed components. In addition, general cleanup is pe rformed to remove program l inks and common files that may not belong to any specific component. ⢠Adapter configur ation components remove d by Primer o HP Advanced Network Control Utility o HP Network Configuration Utility o HP TLAN Network Teaming and Configuration Utility When Primer removes these network uti lities, it dissolves all Windows NT® 4.0 teams. It also removes the teaming drivers that support the utility. As a result, the adapters that were managed by this utility are "unteamed" and function as single, stand-alone adapters. ⢠Other components also removed by Primer o HP Array Configuration Utility o HP Cluster Verification Utility o HP Insight Management Foundation Agents o HP Insight Management Server Agents o HP Insight Management Storage Agents o HP Integrated Management Display Utility o HP Integrated Management Log Viewer Utility o HP Integration Maintenance Utility o HP NT Management o HP Online Storage Controller Recovery Utility o HP PCI Hot Plug Utility o HP Power Down Manager o HP Power Supply Viewer o HP Remote Monitor Service o HP System Management Services
Team properties 132 o HP System Partition Upgrade Utility Using HP Primer Utility from the graphical user interface To remove the Windows NT® components either befo r e or after upgrading to Windows® 2000, follow these steps: 1. From the location where you have downloaded Primer, double-click PRIMER.EXE. 2. Click CLEANUP. The utility will now remove the previously-in stalled Windows NT® components. 3. Exit the Windows® 2000 Primer Util ity from HP. 4. Reboot your system for the changes to take effect. Using HP Primer Utility from the command line You may run this utility in silent mode to perform th e cleanup without us ing the graphical user interface by following these step s: 1. Open the MS-DOS command prompt. 2. Go to the directory where the Windows® 200 0 Prim er Utility from HP, PRIMER.EXE, is located. 3. Type primer s and press Enter. The utility will run in silent mode. Viewing the HP Primer Utility log results After clicking the Cleanup button or performing a command-line in stall using "primer s," a log file is generated that allows you to view the uninstalled comp onents and any non-critical errors received during cleanup, such as components that the utility was unable to uninstall. You can easily view the results by clicking the View Log button or you can go to the directory where the log file resides, which is usually th e current working directory (where the utility is run). However, if the media from which you are running the utility is writ e-protected, the log file will reside in the %systemdir% directory. NOTE: You must have Administrator privileg es to use the Log. Teaming considerations when upgrading to Windows® 2000 Under Windows NT® 4.0, adapter teaming ma y be managed using CPQNTAC (HP Network Configuration Utility) for HP NCxxxx Ethernet , Fast Ethernet, and Gigabit adapters Because different utilities m anage teaming under Wi ndows NT® 4.0, in upgrading to Windows® 2000, certain steps are necessary to maintain teaming for redundant adapter pairs. Removing previous forms of teaming on Windows NT® 4.0 Both CPQNTAC.SYS and CPQTTAC.SYS are NDIS4 Inte rmediate drivers that do not work with Windows® 2000. For a Windows® 200 0 upgrade, yo u must remove them with one of the following methods: Method 1: Use HP Primer Utility for Windows® 2000 1. Run the HP Primer Utility. Primer remov es the installed files, program links, registry e ntries, and generally cleans files that may not belong to an y specific component. For further information about Primer, refer to HP Softpaq #10652 provided at HP support website ( http://h18000.www1.hp.com/s upport/files/index.html ). 2. Upgrade the OS to Windows® 2000. 3. Install the new HP Network Configuration Utility (NCU) for Windo ws® 2000.
Team properties 133 Method 2: Manual Remova l 1. Dissolve all teams and controller-pairs wit h their respective utilities. 2. Upgrade the OS to Windows® 2000. 3. Install the new HP Network Configuration Utility (NCU) for Windo ws® 2000. Installing the teaming driver NOTE: Before installing the Teaming protocol (T eaming driver), you may want to back up the Windows® 2000 Registry. To install the teaming driver: 1. Go to the HP website ( http://www.hp.com ). 2. Click Software & Driver Downloads from the left menu bar. 3. Type the product name in the For product box and press Enter . For example, type NC370T. 4. Select an operating system. 5. Click HP ProLiant Networking Software . 6. Click download and save the HP SoftPaq (sp#####.exe) file to a directory on your hard drive. The SoftPaq file is a self-extracting executable wi th a file name based on the SoftPaq number. 7. Click the SoftPaq file to extract the files and then open the cmponent.htm file. 8. Select the win2000, win2003 or win2003x64 folder that is appropriate for your operating system. 9. Run the HP ProLiant Network Configuration Ut ility Component Package (cpxxxxxx.exe file). The component is a self-extracting executable with a file name based on the componen t number. 10. Click the Install button to proceed with the installation. 11. After the drivers have been installed, you may delete the HP SoftPaq file. If this is a first-time installation, proceed to step 1 in the "Configuring (or Adding) a Team" section. Additional documentation can be found in Soft Paq SP24889 available at the HP support web site ( http://h18000.www1.hp.com/s upport/files/index.html ). Configuring (or adding) a team NOTE: You can configure a maximum of eight teams with eight adapters in each team for a maximum of 64 ports. NOTE: If you are adding adapters to an existing team, you do not need to install the Teaming driver. To configure a team or add an adapter to an existing team: 1. Launch the HP Network Configuration Utility by doing one of the following: o Double-click the HP Network tray icon. o Double-click the HP Network icon in the Control Panel window. 2. Select the adapters to be teamed in the NCU Main window. 3. Click Team . 4. Configure the team as needed and click OK . 5. Click Yes to confirm the configuration changes.
Team properties 134 6. A message box appears with a warning that the protocol information for the primary adapter, such as the IP address must be configured properly. Click OK to exit this message box. 7. Restart the HP NCU and verify the adapter protocol s were configured properly. Select the adapter and click Properties . Uninstalling teamed adapters If you need to completely uninstall a teamed adapter (n ot simply replace it), perform these steps to ensure proper operations: 1. Select the Team and click Dissolve . 2. Click OK on the NCU Main window to accept the changes. 3. Physically remove the adapter from the server. Configuring HP adapters The Properties dialog for the HP Network Team Loca l Area Connection contains the following tabs and buttons. General Tab ⢠Connect usingâ HP Network Team. Components that are selected are used by this connection. These include: o Client for Microsoft® Networks o HP Network Configuration Utility o Network Load Balancing o File and Printer Sharing for Microsoft® Networks o Network Monitor Driver o Net BEUI Protocol o Internet Protocol (TCP/IP) NOTE: Other protocols may apply. ⢠Descriptionâ Describes the selected c omponent ⢠Show icon in taskbar when connectedâ Enables/Disables the HP connection tray icon Buttons ⢠Configure âDisplays Windows® 2000 device properties o General (Troubleshooter)âDisable Device option o Driver (Name, Version number, Driver D etails; Install/Uninstall/Update driver buttons) ⢠Install âInstalls the selected protocol ⢠Uninstall âUninstalls the selected protocol ⢠Properties âAccesses dialogs on the protocol for the selected device. Sharing Tab Sharing allows other servers or your local network to access external resource s through this connection. The options on the Sharing tab are:
Team properties 135 ⢠Shared Access (Enable) âIf selected, it allows Internet connection sharing. You can share Applications and Protocols. ⢠Applications tab âDisplays network applications to be enabl ed for servers sharing this connection. ⢠Services tab âDisplays services to be provid ed to the remote network. HP NCxxxx Local Area Connection The HP NCxxxx Local Area Connection window contains the following tabs and buttons: General Tab The options on the General tab are as follows: ⢠Connect using âName of the adapter or module connection. Components that are checked are used by this connection. These include: o Client for Microsoft® Networks o HP Network Configuration Utility o Network Load Balancing o File and Printer Sharing for Microsoft® Networks o Network Monitor Driver o Net BEUI Protocol o Internet Protocol (TCP/IP) NOTE: Other protocols may apply. ⢠Descriptio nâDescribes the select ed component ⢠Show icon in taskbar when connected âEnables/Disables the HP tray icon Buttons ⢠Configure buttonâ Windows® 2000 device properties device. o General Tab Device status Troubleshooter Device usage o Advanced Tab The following options may vary with different adapters. 802.1p QOS (Enable/Disable) Checksum Offload Ethernet@WireSpeed Flow Control Media Connector (AUI/BNC or UTP) Locally Administered Address Number of Receive Descriptors Number of Transmit Descriptors Speed/Duplex
Team properties 136 WakeUp Capabilities WOL Speed ⢠Driver Tab o Name, Version number, Driver Details o Install o Uninstall o Update driver ⢠Resources Tab o Resource settings o Conflicting device list ⢠Install button âInstalls the selected protocol ⢠Uninstall button âUninstalls the selected protocol ⢠Properties button âAccesses dialogs on the protoc ol for the selected device. Sharing Tab Sharing allows other servers or your local network to access external resource s through this connection. The options on the Sharing tab are: ⢠Shared Access (Enable) âIf sele cted it allo ws Inte rnet conne ct ion sharing. You can share applications and protocols. o Applications âNetwork applications to be enabled for servers sharing this connection o Services âServices to be provided to the remot e network Windows NT Teaming The Fast Ethernet and Gigabit Ethernet adapters prov ide the followi ng options for increasing throughput and fault tolerance when running Windows NT® 4.0. ⢠Network Fault Tole rance (NFT). NFT provides the safety of an additional backup li nk between the server and hub/switch. NFT is implemented with a Primary adapter and a Secondary (backup) adapter(s). During normal operations, if the Primary adapter fails, the link to the Secondary adapter automatically takes over. ⢠Transmit Load Balancing (TLB). TLB creates a team of adapters to increase a server's throughput. If you configure TLB, NFT is automatically configured . TLB works with any industry-standard 10BASE-T, 100BASE-TX, or Gigabit switch. All transmit IP fr ames are load balanced among all team members based on a Load Balancing algorithm in the team ing device driver. The Current Primary adapter transmits all non-IP frames, and receives all frames for the team. If a failover event occurs, one o f the non-Primary adapters assumes the role of Current Pr imary adapter, and transmit IP packets are load balanced among all remaining team members. If a fa ilure occurs in any of the non-Primary adapters, the packets are load balanced amon g all remaining team members. To use TLB, you must have at least two NCxxxx a dapters (ports) capable of the same speed and linked to the same Layer 2 (Broadcast) sw itch insta lled in your server. They cannot be connected to the same switch and/or be members of different VLANs.With TLB, as you add adapters to your server, you can group them in teams to provide a si ngle virtual adapter with increased transmission bandwidth. For example, a TLB team containing four HP Fast Ethernet adapters configured for full- duplex operation would provide an aggregate maximum transmit rate of 400 Mb ps and a 100 Mbps receive rate resulting in a total bandwidth of 500 Mbps.
Team properties 137 ⢠TLB load balances IP traffic under Windows NT® and IP and IPX traffic under NetWare. ⢠Switch-assisted Load Balancing (SLB). SLB increases a server's through put using Fast Ethernet or Gigabit Ethernet ports. SLB load balances all traffic . Unlike TLB, SLB can be configured to increase both transmitting and receiving ch annels between your server and switch. For example, an SLB team containing four HP Fast Ethern et adapters configured for full-duplex operation would provide an aggregate maximum transmit rate of 400 Mbps and an aggregate maximum receive rate of 400 Mbps, resulting in a tota l bandwidth of 800 Mbps. SLB works only with SLB-en abled switches, such as the Catalyst 500 0 series. The SLB software continuously analyzes loading on each adapter an d balances network traffic across the adapters as needed. Adapter teams configured f or SLB also provide the benefits of NFT. To use SLB, you must have at least two HP Fast Et hernet or Gigabit adapters installed in your server and linked to the same SLB-enabled switch. For SLB, you can only team 100 Mbps adapters together (no 10 Mbps nor Gigabit), with the exception of the NC71xx and NC77xx server adapters, which are Gigabit and can run at 10/100 Mbps. NOTE: Configuring Switch-assisted Load Balancing typically requi res pairs of ports (two or four ports). Refer to the documentation that sh ipped with your SLB-c apable switch for more information. After your switch is correctly set up, you can configure the adapters in the server with the HP Networking Configuration Utility. Before configuring adapter teaming Before you configure th e HP adapter for adapter teaming, you need to install the adapters and make the appropriate connections to hubs and switches. Installing adapters 1. Install at least two adapters in a Windows NT® 4.0 system as describe d in Windows NT® 4.0 Installation Notes. 2. After installation is comple te, restart Windows NT® 4.0. NOTE: Windows NT® 4.0 Service Pack 5 or late r is required for implementing adapter teaming properly. Install the service pa ck before configuring adapter teaming. Connecting to hubs and switches ⢠For Fast Ethernet adapters, if connected to a hub, each adapter in a team must be connected to a port that is in the same collision domain. ⢠For any adapter connected to a switch, each adapter in a team must be connected to a port which is on the same Layer 2 (Broadcast) network. Installing and configuring teaming NOTE: You can configure a maximum of eight teams with eight adapters in each team for a maximum of 64 ports. 1. Double-click the Network Icon in the Control Panel. This launches the Network Control Panel applet. 2. Select the Adapter tab and double-click the adapter found in the Network Adapters list box. This will launch the CPQNTAC utility. NOTE: To modify the networking configuratio n in Windows NT® 4.0 through CPQNTAC, you must be logged on to the system as an Adminstrator.
Team properties 138 3. Select two to eight of the listed adapters by single -clicking them. Individual stand-alone adapters will remain selected as you click them. After you have selected all the desired ad apters for a team, click the Team button on the upper-right of the main dialog window. IMPORTANT: If the proper Service Pack level (SP5) has not been installed on the system, a popup message displays warning you of this condit ion. If this message is displayed, install the correct Service Pack before proceeding. 4. The Teaming Property Sheet appears for the newly created team. The first page displayed is the Teaming Controls Sheet. The default mode of oper ation for a team is Faul t Tolerant mode. Change this to the desired mode of operation by selecting either Fault Tolerant or Load Balancing at the top of the page. 5. For a Fault Tolerant team, select among the following modes: Manual. Causes a failover when the user clicks Switch Now . In this mode, a failover will also o ccur if the Current Primary adapter is dropped from th e team. Recommended only for troubleshooting purposes. Fail on Fault (Default). If a failover event occurs, a non-Primary adapter takes on the role of the Current Primary adapter. This new Current Primary adapter remains primary unless another failover event occurs, or the system is rebooted. Smart Switch. If a failover event occurs, a non-Primary ad apter takes on the role of Current Prim ary adapter. This new Current Primary adapter remains primary until one of the following occurs. The Preferred Primary adapter becomes available, and a failback occurs, in which case the Preferred Primary adapter resumes its role as Current Primar y adapter. Or, another failover event occurs, in which case, a non-Primary adapter takes on the role of Current Primary adapter. If this mode is selected, make sure the desired Pr imary adapter is displayed in the Preferred Primary NIC list box. A single Fault Tolerant team CAN contain a mixture of Fast Et hernet and Gigabit adapters. 6. For a Load Balancing team, select among the following modes: o Transmit Load Balancing o Switch-assisted Load Balancing 7. Select the Settings tab at the top of the Teaming Property Sheet. On the Settings tab, you may configure the following it ems: Team Network Address. Used to "Locally Administer" the team's un ique receive node address on the network like you would locally administ er a single adapter's node a ddress. Enable Heartbeats. This check box is used to enable/d isable teaming Heartbeats. Heartbeats provide an additional level of adapter failure de tection above simple li nk failure detection. Heartbeat detection, if enabled, will cause Heartb eat packets to be transmitted between the Primary adapter and Secondary adapters to ensure proper transmit and receive functionality among the adapters in the team. Heartbeat timer interval. This value specifies th e amount of time to wait before transmitting Heartbeat packets between adapters in the team. Heartbea ts are not transmitted between the adapters if normal network transmit/receive traffic is passing through the adapters. However, in the absence of network traffic, after the specified Heartbeat time out period expires (the timer interval value), Heartbeat packets are sent. This value MUST be greater than 3000 ms (3 seconds). 8. Click OK on the Teaming Property Sheet dialog to complete the configuration of the team. 9. Repeat Steps 3â8 to create additional teams as desired. 10. Click Close at the bottom of the Main CPQNTAC dial og to finish configuration of teams.
Team properties 139 11. Click Close at the bottom of the Network Control Pane l applet to finish configuration of network devices. 12. Binding analysis will be done to bind newly installed adapte rs and teams to all installed communication protocols. You may be prompted to enter protocol specific information (s uch as IP information) for newly installed adapters and teams. 13. When prompted, click Yes to reboot Windows NT® . NetWare teaming IMPORTANT: Teaming NC150, NC32x, NC37x, NC1020, NC6770, or NC77 xx adapters with NC110x, NC310, NC31xx, NC340x , NC360x, NC364x, NC61xx, or NC71xx adapters is not supported for NetWare. The HP Advanced Network Services driver name is CPQANS.LAN. CPQANS requirements : ⢠Use with NetWare 5.1 and 6.x ⢠Always install the latest Novell Support Pa cks ⢠Always install the latest HP NSSD CPQANS teams with the following HP server adapters : ⢠HP NC110T PCI Express Gigabit Server Adapter ⢠HP NC310F PCI-X Gigabit Server Adapters ⢠HP NC340T PCI-X Quad-Port Gigabit Server Adapter s ⢠HP NC360T PCI Express Dual Port Gigabit Server Ad apter ⢠HP NC364T PCI Express Quad Port Gigabit Server Adapter ⢠HP NC61xx Gigabit Server Adapters ⢠HP NC31xx Fast Ethernet Server Adapters ⢠HP NC71xx Gigabit Server Adapters CPQANS requires the following Novell settings : ⢠SET MINIMUM PACKET RECEIVE BUFFE RS = 2000 (or larger) ⢠SET MAXIMUM PACKET RECEIVE BUFFE RS = 10000 (or larger) ⢠SET MAXIMUM PHYSICAL RECEIVE PAC KET SIZE = 2048 (or larger) Options for increasing throughput and fault tolerance HP Gigabit, Fast Ethernet, and Ethernet server adapters teamed with CPQANS provide the following options for increasing thr oughput and fault tolerance for NetWare 5.1, 6.0, and 6.5. ⢠Network Fault Tole rance (NFT). Creates a team of adapters with the Primary adapter in the active state and the remaining adapters in a standby state. Primary specifies which adapter in the team is the Preferred Primary adapter. If the Preferred Prim ary adapter fails and is later replaced, or if a cable is reattached to fix a failure on the pr imary adapter, the secondary adapter automatically passes control back to the Preferred Primary. Works with any Ethernet switch or hub. NFT is the only team mode that allows for a team of different speed adapters. Refer to the discussion on mixed teams in the "Important teaming notes" section for more informat ion.
Team properties 140 o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=NFT NOTE: NFT is the default mode. Issuing a "LOAD CPQANS COMMIT" command without specifying a "MODE" will result in a team configured for the Network Fault Tolerance mode. ⢠Adaptive Load Balancing (ALB). Creates a team of two or more adapters to increase transmission throughput while providing adapter redundancy. HP recommends that all adapte rs in an ALB team be the same speed (either all Gigabit or all Fast Ethernet). In an ALB team, the primary adapter is used for receiving and transmitting, while all ot her team members transmit. ALB w orks with any Ethernet switch. No performance gain is made with a hub or repeater. o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=ALB ⢠Fast EtherChannel (FEC). Creates a team of HP Fast Ethernet server adapters to increase receive and transmit throughput using a switch with Fast Ethe rChannel technology. As with NFT, FEC provides redundancy. An FEC team uses the switc h FEC tec hnology to perform the load balancing rather than the server. The switch algorithm determin es which ports are used for transmits and receives. All adapters are capable of receiving and transmitting at the same time. o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=FEC NOTE: FEC teaming requires a switch with Fa st EtherChannel technology. Refer to the "Important teaming notes" section for more information about configuring your switch for FEC or GEC teams. ⢠Gigabit EtherChannel (GEC). Creates a team of HP Gigabit server adapters to increase receive and transmit throughput using a switch with Giga bit EtherChannel technolo gy. As with NFT, GEC provides redundancy. A GEC team uses the switch GEC technology to perform the load balancing rather than the server. The switch algorithm determin es which ports are used for transmits and receives. All adapters are capable of receivin g and transmitting at the same tim e. o Syntax: LOAD CPQANS MODE=XXX o Example: LOAD CPQANS MODE=GEC NOTE: GEC teaming requires a switch with Giga bit EtherChannel technology. Refer to the "Important teaming notes" section for more information about configuring your switch for FEC or GEC teams. ⢠802.3ad (Link Aggregation Control Protocol â LACP). Creates a team of HP Gigabit server adapters to increase receive and transmit throughput us ing a switch that supports 802.3ad with LACP protocol. o Syntax : LOAD CPQANS MODE=xxx o Example : LOAD CPQANS MODE=AD Important teaming notes ⢠Configuring Fast EtherChannel (FEC) or Gigabit Et herChannel (GEC) typically requires pairs of ports (two or four ports). However, review the documentation that comes with your FEC- or GEC-capable switch for more information. After yo ur switch is correctly configured, yo u can configure your GEC or FEC team in accordance with your switch. You sh ould always configure your switch first then your server for GEC or FEC tea ms.
Team properties 141 ⢠Every member of a CPQANS team must be connected to the same network segment. If you have multiple CPQANS teams, each team must be on a separate network segment. If you are connected to a VLAN-capable switch, each HP server adapter in a team must be connected to a port with the same VLAN ID number (same network segment). NOTE: VLAN is not supported on heterogeneous teaming. ⢠For mixed teams (different speed adapters), Networ k Fault Tolerance (NFT) is the only team mode that allows for a team of different speed adapters. Such a team can contain any combination of HP NC310x Gigabit Server Adapters, NC61xx Giga bit Server Adapters, NC71xx Gigabit Server Adapters, NC31xx Fast Ethernet Server Adapters an d so on. With a mixed team, you can designate the highest-performance adapter as the preferred primary adapter. To set the preferred primary adapter, use the "PRIMARY" keyword when binding CPQANS to the base driver for the Ethernet server adapter. Adapter performance hierarchy: Gigabit server ad apter (1000 Mbps) is the fastest followed by the Fast Ethernet (10/100 Mbps) server adapter. Additional information The following are some commands that can be help ful in configuring or troubl eshooting teaming . ⢠CPQANS STATUS o Description: This command displays th e current status for the CPQANS team. o Syntax: LOAD CPQANS STATUS o Example: LOAD CPQANS STATUS If you have more than one team, then you can is sue the same command with the team's ID number. o Syntax: LOAD CPQANS STATUS TEAM=<IDnumber> o Example: LOAD CPQANS STATUS TEAM=2 ⢠CPQANS PRIMARY o Description: This command cr eates a preferred primary. o Syntax: BIND CPQANS <adapter name> PRIMARY o Example: BIND CPQANS N100_1_E82 PRIMARY In this example, the adapter "N100_1_E82" is the preferred primary for the CPQANS team. PRIMARY specifies which adapter in the team is the Preferred Primary adapter. If the Preferred Primary adapter fails and is later replaced, or if a cable is reattached to fix a failure on the primary adapter, the secondary adapter automa tically passes control back to the Preferred Primary. For NFT mode teams, if you have a server with an HP Gigabit Server Adapter and an HP Fast Ethernet Server Adapter, you would want the Gigabit Server Adapter to be the Preferred Primary. In this scenario, if the Gigabit Server Adapter fails, the Fast Ethernet S erver Adapter will take over. Then, when the Gigabit Server Adapter is replaced, it will automatically revert to being the Primary Adapter in the team. NOTE: This command is only supported during the BIND command.
Team properties 142 Error conditions If you receive the error messa ge "Failed to create new team," check the following: ⢠All adapters have the same frame types. ⢠The CPQANS team has the same frame type as the adapters. ⢠That CPQANS is loaded once for each team. ⢠A "BIND CPQANS <adapter name>" command has b een issued for each adapter in the team. ⢠All adapters in a team are connect ed to the same network segment. ⢠CPQANS is not being bound to an unsupported adapter. NOTE: The "CONFIG" command will not reflect CPQANS until the command "LOAD CPQANS COMMIT" has be en successfu lly issued. Setting NetWare probe and balance variables NetWare provides several probe and balance variables that provide the optimum performance for most configurations. However, you can modify these to me et your special configuration needs. Each of these variables is described below. You can use the followin g formula to determine th e relative range of many of these parameters: probe_send_interval<(probe_check_interval probe_recheck_interval*probe_retry_count) ⢠load CPQANS balance_interval=xxx [team=nnn] o Default = 180 o Range = 18â1080 o Description: Changes the team's balance interv al. XXX counts 1/18 second ticks. Only even numbers are supported. Odd numbers are round e d down. This can also be used after the COMMIT command. ⢠load CPQANS balance_set_default [team=nnn] Description: Restores the te am's balance interval to the factory default setting. ⢠load CPQANS probes=[on | off] [team=nnn] o Default = Probes enabled o Description: Enables and disables probe packet generation. Probes are not supported in FEC, GEC, and IEEE 803.3ad modes. This can al so be used after the COMMIT command. ⢠load CPQANS probes=[broadcas t|multicast] [team=nnn] o Default = Probes sent in broadcast o Description: Hangs probe addressin g to broadcast or multicast. This cannot be used after the COMMIT command. ⢠load CPQANS probe_retr y_count=xxx [team=nnn] o Default = 7 o Range = Relative to other pa rameters o Description: Changes the team's probe retry coun t value. XXX is the number of retries to send probes, in case the probes are not received by other members of this team. This can also be used after the COMMIT command. ⢠load CPQANS probe_burs t_size=xxx [team=nnn]
Team properties 143 o Default = 3 o Minimum = 1 o Description: Changes the team's number of probes to send in case of a retry. This can also be used after the COMMIT command. ⢠load CPQANS probe_send_i nterval=xxx [team=nnn] o Default = 18 o Range = Relative to other pa rameters o Description: Changes the time interval sending of regular probes. XXX coun ts 1/18 second ticks. Only even numbers are supported. Odd numbers ar e rounded down. This can also be used after the COMMIT command. ⢠load CPQANS probe_check_ interval=xxx [team=nnn] o Default = 2 o Range = Relative to other pa rameters o Description: Changes the time interval betw een probes and checks them. XXX counts 1/18 second ticks. Only even numbers are supported. Odd numbers are rounded down. This can also be used after the COMMIT command. ⢠load CPQANS probe_recheck_ interval=xxx [team=nnn] o Default = 2 o Range = Relative to other pa rameters o Description: Changes the time interval between probe retries and checks them. XXX counts 1/18 second ticks. Only even numbers are supported. Odd numbers are rounded down. This can also be used after the COMMIT command. ⢠load CPQANS probe_set_default [team=nnn] Description: Restores the team's pr obe settings to the factory default settings. (Does not affect On /Off or Multicast/Broadcast settings.) ⢠load CPQANS tx_ecbs_to_use=xxx [team=nnn] o Default = 64 o Minimum = 0 o Description: Number of Tx EC Bs to allocate when the Virtual Adapter is created. ⢠load CPQANS RESET LBN=xxx [team=nnn] Description: Resets a bound adapt er. This accesses the adapter by its Logical Board Number. Supported keywords For a list of supported keywords, see Adv a nced Server Program for NetWare (on page 110 ). Examples of NetWare teaming The following are examples of load and bind statemen ts that you should add to your AUTOEXEC.NCF file to set up adapter Teaming. HP recommends that you try some of these examples from the command line before writing them to the AUTOEXEC.NCF file. EXAMPLE 1: Simple NFT team The following is an example of an NFT Team that co nsists of an HP Gigabit server adapter and an HP Fast Ethernet server adapter. For performance reasons the Gigabit server adapter is the primary adapter and the Fast Ethernet server adapter is the backup.
Team properties 144 Note that ONLY an NFT mode Team can combine an HP Gigabit server adapter (server adapters that us e N1000.LAN) and an HP Fast Ethernet ser ver adapter (use N100.LAN). (Exception: The NC71xx Gigabit over copper server adapters can be teamed with adapters of different speeds.) See Overview of NetWare teaming (" NetWare teaming " on page 139 ) for information about mixing adapters. Note the following: ⢠For simplicity, in this example, only the default frame type 802.2 is used. By default, CPQANS uses the frame type 802.2, so you do not need to specify 802.2 on the command line. ⢠By default, CPQANS uses the NFT mode. In this example, only one instance of each adapter is present in the system, so you do not need to specify a s lot number. ⢠This example checks the status of the Team after each command. Example: 1. Load CPQANS.LAN to form the basis of a Team. LOAD CPQANS FRAME=ETHERNET_802.2 Next, check the status of the Team. LOAD CPQANS STATUS 2. Load the base driver for each adapter and frame type for the Team. As this example uses only 802.2, each adapter has its base driver loaded only once. LOAD N1000 LOAD N100 Next, check the status of the Team. LOAD CPQANS STATUS 3. When you load CPQANS.LAN, it looks like a protocol , so you can bind it to the physical adapter. Bind CPQANS.LAN to each physical adapter in the Team. Notice the key word "PRIMARY." For performance reasons the Gigabit server adapter should be the primary adapter. BIND CPQANS N1000 PRIMARY BIND CPQANS N100 Next, check the status of the Te am. Notice that the Team now has two adapter instances with the Team. Also notice that the Team is in the "BEFORE COMMIT" state. The HP Gigabit server adapter is the "PRIMARY. " LOAD CPQANS STATUS 4. Finalize the Team with the "COMMIT" command. Sinc e we have not chosen the mode of the Team, it is NFT by default. LOAD CPQANS COMMIT Next, check the status of the Team . Notice that NFT is now the Team mode and t hat the Team is now in the "COMMIT-DONE" state. No changes to th e Team can be made when the Team is in the "COMMIT-DONE" state. LOAD CPQANS STATUS 5. Now that the Team is in the "COMMIT-DONE" st ate, the CPQANS Team appears as a virtual adapter. You may now bind a protocol to CPQA NS. ALWAYS ensure that your adapters and CPQANS are using the same frame typ e. BIND IPX CPQANS NET=102 Next, check the status of the Team. LOAD CPQANS STATUS Example 2: TLB Team with Ethernet_II frame type This example is slightly more complicated. In this example, we will only use one frame type, Ethernet _II. Since Ethernet_II and TLB are not the CPQANS defaults , you must specify them both at the command line.
Team properties 145 This example also differs from th e previous example in that the Te am will be in the Transmit Load Balancing (TLB) mode. This example assumes a server with two HP Fast Ethern et server adapters, one in PCI Slot 1 and the other in PCI Slot 2. Note that ONLY an NFT mode Team can combine an HP Gigabit server adapter (server adapters that use N1000.LAN) and an HP Fast Ethe rnet server adapter (server adapters that use N100.LAN). See Overview of NetWare teaming (" NetWare teamin g " on page 139 ) for more information about mixing adapters. Use the "LOAD CP QANS STATUS" command to view the Team status after each of the following commands. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS.LAN to form the ground work for the Team. Since 802.2 is the default frame type we need to specify the frame type Ethernet_II. In a ddition, TLB should be specified, since NFT is the default Team "MODE." LOAD CPQANS FRAME=ETHERNET_II 3. Load the base drivers for each adapter in the Te am. Just as 802.2 is the default frame type for CPQANS, 802.2 is also the default frame type for HP Fast Ethernet. Thus, Ethernet _II must be specified during the driver load. In addition, since two different HP Fast Ethernet server adapters are in use, each one needs to be assi gned a name for later reference. LOAD N100 SLOT=1 FRAME=ETHERNET_II NAME=PRIMARY_N100_II LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=SECONDARY_N100_II 4. Bind CPQANS.LAN to each physical adapter usin g the names you defined when the base drivers were loaded. BIND CPQANS PRIMARY_N100_II BIND CPQANS SECONDARY_N100_II 5. If you have not specified your Te am as an Transmit Load Balance (TLB) Team, you may do it now. LOAD CPQANS MODE=ALB 6. If you forgot to choose the frame type or Te am mode during the or iginal "LOAD CPQANS" command, you can specify the desired frame type and Team mode any time before the "COMMIT" command. 7. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 8. Bind both IPX and IP as if CP QANS were a physical adapter BIND IPX CPQANS NET=112 BIND IP CPQANS ADDRESS=192.168.1.1 MASK=255.255.255.0 Example 3: Two teams (one NFT and one SLB) This example creates two Teams. The first is an NFT Team consisting of one HP Gigabit server adapter (N1000.LAN) and one HP Fast Ethernet server adap ter (N100.LAN). The second Team has four HP Fast Ethernet server adapters (N100.LAN) configured as a Switch-assisted Load Balancing (SLB) Team. SLB is the implementation of port aggre gation. This example uses IPX on Ethernet 802.2 and IP with the Ethernet II frame type. See Overview of NetWare teaming (" NetWare teamin g " on page 139 ) for more information about SLB Team implementation. Use the "LOAD CPQANS STATUS" command to vi ew the Team status after each command. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS.LAN for the first Team. By default yo u do not need to include the first Team's ID number on the command line. However, you will need to include it for every Tea m after the first. If
Team properties 146 you do not include "TEAM =1" CPQANS defaults to T eam 1. Thus, the "TEAM" command line switch is optional for the first Team. LOAD CPQANS FRAME=ETHERNET_802.2 NAME=GIG_FT_E82 TEAM=1 MODE=NFT LOAD CPQANS FRAME=ETHERNET_II NAME=GIG_FT_EII TEAM=1 3. Load the base drivers for the first Team. LOAD N1000 SLOT=1 FRAME=ETHERNET_802.2 NAME=GIG_E82 LOAD N1000 SLOT=1 FRAME=ETHERNET_II NAME=GIG_EII LOAD N100 SLOT=2 FRAME=ETHERNET_802.2 NAME=100_E82 LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=100_EII 4. Bind CPQANS to the base driver for each adapter and each frame type. Bind to the name of the base driver instance (assigned when the driver was loaded) and specify the Team ID number. BIND CPQANS GIG_E82 TEAM=1 BIND CPQANS GIG_EII TEAM=1 BIND CPQANS 100_E82 TEAM=1 BIND CPQANS 100_EII TEAM=1 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT TEAM=1 6. Bind IPX to 802.2 and bind IP to Ethernet II. BIND IPX GIG_FT_E82 NET=102 BIND IP GIG_FT_EII ADDR=128.181.11.1 MASK=255.255.255.0 7. Load CPQANS with the second team. Name each frame type with the "NAME" keyword and specify the second Team with the "TEAM=2" keywor d. You can also specify the Team "MODE" FEC. LOAD CPQANS FRAME=ETHERNET_802.2 NAME=FEC_TEAM_E82 TEAM=2 MODE=FEC LOAD CPQANS FRAME=ETHERNET_II NAME=FEC_TEAM_EII TEAM=2 8. Load the base drivers for the second Team. LOAD N100 SLOT=6 FRAME=ETHERNET_802.2 NAME=FEC_1_E82 LOAD N100 SLOT=6 FRAME=ETHERNET_II NAME=FEC_1_EII LOAD N100 SLOT=5 FRAME=ETHERNET_802.2 NAME=FEC_2_E82 LOAD N100 SLOT=5 FRAME=ETHERNET_II NAME=FEC_2_EII LOAD N100 SLOT=4 FRAME=ETHERNET_802.2 NAME=FEC_3_E82 LOAD N100 SLOT=4 FRAME=ETHERNET_II NAME=FEC_3_EII LOAD N100 SLOT=3 FRAME=ETHERNET_802.2 NAME=FEC_4_E82 LOAD N100 SLOT=3 FRAME=ETHERNET_II NAME=FEC_4_EII 9. Bind CPQANS to the base driver for each adapter and each frame type. Bind to the name of the base driver instance (assigned when the driv er was loaded) and specify the Team number. BIND CPQANS FEC_1_E82 TEAM=2 BIND CPQANS FEC_1_EII TEAM=2 BIND CPQANS FEC_2_E82 TEAM=2 BIND CPQANS FEC_2_EII TEAM=2 BIND CPQANS FEC_3_E82 TEAM=2 BIND CPQANS FEC_3_EII TEAM=2 BIND CPQANS FEC_4_E82 TEAM=2 BIND CPQANS FEC_4_EII TEAM=2 10. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT TEAM=2 11. Bind IPX to 802.2 and bind IP to Ethernet II. BIND IPX FEC_TEAM_E82 NET=202 BIND IP FEC_TEAM_EII ADDR=128.181.22.1 MASK=255.255.255.0 Example 4: One NC series server adapter and two VLANs NOTE: VLAN is not supported on heterogeneous teaming.
Team properties 147 For CPQANS.LAN to work in VLAN mode, the adap ter (or adapters) needs to be connected to a "tagged" port (a "trunk port") on an IEEE 802.1Q compliant switch. This example uses 802.1Q VLAN, which is suppo rted by all HP server adapters that support CPQANS.LAN. (Exception: The NC6134, NC6132, an d NC6133 adapters do not support V LANs.) This example uses an HP Fast Ethernet server adapter an d is configured for two VLANs (VLAN 5 and VLAN 6). It also uses all four NetWare frame types and binds both the IPX and IP protocols. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS and specify the VLANID for each fram e type and each VLAN the server is supposed to be part of. With a single Te am configuration, "Team=1" is optional, since that is the default setting. a. First, load CPQANS for VLAN 5, assign ing a name of 5xx to each instance: LOAD CPQANS VLANID=5 FRAME=ETHERNET_802.2 NAME=582 TEAM=1 LOAD CPQANS VLANID=5 FRAME=ETHERNET_802.3 NAME=583 TEAM=1 LOAD CPQANS VLANID=5 FRAME=ETHERNET_SNAP NAME=5SP TEAM=1 LOAD CPQANS VLANID=5 FRAME=ETHERNET_II NAME=5II TEAM=1 b. Next, load CPQANS for VLAN 6, assign ing a name of 6xx to each instance: LOAD CPQANS VLANID=6 FRAME=ETHERNET_802.2 NAME=682 TEAM=1 LOAD CPQANS VLANID=6 FRAME=ETHERNET_802.3 NAME=683 TEAM=1 LOAD CPQANS VLANID=6 FRAME=ETHERNET_SNAP NAME=6SP TEAM=1 LOAD CPQANS VLANID=6 FRAME=ETHERNET_II NAME=6II TEAM=1 3. Load the base driver for each frame type. LOAD N100 SLOT=5 FRAME=ETHERNET_802.2 NAME=E82 LOAD N100 SLOT=5 FRAME=ETHERNET_802.3 NAME=E83 LOAD N100 SLOT=5 FRAME=ETHERNET_SNAP NAME=ESP LOAD N100 SLOT=5 FRAME=ETHERNET_II NAME=EII 4. After loading CPQANS for all the desired VLANs, bind CPQANS to the base driver using the name you defined when the base driv ers were originally lo aded. BIND CPQANS E82 TEAM=1 BIND CPQANS E83 TEAM=1 BIND CPQANS ESP TEAM=1 BIND CPQANS EII TEAM=1 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 6. Bind the protocols to each instance (VLAN and fram e type) of CPQANS in the same manner as if it were a physical adapter. BIND IPX 582 NET=502 BIND IPX 583 NET=503 BIND IPX 5SP NET=505 BIND IPX 5II NET=511 BIND IPX 682 NET=602 BIND IPX 683 NET=603 BIND IPX 6SP NET=605 BIND IP 6II ADDR=192.168.62.1 MASK=FF.FF.FF.0 Example 5: TLB team with multiple VLANs
Team properties 148 NOTE: VLAN is not supported on heterogeneous teaming. You can implement Teaming and VLAN at the same ti me. Because the script gets large quickly, this example only uses one frame type. To have both IP an d IPX bound to it, the Ethernet_I I frame type will be used instead of Novell's default 802.2. A homogene ous Team of two load-balanced HP Fast Ethernet server adapters configured as members of five different VLANs will be used here. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP 2. Load CPQANS.LAN with the VLANID key word once fo r each VLAN that the server is intended to communicate with. Assign a name so each instance can be identified later. LOAD CPQANS FRAME=ETHERNET_II VLANID=10 NAME=V10 TEAM=1 MODE=ALB LOAD CPQANS FRAME=ETHERNET_II VLANID=20 NAME=V20 TEAM=1 LOAD CPQANS FRAME=ETHERNET_II VLANID=30 NAME=V30 TEAM=1 LOAD CPQANS FRAME=ETHERNET_II VLANID=40 NAME=V40 TEAM=1 LOAD CPQANS FRAME=ETHERNET_II VLANID=50 NAME=V50 TEAM=1 3. Load the base drivers for each adapter in the Team. LOAD N100 SLOT=1 FRAME=ETHERNET_II NAME=PRIMARY_II LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=SECONDARY_II 4. Bind CPQANS to each physical adapter. BIND CPQANS PRIMARY_II BIND CPQANS SECONDARY_II 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 6. Bind both IPX and IP to CPQANS for each VLAN. 7. Bind IPX first. BIND IPX V10 NET=1011 BIND IPX V20 NET=2011 BIND IPX V30 NET=3011 BIND IPX V40 NET=4011 BIND IPX V50 NET=5011 Then bind IP. BIND IP V10 ADDR=192.168.1.1 MASK=FF.FF.FF.0 BIND IP V20 ADDR=192.168.2.1 MASK=FF.FF.FF.0 BIND IP V30 ADDR=192.168.3.1 MASK=FF.FF.FF.0 BIND IP V40 ADDR=192.168.4.1 MASK=FF.FF.FF.0 BIND IP V50 ADDR=192.168.5.1 MASK=FF.FF.FF.0 Example 6: IEEE VLAN NOTE: VLAN is not supported on heterogeneous teaming. This example creates two IEEE VLANs on a single HP Fa st Ethernet server adapter, loads two frame types (Ethernet_802.2 and Ethernet_II) on each VLAN, bind s IPX to each loaded frame type, and binds IP to each instance of the Ethernet_II frame type. 1. Load TCPIP.NLM. This can be done at any point before binding IP. LOAD TCPIP
Team properties 149 2. Load CPQANS twice for each frame type, once for each VLAN. As sign the VLANID number keyword for each instance. LOAD CPQANS FRAME=ETHERNET_802.2 NAME=E82_Vl7 TEAM=1 VLANID=7 LOAD CPQANS FRAME=ETHERNET_II NAME=EII_Vl7 TEAM=1 VLANID=7 LOAD CPQANS FRAME=ETHERNET_802.2 NAME=E82_Vl8 TEAM=1 VLANID=8 LOAD CPQANS FRAME=ETHERNET_II NAME=EII_Vl8 TEAM=1 VLANID=8 3. Load the base driver for each frame type: LOAD N100 SLOT=2 FRAME=ETHERNET_802.2 NAME=PRI_E82 LOAD N100 SLOT=2 FRAME=ETHERNET_II NAME=PRI_EII 4. Bind CPQANS to each frame type on the physical adapter. BIND CPQANS PRI_E82 TEAM=1 BIND CPQANS PRI_EII TEAM=1 5. Finalize the Team with the "COMMIT" command. LOAD CPQANS COMMIT 6. Bind the protocols to CPQANS for each VLAN and frame type. BIND IPX E82_Vl7 NET=702 BIND IPX EII_Vl7 NET=711 BIND IP EII_Vl7 ADDR=192.168.72.200 MASK=255.255.255.0 BIND IPX E82_Vl8 NET=802 BIND IPX EII_Vl8 NET=811 BIND IP EII_Vl8 ADDR=192.168.82.200 MASK=255.255.255.0
VLAN properties 150 VL AN pr oper tie s In this section VLAN info rmation ............................................................................................................... .................. 150 VLAN information A Virtual LAN (VLAN) is a collection of network node s that share the same broadcast domain regardless of their physical location or connection point to the network. A VLAN serves as a logical workgroup with no physical barriers and allows users to share inform ation and resources as thou gh located on the same LAN. The HP Network Configuration Utility supports VL ANs in Windows S erver 2003 and Windows Server 2003 x64 environments. Before creating VLANs, you should be aware of the following: ⢠This feature supports IEEE 802.1Q VLAN tagging only. ⢠This feature supports a VLAN identifier ra nge of 1 to 4094. ⢠Up to 64 Ethernet VLANs can be defined for a team or a single, non-teamed adapter. ⢠Only one iSCSI VLAN can be defined on a multifunction adapter ⢠VLANs cannot be defined for individual adapters th at are members of a team. VLAN defin ition must be done at the Team level. ⢠If an adapter with VLANs is added to a team, th e VLANs defined for the adapter will be deleted. ⢠When a VLAN is defined for an adap ter, the properties of that adap ter will be modifi ed if necessary to allow proper VLAN functionality. For example, 802.1p QoS must be enabled on an adapter or team with VLANs. Likewise, when an adapter is ad ded to a team that has VLANs defined, then that adapter will have 802.1p QoS enabled. ⢠The VLAN defined as the âDefault VLANâ cannot be deleted if more than one VLAN is defined for a team. For example, if VLAN 1 needs to b e deleted and it is defined as the âDefault VLANâ, chan ge the âDefault VLANâ to a different VL AN. VLAN 1 can then be deleted. ⢠When creating a VLAN for a team, the Fast Path VLAN ID is disabled if the Spanning Tree Protocol is set to IEEE 802.1D. ⢠When the last VLAN is removed from an adapter or the adapter is removed from a team with VLANs, the user will be given the option to disabl e 802.1p QoS on the adapter or allow it to remain enabled. For more in formation: ⢠Removing a VLAN
ProLiant Essentials 151 Pr oL iant E ss ent ials In this section Networking pack installa tion ................................................................................................... .............. 151 Intelligent Networking Pack-Windows Edition .......................................................................................... 151 Intelligent Networking Pack-Linux Edition ...................................................................................... ........... 152 Accelerated iSCSI for stand-alone and c-Class BladeSystem Multifunctio n server ad apters ............................ 153 Accelerated iSCSI for embedded HP Multifunction serv er adapt ers............................................................. 153 Networking pack installation Included with the networking software that shippe d with your HP server adapter are a number of advanced features that extend the basic networking capabilities of your eq uipment. You can activate these advanced features through the purchase and in stallation of the foll owing ProLiant Essentia ls networking pack licenses: ⢠Intelligent Networking PackâWindows Edition ⢠Intelligent Networking PackâLinux Edition ⢠Accelerated iSCSI Pack for Embedded Multifunction Server Adapters Installing a networking pack license from the NCU The HP ProLiant Essentials Network License Manager a llows you to install and track network licenses on your system. To access the License Mana ger dialog: 1. From the NCU Main window, click License Manager. The License Manager dialog appears. 2. Click Help to access the License Manage r help file. NOTE: The HP Accelerated iSCSI feature comes st andard with stand-alone HP Multifunction Server Adapters and Multifunction Se rver Adapters for HP BladeSystem. Intelligent Networking Pack-Windows Edition The HP Network Configuration Utility (NCU) allows you to configure and monitor HP server network adapters running under Windows® 2000, Windows Server 2003, or Windows Server 2003 x64. Although the NCU provides an extensive list of features through the Foundation Pack, additional advanced networking capabilities are available with the installation of an HP ProLiant Essentials Intelligent Networking Pack license. The Intelligent Networking Pack (INP)âWindows Edition includes: ⢠Virus Throttle, a network pack et-filtering feature that helps slow down the spread of viru ses on your system. Virus Throttle monitors all outbound connect ion requests and counts the number of un ique connections. It detects abnormal ( âvirus-likeâ) behavior in the re quests, and slows down excessive connection requests to new hosts un til you can determine if they are viral in nature and take action. ⢠Advanced networking capabilities that allow ProLia nt servers to adapt and change the network path to achieve maximum reliability and performance. These features detect and analyze network
ProLiant Essentials 152 bottlenecks or broken network linkages, and steer traffic to the optimum route based on network conditions. Advanced networ king features include: o Fast Path FailoverâAllows a ProLiant ser ver to use the quickest available path to the core network for all server traffic. o Active Path FailoverâAllows a ProLiant server to detect blocked paths an d to redirect data along an unblocked path to the core network. o Router Path FailoverâProvides red undancy by grouping routers into one router g roup. One router serves as the active router. If that router fails, a second ro uter assumes rout er functionality. ⢠Dual Channel Load Balancing, a special type of te aming that allows the creation of two teams, called groups, inside of a single team. Each group is assigned one or more teamed ports and can be connected to a different switch to provide swit ch fault tolerance. Full inbound and outbound load balancing is provided across both groups. Should an y group completely fail, caus ed by a failure of all teamed ports in the group or by a failure of the groupâs switch, the team remains available through the other group. These types of Dual Channel Load Balancing teams can be configured : o Switch-assisted Dual Channel Load Balancing o Switch-assisted Dual Channel Network Fault Tolerance o Switch-assisted Dual Channel NFT and Preference Order o 802.3ad Dynamic Dual Channel Load Balancing o 802.3ad Dynamic Dual Channel Network Fault Tolerance o 802.3ad Dynamic Dual Channel NFT and Preference Order NOTE: Dual Channel Network Fault Tolerance te aming capabilities are not supported on systems running Windows® 2000. ⢠Cisco Discovery Protocol (CDP) and Link Layer Discov ery Protocol (LLDP) are both layer two protocols that are neighbor device discovery protocols that run on supported networ king devices such as routers, access servers, bridges, and switch es. Using CDP or LLDP, an HP adapter team can passively discover configuration information about directly connected networ k devices that have CDP or LLDP enabled. NOTE: LLDP is not supported on systems runnin g Windows® 2000. For additional information about INP features and how to install a license on your system, refer to the HP ProLiant Essentials Intellig ent Networking Pack License â Windows Edition User Guide located on the HP website ( http://www.hp.com ). For information on how to purchase an HP ProLia nt Essentials Intelligent Networking PackâWindows Edition license, go to the HP website ( http://www.hp.com ) or contact your HP reseller. Intelligent Networking Pack-Linux Edition The Intelligent Networking PackâLinux Edition includes Virus Throttle, a network packet-filtering feature that helps slow down the spread of virus es on your sy stem. Virus Throttle monitors all outbound connection requests and counts the number of unique connections. It detects abnorma l (âvirus-likeâ) behavior in the requests, and slows down excessive co nnection requests to new hosts until you can determine if th ey are viral in nature and take action.
ProLiant Essentials 153 To enable Virus Throttle for Linux, an Intelligent Netw orking PackâLinux Edition lice nse must be installed on the system. For additional information about Virus Throttle and how to install an Intelligent Networking Pack license on your system, refer to the HP ProLiant Essentials Intelli gent Networking Pack License â Linux Edition User Guide located on the HP website ( http://www.hp.com ). For information on how to purchase an HP ProLiant Essentials Intelligent Netw orking PackâLinux Edition license, go to the HP website ( http://www.hp.com ) or contact your HP reseller. Accelerated iSCSI for stand-alone and c-Class BladeSystem Multifunction server adapters The HP Accelerated iSCSI feature comes standard with HP stand-alone Multifunction server adapters and Multifunction server adapters for c-Class BladeSystem. The Accelerated iSCSI feature runs under Wind ows® 2000, Windows Server 2003, and Windows Server 2003 x64 and allows you to complete the following: ⢠Enable an iSCSI device ⢠Adjust the SRB Timeout Delta for an adapter ⢠Configure an iSCSI VLAN For additional information about the HP A ccelerated iSCSI feature, refer to the HP Accelerated iSCSI for Multifunction Network Adapters User Guide located on the HP website ( http://www.hp.com ). Accelerated iSCSI for embedded HP Multifunction server adapters The HP Accelerated iSCSI capability is available for embedded HP Multifunction server adapters through the purchase and installation of a ProLiant Essent ials Accelerated iSCSI Pack license. The ProLiant Essentials Accelerated iSCSI Pack li cense is part of the ProLiant Ess entials software family and is integrated with HP Systems Insight Manager (SIM), Sm artStart, and HP Management agents. It is included with the networking software (releas e 8.35 and higher) found on the HP website ( http://www.hp.com ). NOTE: The HP Accelerated iSCSI feature comes st andard with stand-alone HP Multifunction Server Adapters and Multifunction Se rver Adapters for HP BladeSystem. The Accelerated iSCSI feature runs under Wind ows® 2000, Windows Server 2003, and Windows Server 2003 x64 and allows you to complete the following: ⢠Enable an iSCSI device ⢠Adjust the SRB Timeout Delta for an adapter ⢠Configure an iSCSI VLAN for a Multif unction adapter For additional information about the HP A ccelerated iSCSI feature, refer to the HP Accelerated iSCSI for Multifunction Network Adapters User Guide located on the HP website ( http://www.hp.com ). For information on how to purchase HP ProLiant Essentials value pack li censes for networking, go to the HP website ( http://www.hp.com/servers/proliantessentials ) or contact your HP reseller.
Diagnostics 154 Di agnos ti c s In this section PCI shared interrupts .......................................................................................................... ................... 154 HP NC-Series Intel adapter diag nostics ......................................................................................... .......... 154 HP NC-Series Broadcom adapter diag nostics ...................................................................................... .... 156 HP NC-Series Broadcom Multifunction adapter di agnostics ....................................................................... 1 67 PCI shared interrupts Shared interrupts are normal in PC I servers, but not all PCI device dr ivers support them. Although these adapters support shared interrupts, other PCI drivers used by your server may not. If you are having problems with a shared interrupt, try one of the following: ⢠Use the BIOS setup program for your ser ver to ch ange the interrupt assign ed to this adapter. ⢠Be sure that the adapter device driver loads AFTER the driver or any other device sharing this interrupt. ⢠Contact the manufacturer of the other adapter or add -in card that is sharing this interr upt for a new driver. HP NC-Series Intel adapter diagnostics Overview The diagnostic software lets you test adapters to see i f there are any problems with the adapter hardware, cabling, or network connections. HP recommends that you run diagnostics every time you in stall an adapter. Prerequisites The diagnostic utility runs under MS-DOS® v6.22 or later. It will not run from a DOS window in IBM OS/2® or any Microsoft Windows prod ucts. This util ity is designed to test hardware operation and confirm the adapter's ability to communicate with an other adapter in the same network, which primarily confirms cabling status. This diagostic utility i s not a throughput measurement tool. The utility can test the adapter wheth er or not there is a responder on the network; however, to co nduct a thorough test you should set up a second system on the network as a responder prior to starting a test. Starting the DIAGS.exe test s NOTE: If there is a DOS network driver present, such as NDIS2 or DOS-ODI, the t est utility and/or the network driver could become unstable. You should reboot and ensure that there are no network drivers loaded.
Diagnostics 155 NOTE: The DOS diagnostic utility, DIAGS.EXE, wi ll always use auto-negotiation to determine the link type listed under the "View Adapter Co nfiguration" section. If the network device's port to which your adapter is connected is set to auto-negotiate, then the link type will reflect an auto-negotiated value. If the network device's port has its dupl ex value forced to full, then the link type will show a half-duplex value. 1. If you have not already done so, download and extract the HP SoftPaq file and then run MAKEDISK.BAT to create driver diskettes. 2. Boot to DOS or the EFI shell. 3. From the DOS prompt navigate to the \APPS\DIAGS\N100X directory. 4. Type DIAGS and press the Enter key. The test utility program automatically scans the hardware and lists all NC110X, NC310F, NC340x, NC36 0x, NC364X, NC31xx, NC61xx, and NC71xx adapters. If you have a multiport adapter, each port is listed separately starting with "Port 1." 5. Using the arrow keys, select the adapter to test then press the Enter key. The main menu of the diagnostic utility appears. Viewing adapter configuration Selecting the View Adapter Configuration option displays the adapter configuration screen, which lists the adapter properties such as MAC address, media type, bus number and so on. 1. Press the F5 key to view additional information on th e PCI slot occupied by the adapter. This is primarily used for troublesh ooting by technical support. 2. Press the Enter key twice to return to the main menu. Accessing the test adapter menu Selecting Test Adapter from the main menu displays the test a dapter menu. The test me nu allows the user to begin the adapter tests, change test options, di splay diagnostic log files, and access the networking submenu. Beginning adapter tests Selecting Begin Adapter Tests brings up the test scr een. While tests are being performed, a rota ting spinner is shown letting the user know that the applic ation is still running. The result s of the tests are displayed as each test is performed. If multiple test passes are selected, then the results contain a count of test failures. A list containin g zeros means that all tests have passed. If there is no responder on the network, then the Ne twork Test will indicate a failure. To correct this situation, set up a system on the network to act as a responder then re-run the test. Changing test options The test setup screen allows you to select and conf igure the specific t ests desired. Toggle betwee n options by using the arrow keys and pressing the Enter key. The utility displays only those test options that are supported by your adapter. The following is a list of the available test options. ⢠Number of test passes âThis option determines how many times a test runs. While the test is running, the information on the screen is continuously updated. When you pres s the Esc key, the test is cancelled and control is returned to the Test Adapter menu. ⢠Diagnostic Log âThis test is disabled by default. If enab led, the program will ask for a directory for the log file. The file it places here is named DIAGS.LOG.
Diagnostics 156 ⢠Device Registers âTest patterns are written, read, and verified through the adapter's device registers to check proper functionality. ⢠FIFO âThis option writes test bit patterns to the a dapter's FIFO buffer s to check the FIFOs proper functionality. ⢠EEPROM âThis option tests both the readability of th e EEPROM as well as the integrity of the data stored in the EEPROM. It reads EEPROM and calculates the checksum. This checksum is then compared to the checksum stored in the EEPROM. If the values are not identical, the test r eports failure. ⢠Interrupt âThis option tests the adapter's ability to ge nerate an interrupt and have it propagated through the system to the Programmable Interrupt Contr oller (PIC). The test triggers an interrupt by setting the interrupt cause register and then verifies that an interrupt has been triggered. On EFI, interrupts are not supported , Therefore, this test sets the interrupt cause regist er and reads the interrupt set register, which verifies the c a rd internally registered an interrupt. ⢠Loopback Tests âThese options are internal loopback test s. These tests set the adapter in the appropriate loopback mode and send packets back through the adapter's receive circuitry and logic. ⢠Link âThis option checks to see whether or not the adapter has link. ⢠Network Test âThis option tests netwo rk communication. It looks for a responder a nd then sends packets. If no responder is found, then the test re ports failure. If packets are received back from the responder, the test reports succe ss. Displaying Diagnostic Log The Display Diagnostic Log option allow s you to view a detailed report of the tests you just ran. When Diagnostic Log is enabled, test re sults are recorded in a log file named DIAGS.LOG. If it does not already exist, the test utility creates it. If it already exist s , new data is appended to it. The DISPLAY DIAGNOSTIC LOG command displays the contents of the log file for your convenience. Each entry in the log fil e is time stamped. The test run banner identifies the test ed adapter according to its bus slot address. Accessing the networking submenu The networking submenu allows you to setup the adapter as a responder and to de tect a spanning tree on the network. ⢠Setting up as a responder âThis allows the user to set up th e adapter as a responder so another system can perform the continuous network test. Sele cting this option displays the tr ansmit/receive screen. This test will fail if the adapter does no t have link. Although you can use a variety of adapters as responders, and either connect directly (with a crossover cable) or through a switch, ideal results are obtained with a same-type adapter. When you press the Esc key, the responder operation is cance lled and control is immediately returned to the Test Adapter menu. ⢠Detecting spanning tree âThis allows the user to detect if a spanning tree is used on the network. HP NC-Series Broadcom adapter diagnostics Overview The diagnostic software lets you te st adapters to see if there are any problems with the adapter hardware, cabling, or network connections. HP recommends that you run diagnostics every time you in stall an adapter.
Diagnostics 157 Prerequisites The diagnostic utility runs under MS-DOS® v6.22 or later. It will not run from a DOS window in IBM OS/2® or any Microsoft Windows prod ucts. This util ity is designed to test hardware operation and confirm the adapter's ability to communicate with an other adapter in the same network, which primarily confirms cabling status. This diagostic utility is not a throughput measurement tool. The utility can test the adapter whether or not there is a responder on the network; however, to co nduct a thorough test you should set up a second system on the network as a responder prior to starting a test. These diagnostics use pmode/w DOS Extender V1.33. ⢠Input file list: The following files sho uld be foun d in the same location as the q57diag.exe file. o 5701.bin o 5703.bin o config.sys o cpu05.bin o cpu.bin o cpu14a.bin o cpu14b.bin o cpudg05.bin o cpudiag.bin o cpusj.bin o cpumem.bin o diagcfg.bin o flshdg05.bin o flshdg5x.bin o flshdiag.bin o flshd14a.bin o flshd14b.bin o flshdgsj.bin o q57diag.exe o ump14a.bin o ump14b.bin Do not use emm386.exe. Himem.sy s and emm386.exe are both memory managers. If they are both loaded, the diags aborts and the following error may occur: Exception: OE Error Code: 00000000 About the Q57DIAG.exe diagnostic tests The tests are divided into four group s: Register tests, Memory test s, Miscellaneou s tests, and Data tests. They are identified as group A, B, C, and D. Group A: Register tests ⢠A1. Indirect register test
Diagnostics 158 Using indirect addressing method, this test write s increment data into the MAC Hash Register table and reads back for verification. Th e memory read/write is done 100 times while incrementin g test data. ⢠A2. Control register test Each register specified in the configurati on contents are defined as read only bit and read/write bit. The test writes zero and one into the test bits to ensure the read only bits are not changed and read/write bits are changed accordingly. ⢠A3. Interrupt test This test verifies the interrupt functionality. It enables interrupt and then waits for the interrupt to occur. It waits for 500ms and reports an erro r if it could not generate interrupts. ⢠A4. Built-In-Self test This test initiates Hardware Built-In-S elf-Test (BIS T) and then waits for the test result returned by hardware. The hardware could not generate interrupts. ⢠A5. PCI Cfg register test This test verifies the access inte grity of the PCI config register s. Group B: Memory tests ⢠B1. Scratch pad test This tests the scratch pad SRAM on boar d. The following tests are performed: o Address test: Writes each address with unique increment data. Reads back data t o ensure data is correct. After filling the entire data with th e unique data, the program reads back data again to ensure data stays the same. o Walking one bit test: For each address, data one is written and read back for testing. Then the data is shifted left one bit, so the data becomes two and the same test is run again. It repeats for 32 times until the test bit is shifted out of test da ta. The same test is repeated for the entire test range. o Pseudo random data test: A pre-calculated pseudo random data is used to write a unique data into each test RAM. After the first pass of th e test, the program reads back one more time to ensure data stays correct. ⢠B2. BD SRAM test This tests the BD SRAM by performing th e tests as described in test B1. The Scratch pad test. ⢠B3. DMA SRAM test This tests DMA SRAM by perfo rming the tests described in test B1. The Scratch pad test. ⢠B4. MBUF SRAM test This tests DMA SRAM by perfo rming the tests described in test B1. The Scratch pad test. ⢠B5. MBUF SRAM via DMA test Eight test pattern data are used in the test. They are described below. A 0x 1000 sized data buffer is used for this test. Before each pattern test, the buff er is initialized and filled with the test pattern . It then, performs size 0x1000 transmit DMA from host buffer to adapter MBUF memory. It verifies the data integrity in MBUF against host memory and repe ats the DMA for the entire MBUF buff er. Then it performs receive DMA from adapter to host. The 0x 1000-byte test buffer is cleared to zero before each receive-DMA. It verifies the data integrity an d the test is repeated for the enti re MBUF SRAM range. Test Pattern Description:
Diagnostics 159 "16 00's 16 FF's" Fill the entire host DMA buffer with 16 bytes of 00's and then 16 bytes of FF's. "16 FF's 16 00's" Fill the entire host DMA buffer wi th 16 bytes of FF's and then 16 bytes of 00's. "32 00's 32 FF's" Fill the entire host DMA buffer with 32 bytes of 00's and then 32 bytes of FF's. "32 FF's 32 00's" Fill the entire host DMA buffer wi th 32 bytes of FF's and then 32 bytes of 00's. "00000000's" Fill the entire host DMA buffer with all zeros. "FFFFFFFF's" Fill the entire host DMA buf fer with all FF's. "AA55AA55's" Fill the ent ire host DMA buffer with da ta 0xAA55AA55. "55AA55AA's" Fill the ent ire host DMA buffer with da ta 0x55AA55AA. Group C: Miscellaneous tests ⢠C1. NVRAM test An increment test data is used in the EEPROM test. It fills the test data into th e test range and reads it back to verify the content. After the test, it fills data with zeros to clear the memory. ⢠C2. CPU test This test opens the file cpu.bin. If the file exists and the content is good, it loads code to the Rx and Tx CPU and verifies CPU execution. ⢠C3. DMA test ⢠This tests both high and low prio rities DMA. It moves data from host memory to adapter SRAM, verifies data, and then moves data back to the host memory again to verify data. ⢠C4. MII test This function is identical to A2. Control Register Test. Each Register specified in the configurati on contents is define d as re ad only bit and read/write bit. The test writes zero and one into the test bits to ensure the read only bits are not changed and read/write bits are changed accordingly. ⢠C5. VPD test The content of VPD is saved first before performing the test. After it is saved, th e test writes on e of the five pattern test data, 0xff, 0xaa, 0x55, incremen t data, or decrement data, into VPD memory. By default, increment data pattern is used. It writes and reads back the data for the entire test range, and then it restores the original content. ⢠C6. ASF test The function of this test is as follows: o Reset test. Sets the reset bit and polls for self-clear ing. Verifies the reset value of th e registers. o Event Mapping Test. Sets SMB_A TTN bit by changing ASF_ATTN LOC bits. Verifies the mapping bits in TX_CPU or RX_CPU event bits. o Counter Test. Clear s WG_TO, HB_TO, PA_T O, PL_TO, RT _TO bits by setting those bits. Makes sure the bits are clear. Clears the Timestamp Coun ter. Writes a value 1 into each of the PL, PA, HB, WG, RT counters. Sets the TSC_EN bit. Polls each PA_TO bit and counts up t o 50 times. Checks if PL_TO gets set at the end of 50 times. Continues to count up to 200 times. Checks if all other TO bits are set and verifies th e Timestamp Counter is incremen ted. ⢠C7. Expansion ROM test This tests the ability to enable/disabl e/access the expansion ROM on the device . Group D: Driver associated tests ⢠D1. Mac loopback test
Diagnostics 160 This is an internal loopback data transmit/receive test. It initializes M AC into internal loopback mode, and transmits 100 packets. The data should be routed back to the receive channel and is received by the receive routi ne, which verifies the in tegrity of data. One Gigabit rate is used for this test. ⢠D2. Phy loopback test This test is the same as D1. Mac L oopback Test ex cept the data is routed back via physical layer device. One Gigabit rate is used for this test. ⢠D5. MII miscellaneous test (copper only) This function tests the auto-polling and phy-interrupt capabilities. Th ese are the functionalities of the phy. ⢠D6. MSI test This tests the Testing Message Interrupt Function to se e if it handles t his interrupt correctly. The default is disabled. Starting the Q57DIAG.exe tests 1. If you have not already done so, download and extract the HP SoftPaq file and then run MAKEDISK.BAT to create a driver disk. 2. Boot to DOS or the EFI shell. 3. From the DOS prompt or shell naviga te to the \APPS\DIAGS\Q57 directory. 4. Type Q57DIAG. exe and press the Enter key. The diagnostic te sts run automatically. Command line options The following are the available options in q57diag. They are case sensitive. Usage: q57diag.exe [options] ⢠-c <num> o Specifies UUT device number When more than one device is in the system, the devices are numb ered starting from zero. For example, if there are three devices det ected, the devices are numbered as 0, 1, and 2. In this case, entering the parameter -c 2 will sele ct the last found device as default UUT. In manufacture testing mode, by default, all device s are tested; however, if this opti on is used, only that selected device is tested. o Example: q57diag.exe -c 2 ⢠-smbaddr <num>- o Specifies the ASF SMBus A ddress ⢠-h o Specifies high resolution (80x50) vide o mode o Example: q57diag.exe -h ⢠-l <file> o Specifies the log file in which all diagnostic ou tput is saved. The default is no lo g file. o Example: q57diag.exe -l mylogfile.txt ⢠-p o Specifies Print on Error
Diagnostics 161 o Example: q57diag.exe -p ⢠-npol o Selects negative link polarity in TBI test o Example: q57diag.exe -npol ⢠-q Selects quick diagnostic mode ⢠w <value> o Enables WOL programming in manufacture mode After a successful manufacturing testing, the prog ram will set WOL to either enable or disable mode. Value = 0 to disable WOL and 1 to enable WOL. When -f is entered, software uses eeprom.bin's content for WOL setting. When -w is entered with -f, -w is ignored. o Example: q57diag.exe -w 1 o To disable/enable WOL for LOM design s, Lan On Motherboard, use the Rom Bios Setup Utility, RBSU. ⢠-x <value> o Enables PXE in manufacture mode After a successful manufacturing testing, the prog ram will set PXE to either enable or disable mode. Value = 0 to disable PXE and 1 to enable PXE. When -f is entered, software uses eeprom.bin content for PXE setting. When -x is entered with -f, -x is ignored. o Example: q57diag.exe -x 0 o To disable/enable PXE for LOM de signs, Lan On Motherboard, use the Rom B ios Setup Utility, RBSU. ⢠-asf <value> o Enables asf in manufacture mode o Value = 0 to disable; Value = 1 to enable ⢠-t <id> Disables test ⢠-T <id> o Enables test A certain test is enabled or disabled by default. The user can overwrite the enabling status by those options. The test ID must start with a letter 'A ', 'B', 'C', or 'D' to indicate the group and be followed by test numbers. Each digit of the numb er represents the sub-test numb er. For example, if the user wants to disable test A1 and A3. The option -t A13 should be entered. If no test numbers are entered, all tests in that group are selected. For the tests no t specif ied, the default setting will be used. o Example: q57diag.exe -t A15BC1 -T C4 -t D2 This disables A1, A5, B1, B2, B3, B4, B5, B6, C1, D2, and enables C4
Diagnostics 162 o Default Settings Enabled Tests: A1. Indirect Register Test A2. Control Register Test A3. Interrupt Test A4. Built In Self Test A5. PCI Cfg Register Test B1. Scratch Pad Test B2. BD SRAM Test B3. DMA SRAM Test B4. MBUF SRAM Test B5. MBUF SRAM via DMA Test C1. NVRAM Test C2. CPU Test C3. DMA Test C4. MII Test C5. VPD Test C6. ASF Test C7. Expansion ROM Test D1. Mac Loopback Test D2. Phy Loopback Test D4. MII Miscellaneous Test (Copper only) Disabled Tests: B6. External SRAM Test D3. External Loopback Test ⢠-I <num> o Use this option to specify the number (iteration) of times the tests ar e to be run. The default is to run one time. A number zero indicates loop fore ver. A control-C or control-break key can be used to break the loop. Any error detected will also stop testing after reporting the erro r. o Example: q57diag.exe -I 5 Runs tests five times ⢠-ver o Displays the version of the current test software/eeprom .bin o If this option is entered, it displays the softwa re version number/silkscreen r evision and then exits the program. o Example: q57diag.exe -ver ⢠-bus <num> o Tests UUT location o If this option is specified, the program will test all UUTs on the specified bus number. This option is ignored if -c option is entered.
Diagnostics 163 o Example: q57diag.exe -bus 5 Program tests UUTs on bus 5 ⢠-dpmi o Specifies the use of DPMI memory alloc ation o Uses DPMI memory allocation method to allo cate memory instead of malloc() or free(). ⢠-dir o Displays the version of the IPMI (Inte lligent Platform Ma nagement Interface) ⢠-errctrl <c> Specifies action on error -> a:abort , w:wait, l:loop, c:continue, s:skip ⢠-com <value> Enables com port value (1..4) ⢠-ems <size> Specifies external memory size in HEX to tst ⢠-ref Runs test with reference device ⢠-findref Detects reference device ⢠-fmac Enables input of Mac address from a file ⢠-lbm <n> Sets Mac loopback packets ⢠-lbp <n> Sets Phy loopback packets ⢠-lbe <n:n:n> o Sets external loopback packets o Format: <1000Mbps:100Mbps:10Mbps> ⢠-mba <value> o Enables/disables MBA o Value = 0 to disable; Value = 1 to enable ⢠-mbap <value> o Selects MBA protocol o Value = 0 for PXE; Value = 1 for RPL; Value = 2 for BootP ⢠-mbas <value> o Selects MBA speed o Value = 0 for auto; Value = 1 for 10HD; Value = 2 for 10FD; Value = 3 for 100H D; Value = 4 for 100FD; Value = 6 for 1000FD (fiber ) ⢠-firm <file> Updates device eeprom based on <file> image match
Diagnostics 164 ⢠-firmall <file> Updates device eeprom based on <file> ; Only MAC address will be preserved ⢠-pxe Programs PXE firmware from file ⢠-elog <file> Produces a log file with only error information ⢠-dids Shows PCI DID, VID, SDID, and SVID of present devices ⢠-pasf <file> o Programs ASF firmware o <file> is a text file that contains ASF bin file names ⢠-pxes <value> o Sets PXE speed o Value = 0 for auto; Value = 1 for 10HD; Value = 2 for 10FD; Value = 3 for 100H D; Value = 4 for 100FD; Value = 6 for 1000FD (fiber ) ⢠-hlb <spd> o Sets device host loopback mode o spd = 10/100/1000 o dflt = 1000 ⢠-tr <pkts> <spd> o Sets Tx and Rx pkts packets o dflt Tx pkts = 1e6 o spd = 10/100/1000 ⢠-lbsped <spd> o Runs D3 loopback test at <spd> o spd = t/h/g => 10/100/1000 ⢠-l200 <m200> Runs D4 loopback test with 200*m200 packets ⢠-dispgbend Runs D4 loopback test and displays packets at end ⢠-disppkts Runs D4loopback test and di splays packet count ⢠-errlimit <lim> Runs D4 loopback test. Fail afte r errors greater than <lim>. ⢠-pktf0 Runs D4 loopback test. Packet Data Alternating F 0. ⢠-rsttm <x> o <x> = 0 dis GRC rst. <x>!=0, post rst dly in ms
Diagnostics 165 o dflt = 1 ⢠-rstpl <x> <y> o <x> = post GRC rst poll interval, dly in ms; dflt = 10 o <y> = post GRC rst poll count; dflt = 0 (disable) ⢠-rstpld Above post reset polling counts display enable. ⢠-addref (x) Added a ref card info of dev <x> to refinfo.txt ⢠-blink <x> Blinks the LED on the selected device ⢠-rf <x> Selects a reference device for test D3 ⢠-sil Used before the -firm and -fir mall options to hide warnings ⢠-exe Executes diags commands wi thout a device present ⢠-updateasfcfg Used with the -firmall option to include updating ASF CFG block ⢠-updateumpcfg Used with the -firmall option to include updating UMP CFG block ⢠-pump <file> Option to program UMP firmware ⢠-updatesecfg Used with the -firmall option to include updating SE CFG block ⢠-mfct <file> Programs a new SSID and SVID based on <file> ⢠-fail2 On failure offsets failed message printo ut ⢠-logfl Log file buffer flush per line ⢠-nosz Do not write the NVRAM with the TPM/ NVRAM size Error codes and messages 1. Got 0xx @ 0xx. Expected 0xx 2. Cannot run test while chip is running 3. Invalid adapter device 4. Read only bit %s got changed afte r writing zero at offset 0x%X
Diagnostics 166 5. Read only bit %s got changed afte r writing 1's at offset 0x%X 6. Read/Write bit %s did not get cleared after writing zero at offset 0x %X 7. Read/Write bit %s did not get set after writing 1's at offset 0x %X 8. BIST failed. 9. Could not generate interrupt 10. Test aborted by user 11. Tx DMA:Got 0xx @ 0xx. Expected 0xx 12. Rx DMA:Got 0xx @ 0xx. Expected 0xx 13. Tx DMA failed 14. Rx DMA failed 15. Data error, got 0xX at 0xX, expected 0xX 16. Second read error, got 0xX at 0xX, expected 0xX 17. Failed writing EEPROM at 0xX 18. Failed reading EEPROM at 0xX 19. EEPROM data error, got 0x08X at 0x04X, expected 0xX 20. Cannot open file %s 21. Invalid CPU image file %s 22. Invalid CPU image size %d 23. Cannot allocate memory 24. Cannot reset CPU 25. Cannot release CPU 26. CPU test failed 27. Invalid Test Address Range Valid adapter address is 0xx-0xx and exclude 0xx-0xx 28. DMA:Got 0xx @ 0xx. Expected 0xx 29. Unsupported PhyId X:X 30. Too many registers specified in the file, max is %d 31. Cannot write to VPD memory 32. VPD data error, got X @ 0x04X, expected X 33. No good link! Check Loopback plug 34. Cannot TX Packet! 35. Requested to Tx %d. Only %d is transmitted 36. Expected %d packets. Only %d good packets are received %d unknown packets have been received %d bad packets have been received 37. %c%d is an invalid Test 38. EEPROM checksum error 39. Error in reading WOL/PXE 40. Error in writing WOL/PXE
Diagnostics 167 41. No external memory detected 42. DMA buffer X is large, size must be less than %0 4X 43. File size %d is too big, max is %d 44. Invalid %s 45. Failed writing 0x%x to 0x%x 46. *1 47. *1 48. *1 49. *1 50. Cannot perform task while chip is not running. (need driver) 51. Cannot open register define file or content is bad 52. ASF Reset bit did not self-cleared 53. ATTN_LOC %d cannot be mappe d to %cX CPU event bit %d 54. %s Register is not cleared to zero after reset 55. Cannot start Register Timer 56. poll_ASF bit did not get re set after acknowledged 57. Timestamp Counter is not counting 58. %s Timer is not working 59. Cannot clear bit %s in %cX CPU event register 60. Invalid "EEPROM_FILENAME" file size, expected %d but only can read %d bytes 61. Invalid magic value in %s, expected x but found x 62. Invalid manufacture revision, expected %c but found %c 63. Invalid Boot Code revision, expected %d.%d but found %d.%d 64. Cannot write to EEPROM 65. Cannot read from EEPROM 66. Invalid Checksum 67. Invalid Magic Value 68. Invalid MAC address, expected X-X-X -X-X-X 69. Slot error, expected an UUT to be found at location X:X:00 70. Adjacent memory has been corrupted while testing block 0xx-0xx Got 0xx @ address 0xx. Expected 0xx *1 Internal Use. Program will not generate this error. HP NC-Series Broadcom Multifunction adapter diagnostics Overview
Diagnostics 168 The diagnostic software lets you te st adapters to see if there are any problems with the adapter hardware, cabling, or network connections. HP recommends that you run diagnostics every time you in stall an adapter. Prerequisites The diagnostic utility runs under MS-DOS® v6.22 or later. It will not run from a DOS window in IBM OS/2® or any Microsoft Windows prod ucts. This util ity is designed to test hardware operation and confirm the adapter's ability to communicate with an other adapter in the same network, which primarily confirms cabling status. This diagostic utility i s not a throughput measurement tool. The utility can test the adapter wheth er or not there is a responder on the network; however, to co nduct a thorough test you should set up a second system on the network as a responder prior to starting a test. The DOS extender PMODE/W is embedded into the e xecutable. About the XDIAG.exe diagnostic tests The xdiag.exe diagnostic tests are di vid ed into four groups: Group A: Register tests; Group B: Memory tests; Group C: Block tests; an Group D: Miscellaneous tests. Group A: Register tests ⢠A1. Register test This tests the ch ip registers, accessible th rough PCI/PCI-E bus, for their read-only and read/write attributes. Some critical register s are not tested as the system and/or the chip becomes unstable when the values change. ⢠A2. PCI configuration test This test checks the functionality of the BAR size configuration by examining the BAR value as the BAR size varies. ⢠A3. Interrupt test This test checks to see if the system (OS) receive s the interrupt artificially generated by the chip and if the software ISR is properly invoked. ⢠A4. Not used ⢠A5. MSI test This test checks for the correct behavior of the MSI, making sure no interrupt is generated other than the message. It also runs the negative test to ma ke sure no message is generated when interrupt is masked off. ⢠A6. Memory BIST This tests all memory modules inside the chip using B uilt-In-Self-Test (BIST). ⢠A7. Network link test This tests the external link connection. For the fibe r medium, this is simply another external loopback test. For the copper medium, this is not applicable. Group B: Memory tests Various patterns (0x55aa55aa, 0xaa55a a55, & address) are used to te st each of the memory blocks. ⢠B1. TXP scratchpad ⢠B2. TPAT scratchpad
Diagnostics 169 ⢠B3. RXP scratchpad ⢠B4. COM scratchpad ⢠B5. CP scratchpad ⢠B6. MCP scratchpad ⢠B7. TAS header buffer ⢠B8. TAS payload buffer ⢠B9. RBUF via GRC ⢠B10. RBUF via indirect access ⢠B11. RBUF Cluster list ⢠B12. TSCH list ⢠B13. CSCH List ⢠B14. RV2P scratchpads ⢠B15. TBDC memory ⢠B16. RBDC memory ⢠B17. CTX page table ⢠B18. CTX memory Group C: Block tests ⢠C1. CPU logic and DMA interface tests The tests check the basic logic functionalities of ea ch of the on-chip CPUs. The tests also c over the DMA interface exposed to the CPUs. These tests r equire the presence of a test fi rmware file inside the âdiagfwâ directory. ⢠C2. RBUF allocation test This tests the Rx buffer allocation interface. ⢠C3. CAM access test This tests read, write, add, modify, and cache hit functionalities of the associative memory. ⢠C4. TPAT cracker test This tests the packet cracking logic block as well as the checksum/CRC offl oad logic. This test requires the presence of a test firmware file inside the â diagfwâ directory. ⢠C5. FIO register test This is another register test dedicated for register interface only exposed to the internal CPUs. This test requires the presence o f the test firmware files in the âdiagfwâ directory. ⢠C6. NVM access and reset-corruption tests This tests the non-volatile memory access (both re ad and write). It also tests for appropriate access arbitration among multiple entities (CPUs). Another test is to check to issue chip reset while NVM block is servicing data to look for any NVM data corruption. This test requires the presence of a test firmware file inside the âdiagfwâ directo ry. ⢠C7. Core-reset integrity test
Diagnostics 170 This test issues multiple chip resets and constant driver lo ad/unload to check for boot ROM appropriately loads the bootcode. This test requires a proper bootcode to be programmed into the non-volatile memory. ⢠C8. DMA engine test This tests the internal DMA engine by performing both DMA reads and writes at various location and various sizes. CRC check is performed to ensure data integrity. DMA write test also checks to ensure that the DMA writes do not corrupt the neighboring host memory. This test re quires the presence of test firmware files inside the âdiagfwâ direc tory. ⢠C9. VPD test This tests the VPD interface used by the bootcode . This test requires a proper bootcode to be programmed into the non-volatile memory. ⢠C10. Not used ⢠C11. FIO Events test This test checks for the event bits in the CPUâs Fa st IO interface, making sure appropriate bits are triggered when a particular event occurs (GP IO bit changes, NVM access, and so on). Group D: Miscellaneous tests ⢠D1. MAC loopback test This test puts the chip in the MAC loopback mode and transmits 5000 layer two packets of various sizes and receives them and checks the packet integrity. ⢠D2. PHY loopback test This test puts the chip in the PHY loopback mode and tr ansmits 5000 layer two packets of various sizes and receives them and checks the packet integrity. ⢠D3. External loopback test This test puts the chip in the PHY loopback mode and transmits various number of layer two packets of various sizes and receives th em and checks the packet integrity. The number of packets is determined by the speed. For 10Base-T, only 1000 packets are used; for 100Base-T, 5000; for gigabit traffic, 20000 packets are used. This test re quires an external loopback terminator to t he traffic to be returned, and is turned off by default. ⢠D4. LSO test This test checks the functionality of the large send offload by submit big TCP packets to the chip and expects the chip to segment them into multiple sm aller TCP packets (based on the MSS). The packets are returned in the MAC loopback mode and their integrity is check ed upon receive. ⢠D5. EMAC statistics test This test checks the basic statistics information maintained by the chip by sending and receiving packets of various sizes. ⢠D6. RPC test (D06) This test checks the receive catch-up path by sendin g packets to a different Tx chai n. The packets will go through the RPC logic and return to th e receive buffers as Rx packets. The integrity of each packet is checked to ensure no data corruption. Starting the XDIAG.exe tests
Diagnostics 171 1. If you have not already done so, download and extract the HP SoftPaq file and then run MAKEDISK.BAT to create a driver disk. 2. Boot to DOS or the EFI shell. 3. From the DOS prompt navigate to the \APPS\DIAGS\MFDIAG directory. 4. Type XDIAG and press the Enter key. The diagnostic tests r un automatically. Command line options The following are the available options in xdiag. They are case sensitive. Usage: xdiag [options] ⢠-t <grps/tests> Disables certain tests/grou ps (for example: a1c2) ⢠-T <grps/tests> Enables certain tests/groups (for example: a1c2) ⢠-c <card#> Selects device number on wh ich tests will be running ⢠-I <iteration#> Specifies how many iterat ions tests need to run ⢠-wol <1|0> Enables (1) or disables (0) magic pkt wol ⢠-pxe <1|0> Enables (1) or disables (0) pxe ⢠-asf <1|0> Enables (1) or disables (0) asf ⢠-cof Allow tests to continue test s on failure ⢠-ver Displays information on devices and xdiag version ⢠-log <logfile> Logs the tests execution into the specified file ⢠-help Prints out the screen
Technical support 172 T ec hni cal suppo r t In this section Before you contact HP.......................................................................................................... ................. 172 HP contact in format ion ......................................................................................................... ................. 172 Free automated custom er support services ....................................................................................... ........ 172 Before you contact HP Be sure to have the following info rmation available before you call HP: ⢠Technical support registrati on number (if applicable) ⢠Product serial number ⢠Product model name and number ⢠Applicable error messages ⢠Add-on boards or hardware ⢠Third-party hardware or software ⢠Operating system type and revision level HP contact information For the name of the nearest HP authorize d reseller: ⢠In the United States, see the HP US service locator webpage ( http://www.hp.com/service_l ocator ). ⢠In other locations, see the Contac t HP worldwide (in English) webpa ge ( http://welcome.hp.com/country/us/en/wwcontact.html ). For HP technical support: ⢠In the United States, for contact options see the Contact HP United States webpage ( http://welcome.hp.c om/country/us/en/contact_us. html ). To contact HP by phone: o Call 1-800-HP-INVENT (1-800-474 -6836). This service is available 24 hours a day, 7 days a week. For continuous quality improvemen t, calls may be recorded or monitored. o If you have purchased a Care Pack (servi ce upgrade), call 1-800-633-3600. For more information about Care Packs, refer to the HP website ( http://www.hp.com ). ⢠In other locations, see the Contac t HP worldwide (in English) webpa ge ( http://welcome.hp.com/country/us/en/wwcontact.html ). Free automated customer support services Free Automated Customer Support Services
Technical support 173 The following sites offer tro ubleshooting information, compatibility notes, and softw are upgrades (including Softpaqs and drivers). HP Worldwide Web Server ⢠Navigate to a specific product, and then look for support information from the list of support resources at the HP support websit e ( http://h18007.www1.hp.com/support/files/serv er ). ⢠For downloadable support software for HP Digital Networking Products, Hu bs, Integrated Access Devices, Modems and ISDN, Adap ters, Remote Access Concentrators/Servers, Software, and Switches, go to the HP software and drivers website ( http://h18007.www1.hp.com /support/files/server ). ⢠All SoftPaqs sorted by SoftPaq number can be found at the HP ftp support website ( ftp://ftp.compaq.com/pub/softpaq/ ). An ASCII version of a SoftPaq can be found by selecting a SoftPaq at the HP ftp support website. ( ftp://ftp.compaq.com/pub/softpaq/ ) ⢠An index of available software sorted by produc t can be found at the HP software and drivers website ( http://h18007.www1.hp.com/support/files/serv er ). HP FTP Server Navigate to a specific product, and then look for supp ort information from the list of support resources at the HP ftp support website ( ftp://ftp.compaq.com/pub/softpaq/ ).
Acronyms and abbreviations 174 A c r on y ms and abbr e v iati ons BIOS Basic Input/Output System DHCP Dynamic Host Configuration Protocol IPL initial program load iSCSI Internet Small Computer System Interface PCI peripheral component interface PDU power distribution unit PXE Preboot Execution Environment ROM read-only memory RSS Receive-Side Scaling SAN storage area network TCP/IP Transmission Control Protocol/Intern et Protocol TOE TCP/IP Offload Engine
Acronyms and abbreviations 175 UNDI Universal Network Driver Interface VLAN virtual local-area network
Index 176 A Adapter teaming 129, 139 Advanced Server Program for NetWare 110 B Boot Agent 69 D diagnostics 155, 157 Diver and software information 7 DXE configuration NC37xx and NC380x 67 F Firmware upgrade 48 H HP ProLiant Essentials value packs for networking 151 I installation 13, 15, 20, 21, 23, 28 K Keywords for B06 NDIS2 driver 63 Keywords for N100 NDIS2 driver 81 Keywords for N1000 NDIS2 driver 87 Keywords for Q57 NDIS2 driver 62 L Linux installation 20 N NCU properties 91 NetWare Client 32 driver installation for HP adapters 120 NetWare installation 28 NetWare teaming 139 P PCI Hot Plug 9 PCI shared interrupts 155 properties 91, 98, 106 PXE configuration HP NC-series Broadcom adapters 64, 67 HP NC-series Intel adapters 89 S Solaris driver information guide 23 supported operating system s 5 T teaming adapters 126 U Unix driver software Gigabit Ethernet Adapter user's guide 21 W Windows 2000 driver software NC150, NC32x, NC37x, NC380, NC1020, NC6770, and NC77xx Gigabit Ethernet Adapter user's guide 98 Windows 2000 installation notes 15 Windows Server 2003 NC-series Intel adapter properties 103 Windows Server 2003 and Windows Server 2003 x64 installation notes 13 Windows Server 2003 NC31xx Fast Eth ernet Adapter user's guide 106 Windows Server 2003 NC -s eries Broadcom adapter properties 98 Inde x