HP Hewlett Packard Network Card 441877 00B User Manual

HP Network Adapter  
Software and Configuration Guide  
Part Number 441877-00B  
March 2007 (Second Edition)  
 
Contents  
Contents  
3
 
Contents  
4
 
Overview  
In this section  
Supported adapters and operating systems ................................................................................................. 5  
PCI Hot Plug ............................................................................................................................................ 9  
Supported adapters and operating systems  
Supported adapters  
The Network Configuration Utility (NCU) supports the 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 Adapter  
HP NC364T PCI Express Quad Port Gigabit Server Adapter  
HP NC370T PCI-X Multifunction Gigabit Server Adapter  
HP NC370F PCI-X Multifunction Gigabit Server Adapter  
HP NC373F PCI Express Multifunction Gigabit Server Adapter  
HP NC373T PCI Express Multifunction Gigabit Server Adapter  
HP NC380T PCI Express Dual Port Multifunction Gigabit Server Adapter  
HP NC1020 Cu Gigabit Server Adapter 32 PCI Single Port  
HP NC3123 Fast Ethernet Server Adapter *  
HP NC3132 Fast Ethernet Upgrade Module (Dual 10/100) for the NC3134 Server Adapter *  
HP NC3133 100FX Upgrade Module for the NC3134 Server Adapter *  
HP NC3134 Fast Ethernet Server Adapter (Dual Base 10/100) *  
HP NC3135 Fast Ethernet Upgrade Module (Dual 10/100) for the NC3134 Server Adapter *  
HP NC6132 1000 SX Gigabit Upgrade Module for the NC3134 Server Adapter*  
HP NC6133 1000 LX Gigabit Upgrade Module for the NC3134 Server Adapter *  
HP NC6134 Gigabit Server Adapter (PCI 1000 SX)  
Overview  
5
 
     
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 NC3134 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 Server Adapter  
HP NC371i PCI-X Multifunction Gigabit Server Adapter  
HP Embedded NC373i Multifunction 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 Adapter for c-Class BladeSystem  
HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem  
HP NC373m PCI Express Dual Port Multifunction Gigabit Server Adapter for c-Class BladeSystem  
HP NC374m PCI Express Dual Port Multifunction Gigabit Server Adapter  
* Not supported in Windows Server 2003 x64  
Supported operating systems  
These server adapters are supported on the following operating systems.  
Microsoft  
Windows Server 2003 x64 Editions  
Windows Server 2003  
Overview  
6
 
Windows® 2000 (NDIS5)  
Windows NT® 4.0 (NDIS4) (only supported on adapters that shipped before 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 AMD64 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 NDIS2 is not supported for NC3133 adapters.  
Adapter driver and software information  
To obtain driver and software information for Microsoft® 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.  
2.  
3.  
4.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Overview  
7
 
   
5.  
6.  
Click HP ProLiant Networking Software.  
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.  
Click the SoftPaq file to extract the files and then open the cmponent.htm file.  
Drivers  
Drivers are found in the following folders:  
Microsoft® Windows Server 2003 \WIN2003  
Microsoft® Windows Server 2003 x64 \WIN2003x64  
Microsoft® Windows® 2000 \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.exe 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 list of all files in the SoftPaq.  
Release notes and other documentation in \docs  
HP Accelerated iSCSI for Multifunction Network Adapters (AiSCSIUG.pdf)  
HP Network Adapter License Utility for Windows (NALICNSE.pdf)  
HP Network Adapter Scripting Utility (NICSCRPT.pdf)  
HP Network Configuration Agent (ncuasr.pdf)  
HP ProLiant Essentials Intelligent Networking Pack–Linux Edition (PEINPL.pdf)  
HP ProLiant Essentials Intelligent networking Pack–Windows Edition (PEINPW.pdf)  
HP Virus Throttle for Windows License Utility (VTLICNSE.pdf)  
HP NC510x PCIe 10 Gb Software Configuration Guide (NC510x PCIe 10 Gb Software Guide.pdf)  
Network adapter user guides (\docs\hw)  
Release notes (Relnotes.txt)  
SetLACState Utility (SetLAC.pdf)  
Overview  
8
 
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-critical computing environments. PCI Hot Plug enables PCI networking adapters  
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. PCI Hot Plug technology also has a broad range of  
industry support from leading operating system suppliers including Microsoft® and 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 supports PCI Hot Plug. To implement PCI Hot Plug, use the  
configuration instructions that come with your system. If you use an HP system that supports PCI Hot Plug,  
refer to the documentation for configuring and installing 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 documentation that came with your  
system or contact technical support for the vendor.  
Support for enabling PCI Hot Plug  
Different operating systems support PCI Hot Plug in different ways. The following list describes how  
different operating systems enable PCI Hot Plug devices.  
Windows® 2000, Windows Server 2003, and Windows Server 2003 x64  
o
o
o
Can replace like item for like item  
Can add a new adapter to an empty slot  
Can add an option module to an existing NC3134 adapter  
Windows NT® 4.0  
o
o
Can replace like item for like item  
Cannot add a new adapter to an empty slot  
NetWare 6.x  
o
o
o
o
Can replace like item for like item  
Can remove an adapter  
Can add a new adapter to an empty slot  
Can add an option module to an existing NC3134 adapter (This feature is not supported on an  
NC3134 adapter that has a persistent driver installed.)  
UnixWare 7.x and OpenServer 6.x  
Requires hardware that implements PCI Hot Plug failover and a DDI 8 MDI driver that implements the  
CFG_SUSPEND and CFG_RESUME subfunctions to the 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 without bringing the machine down. For further information  
about PCI Hot Plug support, refer to your UNIX documentation.  
SCO OpenServer 5.x  
Overview  
9
 
   
No support  
Linux  
No support  
Upgrading the adapter with a persistent driver  
To upgrade the adapter with a persistent driver:  
1.  
2.  
3.  
4.  
5.  
Manually remove (unload) the driver.  
Power down the slot and remove the adapter.  
Upgrade the adapter (add the option module).  
Reinstall the upgraded adapter.  
Power up the slot to let NetWare automatically detect the (upgraded) 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-duplex switch to send and receive  
packets at the highest rate. The switch must support auto-negotiation. If the switch does not support  
auto-negotiation, the adapter defaults to half-duplex.  
NOTE: An exception is the HP NC3133 adapter, which defaults to full-duplex.  
Full-duplex. (Requires full-duplex switch) The adapter sends and receives packets simultaneously by  
using a different wire to perform each function. This feature can improve adapter performance by  
eliminating collisions. Set duplex mode to full-duplex ONLY if you have a hub/switch that supports  
full-duplex.  
Half-duplex. The adapter performs one operation at a time (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 automatically matches the duplex setting at the switch.  
If you do not have an auto-negotiating switch, you must manually configure the adapter for full-duplex  
support.  
Configuration is specific to the driver you 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 that  
performance may suffer if your switch is not full-duplex and you configure the adapter to full-duplex. Leave  
the adapter on half-duplex if you are not sure what 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 Neighborhood  
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 10  
 
 
3.  
Set Speed/Duplex on the adapter to Auto/Auto, 10/Half, 10/Full, 100/Half, 100/Full, or  
1000/Full.  
4.  
5.  
Click OK until prompted to restart.  
Restart Windows NT® 4.0.  
Windows® 2000, Windows Server 2003, and Windows Server 2003 x64  
For new installations, Windows® 2000 automatically installs the adapter drivers. Exception: For the  
NC150T, NC310F, NC32xx, NC37xx, NC340x, NC1020, NC6136, NC6170, NC71xx and  
NC77xx adapters, you must install the appropriate driver components located in the \WIN2000  
directory in the HP SoftPaq. Complete the following steps to download the SoftPaq file.  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
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  
To set the duplex options:  
1.  
2.  
Launch the Network Configuration Utility.  
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 FORCEDUPLEX and SPEED.  
NOTE: The NC340T, NC1020, NC71xx, and NC77xx adapters support full-duplex and half-  
duplex. The fiber Gigabit adapters support full-duplex mode only and cannot be changed.  
Syntax: FORCEDUPLEX=n  
Where n =  
0–auto-negotiation (You must have an auto-negotiating switch/hub (an Nway* switch) to get full-  
duplex support with the FORCEDUPLEX parameter set to 0 (auto-negotiation).  
1–half-duplex  
2–full-duplex  
NOTE: To set the adapter to half- or full-duplex, you must set the SPEED parameter to either 10  
or 100.  
Default = auto-negotiate  
Overview 11  
 
Examples:  
o
o
100 Mbps full-duplex: FORCEDUPLEX=2 SPEED=100  
10 Mbps full-duplex: FORCEDUPLEX=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 automatically detects  
the network speed. If unable to detect the network speed, the driver defaults to 10 Mbps. If you use  
this parameter, the driver operates at the specified speed instead of auto-detecting network speed.  
NOTE: The NC3133 adapter supports 100 Mbps only. 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) adapters. 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-based clients, edit the PROTOCOL.INI file. Add the  
FORCEDUPLEX and SPEED parameters to the N100 section, or the SPEEDUPLEX parameter to the N1000  
section. For information on these parameters, refer to Keywords for N100 NDIS2 driver (on page 81) ,  
Keywords for N1000 NDIS2 driver (on page 87), or Keywords for Q57 NDIS2 driver (on page 62).  
Overview 12  
 
Installation  
In this section  
Windows Server 2003 and Windows Server 2003 x64 ............................................................................ 13  
Solaris .................................................................................................................................................. 23  
NetWare .............................................................................................................................................. 28  
Download latest drivers, firmware, and  
documentation  
HP updates networking software frequently to include new functionality and features. For the latest driver,  
firmware, and documentation updates go to the HP website  
Windows Server 2003 and Windows Server 2003  
x64  
The following provides information about Windows Server 2003 and Windows Server 2003 x64 support  
for HP NC-series network adapters.  
Windows Server 2003  
The following Windows Server 2003 network adapter drivers are used in this release:  
N100325.sys for NC31xx Fast Ethernet server adapters  
N1000325.sys for NC310xx, NC340x, NC61xx, and NC71xx server adapters  
Q57XP32.sys for NC150x, NC32xx, NC10xx, NC67xx, 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 \WIN2003 directory in the HP SoftPaq. Complete the  
following steps to download the SoftPaq file.  
1.  
2.  
3.  
4.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Installation 13  
 
       
5.  
6.  
Click HP ProLiant Networking Software.  
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.  
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 Server 2003 x64 network adapter drivers are used in this release:  
N1G5132E.sys for NC310F, NC340T, NC61xx, and NC71xx server adapters  
Q57AMD64.sys for NC150T, NC32x, NC1020, NC67xx, and NC77xx server adapters  
BXVBDA.sys for Multifunction server adapters  
N1E5132E.sys for NC110T, NC360T, NC364x server adapters  
NXP2NIC.sys for NC510x 10 Gb server adapters  
Components for these drivers are located in the \WIN2003x64 directory in the HP SoftPaq. Complete  
the following steps to download the SoftPaq file.  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
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 time:  
1.  
2.  
3.  
Connect the cable from the adapter to your network.  
Provide power to your server.  
Start Windows Server 2003 or Windows Server 2003 x64 .  
o
For the NC150T, NC310F, NC32xx, NC340x, NC37xx, NC380x, NC1020, NC6136,  
NC6170, NC6770, NC71xx, and NC77xx server adapters, you must install the appropriate  
driver component located in the \WIN2003 or \WIN2003x64 directory in the HP SoftPaq file.  
o
For all other adapters, Windows® automatically detects 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 drive where you copied it from an  
updated source. See Technical Support for a list of online sources for software and driver updates.  
Installation 14  
 
1.  
Run the appropriate Driver Component Package (listed as CPxxxxxx.exe) in a directory on your hard  
drive. The component is a self-extracting executable with a file name based on the component  
number.  
2.  
3.  
Click the Install button to proceed with the installation.  
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.  
2.  
Before physically removing an adapter from your system, first remove the adapter driver software.  
Start Windows Server 2003 or Windows Server 2003 x64 and log in. You must have Network  
Administrator privileges to remove 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.  
5.  
6.  
Click Device Manager. The Device Manager window appears.  
Click the plus (+) sign to expand the Network adapters. All network adapters appear.  
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 information about Windows 2000 support for HP NC-series network adapters.  
The following Windows 2000 network adapter drivers are used in this release:  
N100NT5.sys for NC31xx Fast Ethernet server 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 Ethernet server adapters  
Components for these drivers are located in the \WIN2000 directory in the HP SoftPaq file.  
Installing network drivers (new installations)  
When you install the adapter in the server for the first time:  
1.  
2.  
3.  
Connect the cable from the adapter to your network.  
Provide power to your server.  
Start Windows Server 2003 or Windows Server 2003 x64 .  
o
For the NC150T, NC310F, NC32xx, NC340x, NC37xx, NC380x, NC1020, NC6136,  
NC6170, NC6770, NC71xx, and NC77xx server adapters, you must install the appropriate  
driver component located in the \WIN2003 or \WIN2003x64 directory in the HP SoftPaq file.  
o
For all other adapters, Windows® automatically detects and installs a driver for the adapter from  
its own library of drivers.  
Updating network drivers from disk  
Installation 15  
 
   
To update the network driver, follow these instructions:  
Locate the driver on the CD, diskette, or directory on your hard drive 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 (listed as CPxxxxxx.exe) in a directory on your hard  
drive. The component is a self-extracting executable with a file name based on the component  
number.  
2.  
3.  
Click the Install button to proceed with the installation.  
After the drivers have been installed, you may delete the Component Package that you copied.  
Windows NT4 unattended install (push)  
The following provides information 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 begins 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 adapters on the Windows NT® 4.0 CD-ROM is restricted and  
special steps must be performed to update the installation source files and modify the scripts to automate  
the process for the adapters not on the list.  
Main steps of the process  
1.  
2.  
3.  
Create a Network Installation Startup disk from the Windows NT® server and modify it for your  
adapter.  
Create a shared directory for the Windows NT® server installation files on the Windows NT®  
server.  
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 parameters for the adapter in the [CPQNic1ParamSection]  
section. The override parameters should not be used unless the default parameters will not work  
properly for a specific installation. The sample parameters are commented out.  
4.  
5.  
Integrate the adapter driver files into the Windows NT® installation source files.  
Perform the Unattended Installation bootup with the 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 shared 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.  
3.  
Use a CD-ROM drive at the Windows NT® server and copy the i386 folder with Windows NT® 4.0  
installation files to the shared folder NT40.INS on this server (for example, using NT Explorer).  
Create or modify the installation answer file UNATTEND.TXT as required by your specific system and  
desired Windows NT® configurations. UNATTEND.TXT may be read-only and may need to be  
made write-able before editing (for example, attrib -r unattend.txt). Refer to the sample  
UNATTEND.TXT file with comments.  
4.  
Create the \CPQANC directory: NT40.INS\i386\DRVLIB.NIC\CPQANC.  
Installation 16  
 
5.  
6.  
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 folder on the  
Windows NT® server). This enables the installation program to find the same file structure as on the  
Drivers disk. Be sure to use the /s /e switches on xcopy.  
On the Windows NT® server, prepare the Microsoft® 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 the diskette as specified in item 3 above  
(preferably to the root (a:\)). The best location to put 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® Administrative Tools (Common), start Network Client Administrator and  
select Make Network Installation Startup Disk.  
d. Set "Existing Path" to use the previously installed software for MS Client or, if it is the first time,  
create the shared directory on the server's hard 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 for MS-DOS and Windows.  
f. Choose 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.DOS 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 N1000$ 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 setup.exe from A:\AUTOEXEC.BAT.  
7.  
Insert the startup disk and boot up your DOS client. After connecting to the network. do one of the  
following:  
o
o
Input USERID as a User name when prompted. Input password as a password when prompted.  
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 17  
 
;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 18  
 
[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 19  
 
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® Windows NT® Resource Kit—Microsoft® 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:  
tg3—Supports NC150x, NC320x, NC324x, NC325x, NC326x, NC1020, NC67xx, and  
NC77xx server adapters  
e1000—Supports NC110T, NC310F, NC340x, NC360x, NC364T, NC61xx, and NC71xx server  
adapters  
hp-pel—Supports ProLiant Essentials Licensing package  
hp-vt—Supports Virus Throttle for NCxxxx adapters  
iSCSI-boot—Supports the HP iSCSI boot feature that allows booting a server over iSCSI  
unm—Supports 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 SoftPaq.  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
Installation 20  
 
   
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  
HP ROMPaqs for specific servers can be found at the HP support website (http://www.hp.com/support).  
The source RPM is suitable for 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 drivers 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  
HP Extended Feature Supplement (EFS) is a collection of device drivers that provide maximum  
performance for HP servers. The EFS also includes several utilities to monitor and increase the  
performance of your HP server.  
SCO OpenServer 5.x  
This procedure describes the installation of the SCO OpenServer driver. This driver is released as a media  
image file containing the driver package. The media image file can be copied to the target machine  
directly for installation or from an installation diskette that you can create.  
Creating a diskette  
1.  
2.  
Copy the file VOL.000.000 to a Caldera system.  
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 install the SCO OpenServer driver from the media image or  
from the diskette created above.  
2.  
3.  
Use netconfig to add the new network adapter.  
Modify the hardware configuration in Advanced Options to change the Line Speed and Flow  
Control if desired. The settings for these parameters are listed below:  
o
Line Speed  
AutoNegotiate (default)  
HalfDuplex10  
FullDuplex10  
HalfDuplex100  
FullDuplex100  
FullDuplex1000  
o
Flow Control  
Off (default)  
AutoNegotiate (Symmetric Pause advertised)  
Installation 21  
 
   
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 parameters, at the command prompt enter:  
man bcme  
Jumbo Frames and other advanced parameters  
Jumbo MTU sizes and other advanced, tunable parameters 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 parameter in space.c, rebuild the kernel, and reboot the system. 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 describes the installation of the UnixWare driver. The driver is released as a  
PKG file.  
Installing the driver locally  
1.  
2.  
Copy the file bcme-<version>.pkg to a UnixWare 7.x system.  
Enter the following command: pkgadd -d <path><filename> where <path> is the directory to which  
the PKG file was copied.  
Creating a diskette  
1.  
2.  
Copy the file bcme-<version>.pkg to a Caldera system.  
Create a diskette using the following command: dd if=bcme-<version>.pkg of=/dev/fd0  
Installing the driver from a diskette  
1.  
2.  
3.  
Use pkgadd -d diskette1 to install the bcme package.  
Use netcfg or scoadmin network to add the new network adapter.  
When prompted, select the Line Speed and then select Advanced Option for Flow Control, MAC  
Address, and Jumbo MTU Size settings, if desired. The settings for these parameters are listed  
below:  
o
Line Speed  
Auto Negotiation (default)  
10 Mbps half-duplex  
10 Mbps full-duplex  
100 Mbps half-duplex  
Installation 22  
 
100 Mbps full-duplex  
1000 Mbps full-duplex  
Flow Control  
o
Disabled (default)  
Auto Negotiation (Symmetric Pause advertised)  
Receive Pause  
Transmit Pause  
Receive & Transmit Pause  
MAC Address  
o
o
No Override (default) - a user-administered MAC address entered with a colon separating each  
hexadecimal byte (for example, 12:34:56:78:9a:bc).  
Jumbo MTU Size  
1500–9000 (default is 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 driver for HP 10/100/1000 Mbps Network Server Adapters. The  
driver is released in two formats:  
BRCMbcme.pkg: Datastream format  
bcmedu.img: Driver Update (DU) binary image which is used to create DU diskette.  
Installing the driver  
1.  
2.  
3.  
4.  
5.  
Change directory to where BRCMbcme.pkg resides.  
pkgadd -d BRCMbcme.pkg  
Execute prtconf to determine instance number of the adapter.  
ifconfig bcme [instance_number] plumb  
ifconfig bcme [instance_number] ip_address netmask ....  
To make these changes permanent:  
1.  
Use your favorite text editor (for example, 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 23  
 
   
In Solaris 7.0 (Intel platform), the operating system only allocates 36 pages of 4K physically contiguous  
memory. The driver needs approximately 130K of physically contiguous memory 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 Solaris 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.  
2.  
3.  
ifconfig bcme [instance_number] down  
ifconfig bcme [instance_number] unplumb  
pkgrm BRCMbcme  
Driver update (DU) diskette instruction (Intel® Platform Only)  
This area contains the Solaris diskette image files, and the instructions to create diskettes from these image  
files.  
1.  
Insert a blank diskette into the diskette drive of your system, and type one of the following commands  
to format it:  
o
o
If you are using DOS, type: format A:  
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 24  
 
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 the Solaris operating system environment 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 drivers 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 Solaris Installation CD-ROM or, for network installation,  
verify with your system administrator that the Solaris network installation image is available on your  
network.  
2.  
3.  
Power on your system.  
When the Configuration Assistant screen is displayed, select the F4 option (on version 2.6, it is  
labeled F4_Driver Update; on Solaris 7, it is labeled F4_Add Driver). The message, "Enumerating  
buses...." is displayed.  
4.  
Remove the Configuration Assistant diskette from the disk drive and insert the first Solaris DU  
diskette.  
5.  
6.  
Select F2_Continue. The Select Solaris System Version screen is displayed.  
Select the appropriate Solaris OS and select F2_Continue.  
The Loading Driver Update Software screen is displayed, along with a progress bar that indicates  
the percentage of drivers that have been extracted from the diskette. Drivers are read into memory  
and survive long enough for the system to successfully boot to its installation program. When all the  
new drivers on the diskette have been processed, the Continue Driver Update Installation screen is  
displayed.  
7.  
8.  
If you have additional DU diskettes to install, remove the DU diskette from the disk drive, insert the  
next DU diskette.  
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 system to successfully boot to its installation program.  
When all the new drivers on the diskette have been processed, the Continue Driver Update  
Installation screen is displayed.  
9.  
Repeat step 7 and 8 until all desired DU diskettes are installed.  
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 Assistant 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 screen is displayed.  
12. Select F2_Continue. The message, "Enumerating buses...." is displayed, followed by the Scanning  
Devices screen. System devices are scanned. When scanning is complete, the Identified Devices  
screen is displayed.  
Installation 25  
 
13. Select F2_Continue. The message "Loading driver...." is displayed, followed by messages about the  
drivers that are required to boot your system. After a few seconds, the Boot Solaris screen is  
displayed.  
14. At the Boot Solaris screen, select the device controller attached to the device that contains your install  
medium.  
15. Select F2_Continue. Drivers display for the device controller you selected. 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 following messages are displayed, prompting you to insert  
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 reinsert the first DU diskette into the disk drive.  
17. Press the Enter key. Packages, patches, or both that contain the new drivers are installed from the  
diskette onto your system. Messages about each installation are displayed. If drivers on other DU  
diskettes are required for your system, the following prompt is displayed.  
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 required, remove the DU diskette from the diskette drive, insert  
the next DU diskette you are prompted to insert. Press the Enter key. Repeat until all the drivers you  
need are installed.  
19. Press the Enter key.  
When installation is complete, the message "Installation complete" is displayed.  
20. Remove the diskette from the disk drive.  
21. Reboot the system.  
When the Solaris operating system is finished booting and running, the new devices whose drivers  
you installed are available for use.  
Adding a DU diskette to an existing Solaris system  
Before adding new or updated drivers, the newly supported hardware devices should be installed and  
configured according to the instructions in the corresponding 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 diskettes as patches on your system, as follows:  
1.  
2.  
Become the superuser on your system.  
Check to see if Volume Management is running on the system you are updating:  
ps -ef | grep vold  
For more information about managing diskette and drives, refer to the System Administration Guide.  
Installation 26  
 
3.  
If Volume Management is running, temporarily stop it by typing:  
# /etc/init.d/volmgt stop  
4.  
5.  
Insert the DU diskette into the disk drive.  
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 update your system  
successfully.  
6.  
Execute the install script on the diskette, using the appropriate Solaris release directory (currently  
sol_26 for Solaris 2.6, sol_27 for Solaris 7, and so on). For example:  
# /mnt/DU/sol_27/i86pc/Tools/install.sh -i  
The install.sh script searches for all new or updated 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 the 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 installs the driver you selected, as well as the bootmod  
and bootbin patches.  
8.  
9.  
When you have finished, and the install.sh script exits, unmount the diskette:  
# cd /  
# umount /mnt  
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 autoboot 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 scanned. 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 device that contains your  
install medium, in this case the main system disk. The /etc/bootrc script is displayed.  
16. At the prompt, type:  
b -r  
Your system boots. You can now use your new hardware.  
Installation 27  
 
NetWare  
A network device driver must be installed before the Gigabit Ethernet adapter can be used with your  
Novell NetWare system. Before you can successfully install the adapter driver for Novell NetWare, the  
adapter card must be physically installed in the server 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 NetWare server, NetWare will automatically detect the new  
adapter and attempt to load the appropriate driver. Ensure that your CD-ROM is mounted and select the  
proper driver.  
To enable the Gigabit Ethernet adapter to function correctly, 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 NWCONFIG. The following  
drivers are supported.  
Driver names: N1000e.LAN, N1000.LAN, and N100.LAN  
HP N1000e.LAN driver requirements  
o
o
o
o
o
o
Use with an HP NC110T, NC360T, and NC364T Gigabit Server Adapter  
Always install the latest Novell support pack  
Always install the latest HP NSSD  
SET MINIMUM PACKET RECEIVE BUFFERS = 2000 (or larger)  
SET MAXIMUM PACKET RECEIVE BUFFERS = 10000 (or larger)  
SET MAXIMUM PHYSICAL RECEIVE PACKET SIZE = 2048 (or larger)  
HP N1000.LAN driver requirements  
o
o
o
o
o
o
Use with any HP NC310F, NC340T, NC61xx, and NC71xx Gigabit Server Adapter  
Always install the latest Novell support pack  
Always install the latest HP NSSD  
SET MINIMUM PACKET RECEIVE BUFFERS = 2000 (or larger)  
SET MAXIMUM PACKET RECEIVE BUFFERS = 10000 (or larger)  
SET MAXIMUM PHYSICAL RECEIVE PACKET SIZE = 2048 (or larger)  
HP N100.LAN driver requirements  
o
o
o
o
o
o
Use with any HP NC31xx Fast Ethernet Server Adapter  
Always install the latest Novell support pack  
Always install the latest HP NSSD  
SET MINIMUM PACKET RECEIVE BUFFERS = 2000 (or larger)  
SET MAXIMUM PACKET RECEIVE BUFFERS = 10000 (or larger)  
SET MAXIMUM PHYSICAL RECEIVE PACKET SIZE = 2048 (or larger)  
Using NWCONFIG or HDETECT to install the adapter  
Installation 28  
 
   
NOTE: Your NetWare server software must be the 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 details 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 MAXIMUM PACKET RECEIVE  
BUFFERS and MINIMUM PACKET RECEIVE BUFFERS are usually too low. The adapter requires a  
minimum of 64 Packet Receive Buffers per adapter port during driver initialization.  
This initial load of the driver may fail because of the two above mentioned settings. The following error  
message is displayed: *** Packet Receive Buffers setting is too low. ***  
To correct:  
1.  
2.  
3.  
4.  
5.  
6.  
7.  
Load MONITOR.  
Select SERVER PARAMETERS.  
Select COMMUNICATIONS.  
Set MAXIMUM PACKET RECEIVE BUFFERS = 10000 (or larger).  
Set MINIMUM PACKET RECEIVE BUFFERS = 2000 (or larger).  
Set MAXIMUM PHYSICAL RECEIVE PACKET SIZE = 2048 (or larger).  
Restart NetWare to complete.  
This can also be set by using the NWCONFIG and selecting the Startup.NCF file. See the Netware  
Installation Guide for specific instructions.  
The maximum and minimum settings of 10000 and 2000 respectively are suggested values only. You  
may need to request a larger number for MAXIMUM PACKET RECEIVE BUFFERS depending on the  
number of supported adapter ports in the system. This message is also displayed on the console if the  
driver cannot allocate the required number of Receive Buffers during driver load.  
Driver configuration parameters  
Parameter  
Options  
Description  
CheckSum =  
Default = ON  
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 CheckSum parameter, you  
need to load it on the first instance.  
Selections are: ON, OFF, Tx,  
Rx  
Frame = type  
node =  
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.  
NNNNNNNNNNNN  
Specifies a node 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 29  
 
Parameter  
Options  
Description  
PDriver =  
Default = OFF  
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.  
Selections are: OFF, ON  
RxBuffers =  
Default = 200  
Pre-allocates ECB resources on the adapter for  
receiving packets  
Recommended Min = 32  
Max = 512  
Min = 1 when used with DOS  
Client32 and when Keywrod  
P3-1 or 2.  
TxDescriptors =  
Default = 200  
Pre-allocates ECB resources on the adapter for  
transmitting packets.  
Recommended Min = 100  
Max = 512  
Min = 1 when  
used with DOS  
Client32 and when  
Keywrod P3-1 or  
2.  
RxFlow =  
TxFlow =  
Slot = n  
Default = OFF  
Allows enabling/disabling of RxFlow control.  
Allows enabling/disabling of TxFlow control.  
Selections are: ON, OFF  
Default = OFF  
Selections are: ON, OFF  
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  
Jumbo =  
If link negotiation has been disabled, specifies  
port speed to be either Auto, 10HD or 10FD,  
100HD or 100FD.  
Set maximum physical receive Enables/disables Jumbo Frame support. When  
packet Size = 18000 in the  
STARTUP.NCF. Choices are  
Jumbo = 1514–9000. This  
keyword is only supported on  
NetWare 6.x.  
enabled, jumbo packets of up to 9000 bytes  
are supported. Not supported on NC1020  
adapters.  
Link=  
Default=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.  
Selections are: AUTO, FORCE  
RxTicks=  
Default = 360  
Enables the use of batching receives within a  
specific time period.  
Min = 0, disabled  
Max = 5000000, 5 seconds  
Units are in micro seconds  
Installation 30  
 
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 parameter in  
the startup.ncf file to 1500 for each adapter in the system. Set the maximum packet receive buffers to  
three times the minimum packet receive buffers. Typically one MB of RAM is required per 1000 receive  
buffers. For more information, refer to the "Configuring Driver Parameters" section.  
In the autoexec.ncf file, delete the packet receive buffers parameter (RxBuffers=32) in the load  
statement for this adapter. Deleting the receive buffers phrase from the load statement resets the receive  
buffers parameter to the default value of 200 for this adapter.  
NOTE: You must reboot the server for the new configuration.  
Example:  
The default maximum number of receive buffers for the system is 500; the default minimum is 128. Edit  
the startup.ncf file to have the following entries. The actual numbers will be a function of the number of  
adapters in the system.  
set maximum packet receive buffers = 30000  
set minimum packet receive buffers = 10000  
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 useful for verifying  
driver configuration. For information on how to use these programs, refer to the utilities  
reference in your Novell NetWare online documentation.  
The parameters that can be defined in the load statements are described in NetWare server driver LOAD  
line parameters for HP server adapters below. A valid autoexec.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 31  
 
# 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 adapter parameters, you must reboot the system before the  
changes will take effect. If you make changes and do not 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 associated with  
the driver and remark them out by inserting the pound (#) 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 NetWare server driver LOAD line parameter settings for HP server adapters for  
N100, N1000, and N1000e.  
FORCEDUPLEX. This keyword specifies one of the following duplex 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 unsuccessful at negotiating the duplex, HP server adapters  
default to half-duplex. You must have an auto-negotiating switch/hub (an Nway switch) to get  
full-duplex support with the FORCEDUPLEX parameter set to 0 (auto-negotiation).  
Installation 32  
 
 
o
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.  
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 Adapters support full-duplex mode only and  
cannot be changed.  
NOTE: The NC3133 default to full-duplex.  
NOTE: To set the adapter to half- or full-duplex, you must set the SPEED parameter to either 10  
or 100.  
o
Syntax: FORCEDUPLEX=n  
Where n =  
0—auto-negotiate  
1—half-duplex  
2—full-duplex  
o
o
Default = auto-negotiate  
Examples:  
100 Mbps full-duplex FORCEDUPLEX=2 SPEED=100  
10 Mbps full-duplex FORCEDUPLEX=2 SPEED=10  
SLOT. (Required) For PCI adapters, SLOT is derived from bus number and device location as defined  
by the PCI specification. One way to determine the slot number is to load the driver from the  
command line. You will be prompted with valid device number(s) for the adapter(s). Select one of  
them.  
o
Syntax: SLOT=n  
Where n = 1, 2, 3, 4,...  
SPEED. This keyword specifies the speed the driver uses. If you do not use this parameter, the driver  
automatically detects the network speed. 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 network  
speed.  
NOTE: The NC3133 adapter supports 100 Mbps only. 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 speed for the Gigabit (fiber) adapters. You can,  
however, change the speed for the NC7170, NC7131, and NC7132 Gigabit (copper)  
adapters to 10 Mbps or 100 Mbps.  
o
o
Syntax: SPEED=n  
Where n = 10 or 100 or 1000  
Default = The adapter automatically detects network speed.  
Installation 33  
 
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 following digits: 2, 6, A, E.  
o
Syntax: NODE=xnxxxxxxxxxx  
Where n = 2, 6, A, E  
Where x = hexadecimal number  
Default = The adapter's assigned address  
o
FRAME. This keyword indicates one of four valid frame types the system is transmitting and receiving.  
o
Syntax: FRAME=n  
Where n = Ethernet_802.2, Ethernet_802.3, Ethernet_II, Ethernet_SNAP  
Default = Ethernet_802.2  
o
POLL. This keyword is used to enable polling mode in the driver with interrupt backup. If polling is  
enabled, interrupts will be reduced, allowing the processor to spend more time performing other  
functions. In polling mode, interrupts will occur only when receive resources have been reduced to  
less than half. If polling is not enabled, the driver will perform in traditional interrupt mode.  
o
o
o
Syntax: POLL=n  
Where n = 0 (interrupt mode), 1 (polling mode with interrupt backup)  
Default = 0 (interrupt mode)  
NOTE: The POLL parameter is for the N100.LAN, 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 automatically loads for 10, 100, and 1000 adapters.  
o
Syntax: SPURIOUS=n  
Where n = 0 or 1  
0—Driver operates in normal mode  
1—Driver operates to reduce the number of spurious interrupts  
Default = 1  
o
NOTE: The minimum threshold for acceptable number of spurious interrupts 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 offload of Receive Packet Checksum  
verification to the adapter. This automatically loads for 10/100 adapters.  
o
Syntax: RXCHECKSUM=n  
Where n = 0 or 1  
0—Does not offload Rx Checksum verification to the adapter  
1—Offloads Rx Checksum verification to the adapter  
Default = 0  
o
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-duplex.  
Installation 34  
 
The following describes the adapter Status Counters and Software Counters and 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 were 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 varies 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 network is heavily  
saturated with traffic. If this value is consistently 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. Normally, late collision events should occur only  
rarely, if at all. A high number of late collisions is often a sign that there is some problem with the  
network topology. Check the topology to be sure that all cables meet IEEE specifications (Category 5  
for 100 Mbps TX), and that all distance requirements for cabling are met. Also, check to be sure 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 were 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. Under-runs occur because a lack of PCI bandwidth  
resulted in the device's internal transmit FIFO running dry during the transmission of a frame. The  
adaptive algorithms within the software keep this number 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 algorithm, refer to the "CurrentThreshold" section.  
Installation 35  
 
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 during 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 conditions, and the number can grow quite large.  
Tx OK with Single Collisions  
This counter contains the number of transmitted frames that encountered one, and only one, collision  
before being successfully sent. This is a normal condition in busy conditions. If single collisions  
accumulate quickly, you may want to consider segmenting your network more. To combat collisions,  
the software can adjust the interframe gap to allow more room between frames. More room between  
frames allows more frames on average. 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 may have had several collisions before being  
successfully sent. This is a normal condition in busy conditions. If multi-collisions accumulate quickly,  
you may want to consider segmenting your network more. To find the average number of multi-  
collisions a frame had to endure before being successful, use the following formula:  
(TxTotalCollisions - TxSingle Collisions - TxMaxCollisions) / TxMultiple Collisions  
If the average is close to 16 (the Ethernet max collisions before a frame 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 collisions that were encountered while attempting to  
transmit. This count includes late collisions and collisions from frames that encountered  
MAXCOL(16). If this counter is growing quickly, you need to segment your network more.  
Rx Good Frames  
This counter contains the number of frames that were 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 is mutually exclusive to the ALIGNMENT_ERRORS and SHORT_FRAMES  
counters. CRC errors are usually caused by late collisions where 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_ERRORS counter is mutually exclusive to the CRC_ERRORS  
and SHORT_FRAMES counters.  
Rx No Resource Errors  
This counter contains the number of good frames discarded because there were no resources  
available. Frames intended for a host whose Receive Unit is in the "No Resources" state fall into this  
category. The number of resource errors can be reduced by increasing the number of receive buffers  
Installation 36  
 
that software uses, or increasing the performance of the host machine's processing power (which  
would allow for faster processing of the resources).  
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 the first  
could also be lost. However, 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 combat this, a faster bus machine, or 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 frames that are shorter than the minimum frame length.  
The SHORT_FRAMES counter is mutually exclusive to the ALIGNMENT_ERRORS 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 received that are not Flow Control Pause frames.  
These frames contain the predefined MAX 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 and 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 frames 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 produces maximum throughput. But under heavy load  
conditions, this means less time for other stations to send safely; thus, more collisions and less  
throughput.  
By adaptively controlling the IFS, more stations can safely send, lowering collisions and helping to  
keep throughput high. The larger this number is, the more collisions your network 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 37  
 
amount of under-runs. The higher this number, the more time the adapter is waiting before starting to  
transmit data onto the wire. Poor bus performance, large amounts of traffic or a large number 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 conditions, the receiving unit does not start normally and  
needs a restart. RSC has no impact on the end user. It is intended for software 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 restarted. This counter tracks  
the number of times it needs to be restarted. RSC is a subset of REC. Most REC events restart  
normally, but the few that do not are REC events. This counter is intended for software development  
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 packets that the hardware has verified, the checksum,  
and if the checksum is valid.  
Checksum Bad  
This counter indicates the number of received packets 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 interrupt situation. A stalled interrupt situation is detected by the  
driver if there are no interrupts for a long time, and a SW interrupt issued by the driver fails to  
arrive.  
Carrier Loss PHY Resets  
This counter indicates the number of times the driver reset the PHY as a result of carrier sense loss.  
N1000 and N1000e custom counters  
Duplex Mode (1=Half, 2=Full)  
This counter indicates the current link duplex configuration. A value of 1 indicates the adapter is  
configured to half-duplex mode. A 2 indicates full-duplex mode configuration. This counter exists in  
v1.12 or higher of the NetWare server driver.  
Line Speed (Mbps)  
This counter displays the current 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-Express)  
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 38  
 
Link Status Change Count  
This counter indicates the number of times the link status has been changed.  
Rx CRC Error Count  
This counter indicates the number of receive packets with CRC errors. Packets less than 64 bytes are  
not counted in this statistic.  
Alignment Errors  
This counter holds the number of receive packets with 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 errors between reads. The count 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 Sequence Error Count when either 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 received in which I_RX_ER was asserted by the PHY. In  
order for a packet to be counted, it must pass address filtering and must be 64 bytes or greater  
(from <Destination Address> through <CRC>, inclusively) 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 missed when the hardware receive  
FIFO has insufficient space to store the incoming packet. This could be the result of having too few  
buffers allocated or because there is insufficient bandwidth on the IO bus. These packets are also be  
counted in the Total Packets Received counter as well as in Total Octets Received.  
Tx Single Collision Count  
This counter indicates the number of times that a transmitted packet encountered a single collision.  
This counter increments only if the device is in half-duplex mode.  
Tx Excessive Collisions Count  
This counter indicates when 16 or more collisions have occurred on a packet. This counter  
increments only if the device is in half-duplex 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 increments only if the device is in half-duplex mode.  
Tx Late Collisions Count  
This counter indicates collisions that occur after 512 bit (64 byte) times. 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 increments  
only if the device is in half-duplex mode.  
Tx DMA Underruns  
Installation 39  
 
This counter contains the number of frames that were 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. Under-runs occur because a lack of PCI bandwidth  
resulted in the device's internal transmit FIFO running dry during the transmission of a frame. The  
adaptive algorithms within the software 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 when 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 successful packet transmissions in which the Carrier Sense input  
from the PHY was not asserted within one slot time of start of transmission. The PHY should assert  
Carrier Sense during every transmission. Failure to do so may indicate that the link has failed, or the  
PHY has an incorrect link configuration. This counter 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 events. 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 the entire packet) and ran  
out of data to DMA before the device was forced to relinquish ownership of the bus. This condition is  
really not an "error" condition, but is nevertheless counted for diagnostic purposes. This statistic is  
valid only if Early Receives are enabled.  
XON Received Count  
This counter indicates the number of XON packets 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 transmitted. XON packets can occur when Flow  
Control is enabled and both the adapter and its link partner are configured to use it.  
XOFF Received Count  
This counter indicates the number of XOFF packets received. XOFF packets can occur when Flow  
Control is enabled and both the adapter and its 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 its link partner are configured to use it.  
Flow Control Received Unsupported Count  
Installation 40  
 
This counter indicates the number of unsupported flow 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 Packet Count are not counted here. This counter does  
not include received flow control packets.  
65–127 Byte Packets Received  
This counter indicates the number of good packets received that are 65–127 bytes in length,  
including CRC. Packets counted in the Missed Packet Count are not counted here. This counter does  
not include received flow control packets.  
128–255 Byte Packets Received  
This counter indicates the number of good packets received that are 128–255 bytes in length,  
including CRC. Packets counted in the Missed Packet Count are not counted here. This counter does  
not include received flow control packets.  
256–511 Byte Packets Received  
This counter indicates the number of good packets received that are 256–511 bytes in length,  
including CRC. Packets counted in the Missed Packet Count are not counted here. This counter does  
not include received flow control packets.  
512–1023 Byte Packets Received  
This counter indicates the number of good packets received that are 512–1023 bytes in length,  
including CRC. Packets counted in the Missed Packet Count are not counted here. This counter does  
not include received flow control packets.  
>1024 Byte Packets Received  
This counter indicates the number of good packets received that are 1024–1522 bytes in length,  
including CRC. Packets counted in the Missed Packet 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 counter does  
not include received flow control packets.  
Broadcast Packets Received Count  
This counter indicates the number of good broadcast packets received.  
Multicast Packets Received Count  
This counter indicates the number of good multicast 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 length. This counter 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 include received flow control packets.  
Installation 41  
 
Good Octets Received Count HI  
This counter, in conjunction with the Good Octets Received Count LO, makes up a 64-bit counter  
that indicates the number of good octets received. This counter reflects the upper 32 bits of the 64-bit  
counter and is incremented each time the Good Octets 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 Octets 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 include received flow control packets.  
Good Octets Transmitted Count HI  
This counter, in conjunction with the Good Octets Transmitted Count LO, makes up a 64-bit counter  
that indicates the number of good octets transmitted. This counter reflects the upper 32 bits of the 64-  
bit counter and is incremented each time the Good Octets Transmitted Count LO has reached it  
maximum value and wraps to zero. This counter does not include received flow control packets.  
Receive No Buffers Count  
This counter indicates the number of times that packets were received when there were no available  
buffers in host memory to store those 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 during load could result in this counter incrementing less frequently.  
Receive Undersize Count  
This counter indicates the number of received packets that were less than minimum size (64 bytes)  
and had a valid CRC.  
Receive Fragment Count  
This counter indicates the number of received packets that were less than minimum size (64 bytes)  
but had a bad CRC.  
Receive Oversize Count  
This counter indicates the number of received packets that were greater than maximum size (1522  
bytes) and had a valid CRC.  
Receive Jabber Count  
This counter indicates the number of received packets that were greater than maximum size (1522  
bytes) but had a bad CRC.  
Total Octets Received LO  
This counter, in conjunction with the Total Octets 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 Octets 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 Octets Received LO has reached it maximum value  
and wraps to zero.  
Total Octets Transmitted LO  
Installation 42  
 
This counter, in conjunction with the Total Octets 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 Octets 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 Octets Transmitted LO has reached it maximum value  
and wraps to zero.  
Total Packets Received  
This counter indicates the total number of packets received. All packets are counted here, regardless  
of their length, whether they are erred, or whether they are flow control packets.  
Total Packets Transmitted  
This counter indicates the total number of packets transmitted. All packets transmitted will be counted  
here, regardless of their length or whether they are flow control packets.  
64 Byte Packets Transmitted  
This counter indicates the number of good packets transmitted that are exactly 64 bytes in length,  
including CRC. This counter does not include transmitted flow control packets.  
65–127 Byte Packets Transmitted  
This counter indicates the number of good packets transmitted that are 65–127 bytes in length,  
including CRC. This counter does not include transmitted flow control packets.  
128–255 Byte Packets Transmitted  
This counter indicates the number of good packets transmitted that are 128–255 bytes in length,  
including CRC. This counter does not include transmitted flow control packets.  
256–511 Byte Packets Transmitted  
This counter indicates the number of good packets transmitted that are 256–511 bytes in length,  
including CRC. This counter does not include transmitted flow control packets.  
512–1023 Byte Packets Transmitted  
This counter indicates the number of good packets transmitted that are 512–1023 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 transmitted that are 1024–1522 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 multicast packets transmitted. This counter does not  
include transmitted flow control packets.  
Rx Checksum Good Pkts  
This counter indicates the number of receive packets with good checksums calculated. This counter  
increments only when RxTcpChecksum has been enabled and packets with good checksums are  
received.  
Rx Checksum Error Pkts  
Installation 43  
 
This counter indicates the number of receive packets 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 Inter Frame Spacing between back-to-back transmits.  
Jumbo Frames Received  
This counter indicates the number of jumbo packets received (packets greater than 1514 in length).  
Jumbo Frames Transmitted  
This counter indicates the number of jumbo packets transmitted (packets 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 packets that were queued due to no available ECBs.  
Number of Probes Sent by All Team Members  
This counter indicates the number of probe packets sent by all team members.  
Number of Probes Received by All Team Members  
This counter indicates the number of probe packets received by all team members.  
Advanced Network Services help - supported keywords  
AGG_SELECTION  
o
o
Syntax: AGG_SELECTION=[BANDWIDTH | COUNT]  
Description: Sets active aggregator selection mode by bandwidth or count  
BALANCE_INTERVAL  
o
o
Syntax: balance_interval=nnn  
Description: Changes Balance interval, nnn counts 1/18 sec  
BALANCE_SET_DEFAULT  
o
o
Syntax: balance_set_default  
Description: Restores Balance interval to factory settings  
COMMIT  
Syntax: commit [team=nnn]  
o
Installation 44  
 
o
Description: Sets a certain mode. Use after binding to base drivers with Team=nnn.  
DELAY  
o
o
Syntax: delay=nnn  
Description: Delays the commit in nnn seconds  
FRAME  
o
Syntax: frame=[ETHERNET_802.2 | ETHERNET_802.3 | ETHERNET_II | ETHERNET_SNAP]  
HELP  
o
o
o
Syntax: -H  
Description: Displays Help  
Example: CPQANS -H  
JOIN_INDIVIDUALS  
o
o
Syntax: join_individuals=[yes | no]  
Description: Joins all individual links to one aggregator  
LBN  
o
o
o
Syntax: Reset LBN=nnn  
Example: cpqans reset lbn=nnn  
Description: Resets the logical board number (LBN) of a bound adapter. Supported only after  
commit command. Default value = 0  
MAX_TX_QUEUE  
o
o
Syntax: max_tx_queue=nnn  
Description: Sets the Max number of TX ECBs queued for send  
MODE  
o
Syntax: mode=[NFT | ALB | FEC | GEC | 802.3AD]  
NAME  
o
o
Syntax: name=[any unique name]  
Description: Sets a unique name  
PRIMARY  
o
o
Syntax: primary | secondary  
Description: Identifies the primary adapter. Supported only in BIND command  
PROBE_BURST_SIZE  
o
o
Syntax: probe_burst_size=nnn  
Description: Changes number of probes to send in a retry  
PROBE_CHECK_INTERVAL  
o
o
Syntax: probe_check_interval=nnn  
Description: Changes probes check interval  
PROBE_RECHECK_INTERVAL  
o
o
Syntax: probe_recheck_interval=nnn  
Description: Changes probe retries check interval  
Installation 45  
 
PROBE_RETRY_COUNT  
o
o
Syntax: probe_retry_count=nnn  
Description: Changes probes retry count  
PROBE_SEND_INTERVAL  
o
o
Syntax: probe_send_interval=nnn  
Description: Changes probes send interval, nnn counts 1/18 sec  
PROBE_SET_DEFAULT  
o
o
Syntax: probe_set_default  
Description: Restores probes settings to factory settings  
PROBES  
o
o
Syntax: probes=[on | off]  
Description: Enables/disables probes  
PROBES  
o
o
Syntax: probes=[BROADCAST|MULTICAST]  
Description: Changes probes addressing  
RECOMMIT  
o
o
Syntax: recommit [team=nnn]  
Description: Resets a certain mode. Use after hot binding to base drivers with Team=nnn  
REMOVETEAM  
o
o
Syntax: removeteam [team=nnn]  
Description: Removes a team. Use after hot binding to base drivers with Team=nnn  
REMOVEVLANID  
o
o
Syntax: RemoveVlanID=nnn  
Description: Removes the selected VLAN  
RESET  
o
o
Syntax: RESET LBN=nnn  
Description: Supported only after Commit command  
SECONDARY  
o
o
Syntax: primary | secondary  
Description: Identifies the secondary adapter. Supported only in BIND command.  
SMPMODE  
o
o
Syntax: SMPMODE=[SMP | NONSMP]  
Description: Enables/disables SMP aware  
STATUS  
o
o
Syntax: status [team=nnn]  
Description: Prints CPQANS status  
TEAM  
Syntax: team=nnnn  
o
Installation 46  
 
o
Description: Identifies the team. nnnn=DecimalNumber.  
TX_ECBS_TO_USE  
o
o
Syntax: tx_ecbs_to_use=nnn  
Description: Sets number of TX ECBs to allocate per virtual adapter  
VLANID  
o
o
Syntax: VlanID=nnn  
Description: Sets team to VLAN mode. Creates MLID edge  
Installation 47  
 
Firmware upgrade utility  
In this section  
Windows utility...................................................................................................................................... 48  
c-Class BladeSystem utility....................................................................................................................... 54  
Linux utility ............................................................................................................................................ 54  
Windows utility  
These notes describe how to use 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 upgrade their servers to the latest server adapter firmware version:  
When HP releases a new server adapter firmware version  
When installing a new server adapter  
After purchasing a new server  
When installing a SoftPaq that specifies the need for the latest version of firmware  
The Gigabit Server Adapter Firmware Upgrade Utility automatically checks for driver, hardware, and  
operating system dependencies, and installs only the correct server adapter firmware upgrades required  
by each target server.  
Supported operating systems  
The upgrade utility supports the following operating systems:  
Microsoft® Windows® 2000 Server  
Microsoft® Windows® 2000 Advanced Server  
Microsoft® Windows® 2000 Data Center  
Microsoft® Windows® 2000 Terminal Server  
Microsoft® Windows 2003 Server  
Microsoft® Windows 2003 (32-bit) Enterprise Server  
Microsoft® Windows 2003 (32-bit) Data Center  
Microsoft® Windows 2003 (x64) Enterprise Server  
Supported Gigabit Server Adapters  
The upgrade utility provides support for the following adapters:  
Firmware upgrade utility 48  
 
     
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 Multifunction Gigabit Server Adapter  
HP NC373F PCI Express Multifunction Gigabit Server Adapter  
HP NC373T PCI Express Multifunction Gigabit Server Adapter  
HP NC380T PCI Express Dual Port Multifunction 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 NC320i Integrated PCI Express Gigabit Server Adapter  
HP NC324i Integrated Dual Port PCI 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 Server Adapter  
HP NC371i Integrated PCI-X Multifunction 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 Adapter for c-Class BladeSystem  
HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem  
HP NC373m PCI Express Dual Port Multifunction Gigabit Server Adapter for c-Class BladeSystem  
HP NC374m PCI Express Dual Port Multifunction 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.  
Firmware upgrade utility 49  
 
IMPORTANT: Before updating NC6770 or NC7770 server adapters, 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 minutes 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 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.  
Installing the firmware utility  
The HP Network Adapter Online Firmware Upgrade Utility is located in the \apps\fwupgrade\windows  
directory obtained by extracting the HP SoftPaq file. Complete the following steps to download and  
extract the SoftPaq file.  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
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 prompt. The Smart Component for the upgrade utility is  
launched using the command line prompt to allow you 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 Component 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 command line arguments. Smart Components are not  
case-sensitive.  
NOTE: All arguments and information enclosed in brackets are optional.  
Firmware upgrade utility 50  
 
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 the 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. This 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.EXE Smart Component on the  
target server using the defaults of the component.  
CPXXXXXX /F  
Deploys the CPXXXXXX.EXE Smart Component on the  
target server, forcing the installation of the server  
adapter firmware.  
CPXXXXXX /S  
Deploys the CPXXXXXX.EXE 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 51  
 
The upgrade installation activity (including errors) is written by the Smart Component to a common  
installation log file called CPQSETUP.LOG in the \CPQSYSTEM\LOG subdirectory on the target server.  
Information regarding subsequent installation activity is appended to the same log file, which provides a  
chronological history of all component installation activity on the target server.  
Using the NicFwUpg.exe command line interface  
It is possible to launch the HP Online Firmware Upgrade Utility by clicking the executable file during  
installation or by using the command line interface. The following information describes how to launch the  
utility using the command line interface.  
Command line syntax  
The utility supports the following commands:  
Command Line Input  
Result  
nicfwupg.exe -c  
Displays the current 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 adapters in the system  
nicfwupg.exe -s  
nicfwupg.exe -f  
nicfwupg.exe -p  
Suppresses console output  
Forces downgrades or restores as needed  
Requests a restore on all adapters  
nicfwupg.exe -e [MAC Address] Requests 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 Upgrade Utility for Multifunction adapters by clicking the  
executable file during installation or by using the command line interface. The following information  
describes how to launch the utility for Multifunction network adapters using the command line interface.  
Command line syntax  
The utility supports the following commands:  
Firmware upgrade utility 52  
 
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 adapters in the system  
nicfwupg_mf.exe -s  
nicfwupg_mf.exe -f  
nicfwupg_mf.exe -p  
Suppresses console output  
Forces downgrades or restores as needed  
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]  
Performs an upgrade and stores the image  
file to be used during a restore. Supply the  
filename.  
Return codes  
When the upgrade utility has finished running, the component reports a return code to the operating  
system or calling application. These return codes are processed by the %errorlevel% operating system  
variable and are used to determine the status of a firmware upgrade procedure.  
The following summarizes the return codes for the Smart Component installer and HP Online Firmware  
Upgrade Utility commands.  
Error level Meaning  
0
1
3
Upgrade component failed to install  
Upgrade component installed successfully  
Upgrade component did not detect any supported network  
adapters  
4
5
6
Firmware upgrade was recommended, but the upgrade  
component was instructed to not update the firmware  
All firmware versions are current and an upgrade is not  
recommended  
At least one network adapter has firmware installed that is  
newer than the available firmware.  
Firmware upgrade utility 53  
 
c-Class BladeSystem utility  
The HP Gigabit Server Adapter Firmware Upgrade Utility for c-Class BladeSystem uses two separate  
utilities (ccfwupg1.bat and ccfwupg2.bat) to upgrade firmware on the c-Class BladeSystem adapters.  
The ccfwupg1.bat command upgrades the Boot Code and Option ROM firmware 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 and Option ROM firmware 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 image on the adapter, complete the following:  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
8.  
Click the SoftPaq file to extract the files and then open the cmponent.htm file.  
Copy all files and folders from the \apps\ccfwupg\ccfwupg1 directory or the  
\apps\ccfwupg\ccfwupg2 directory to a bootable diskette or USB drive.  
9.  
Boot to DOS using the diskette for your adapter.  
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 the Linux Gigabit Server Adapter Firmware Upgrade  
Utility to flash server adapter firmware.  
Supported operating systems  
The firmware upgrade utility supports the following Linux operating systems:  
Linux 32  
Red Hat Enterprise Linux 4  
Red Hat Enterprise Linux 5  
SUSE Linux Enterprise Server 10 x86  
Firmware upgrade utility 54  
 
   
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 file located in the HP SoftPaq file.Complete the following  
steps to download and extract the SoftPaq file.  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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 utility is located in the apps\fwupgrade\linux folder.  
Supported Gigabit Server Adapters  
The upgrade utility provides support for the 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 Multifunction Gigabit Server Adapter  
HP NC373F PCI Express Multifunction Gigabit Server Adapter  
HP NC373T PCI Express Multifunction Gigabit Server 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 Server Adapter  
HP NC371i PCI-X Multifunction Gigabit Server Adapter  
HP NC373i PCI Express Multifunction Gigabit Server Adapter  
Firmware upgrade utility 55  
 
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 Gigabit Server Adapter  
HP NC325m PCI Express Quad Port 1Gb Server Adapter for c-Class BladeSystem  
HP NC326m PCI Express Dual Port 1Gb Server Adapter for c-Class BladeSystem  
HP NC373m PCI Express Dual Port Multifunction Gigabit Server Adapter for c-Class BladeSystem  
HP NC374m PCI Express Dual Port Multifunction 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 adapters, 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 minutes 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 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.  
To download the latest Linux drivers:  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
Open the Linux folder to select drivers.  
Upgrading firmware  
This firmware upgrade utility updates Boot code, PXE code, and CLP upgrades for c-Class BladeSystem  
(NC325m, NC326m and NC373m) server adapter images. The utility is released in RPM format. The file  
Firmware upgrade utility 56  
 
names for the packages are hpnicfwupg-<version>.i386.rpm for Linux 32 and hpnicfwupg-  
<version>.x86_64.rpm for Linux 64. The instructions below describe the x32 version.  
1.  
Before the RPM package is installed, check for the existence of a previous version by issuing the rpm  
-q hpnicfwupg command.  
2.  
3.  
4.  
5.  
6.  
If a previous version is found, remove it using the rpm -e hpnicfwupgcommand.  
Install the RPM package using the hpnicfwupg–<version>.rpm command.  
Launch the adapter interface by entering ifup ethx or ifconfig ethx up.  
Launch the upgrade utility in interactive mode by entering hpnicfwupg.  
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 restore your original firmware interactively by entering  
/usr/sbin/hpnicfwupg -p.  
If errors have occurred, use the hpnicfwupg utility to restore the server adapter firmware from a backup  
copy.  
Command line syntax  
Use the following command line syntax.  
To  
Use command  
Verify previous version of hpnicfwupg # rpm -q hpnicfwupg  
package exists  
Remove an old version of hpnicfwupg # rpm –e hpnicfwupg  
package  
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 following command line arguments.  
Command line  
argument  
Description  
–y  
–s  
–f  
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 –y is used with  
–s, the behavior is the same as if –s had been specified.  
Silent. Executes the utility silently with no user  
interaction. Default action is implied for all prompts.  
Messages are logged but are not displayed on the  
screen. If the –y option is used with –s, –y is ignored.  
Force. Forces the firmware upgrade even if the  
firmware installed on the system is newer than the  
firmware in the RPM. This option is only valid with –s or  
–y. If neither are specified, the option is ignored.  
Firmware upgrade utility 57  
 
Command line  
argument  
Description  
–p  
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.  
–h  
–c  
Help. Displays the usage message. If this option is used  
with any other option, the other options are ignored.  
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_backupdirectory of the target server.  
Use the hpnicfwupg command with the –p option to restore the original firmware image after an  
upgrade or after an upgrade failure. This option can only be used interactively. The firmware backup  
directory is examined for saved images and a message is displayed showing the current firmware  
versions and the versions from the saved image. You will be prompted before the restore is attempted. If  
more than one saved image is found for a particular server adapter, you will be prompted for each  
image.  
The following is a sample screen shot 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 status (including errors) is written by the Linux server adapter upgrade  
package to a firmware upgrade log file called nicfwupg.login the /var/opt/hpsubdirectory on  
the target server. Information regarding subsequent installation activity is appended to the same log file,  
providing a chronological history of all server adapter firmware upgrade activity on the target server.  
Return codes  
When the Linux server adapter firmware upgrade package has finished running, it reports the following  
return codes to the operating system or calling application.  
Error  
level  
Meaning  
0
Hpnicfwupg installed all selected firmware successfully. For  
options –c and –i , 0 means the information was displayed  
correctly.  
Firmware upgrade utility 58  
 
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 quit in response to a prompt. For  
options –c and –i, a 2 means that the information could not be  
displayed.  
3
Hpnicfwupg attempted to upgrade 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 use 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 upgrade their servers to the latest server adapter firmware version:  
When HP releases a new server adapter firmware version  
When installing a new server adapter  
After purchasing a new server  
When installing a SoftPaq that specifies the need for the latest version of firmware  
The Gigabit Server Adapter Firmware Upgrade Utility 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 the 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 59  
 
 
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 adapters, 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 minutes each to  
update both the PXE and Boot Code.  
NOTE: To update the PXE code for embedded adapters update the system BIOS.  
Drivers  
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.  
Upgrading firmware  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
8.  
9.  
Click the SoftPaq file to extract the files and then open the cmponent.htm file.  
Click the SoftPaq file (SP31728) located in the \apps\fwupgrade\dos directory.  
Execute the downloaded file and follow the on-screen instructions. After the files have been  
unpacked, you may delete the SoftPaq 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 60  
 
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.  
Firmware upgrade utility 61  
 
Adapter configurations  
In this section  
HP NC-Series Multifunction adapters ........................................................................................................ 66  
Windows .............................................................................................................................................. 91  
Novell Netware ................................................................................................................................... 110  
Solaris ................................................................................................................................................ 122  
HP NC-Series Broadcom adapters  
Keywords for Q57 NDIS2 driver  
This document describes the keywords for the 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 DEVNUM—These two keywords are used concurrently and have been included for  
manufacturing purposes. Do not use them unless you are familiar with PCI device configuration.  
These two keywords are needed if multiple NC1020, NC67xx, and/or NC77xx boards are in a  
system and a specific adapter (or adapters) needs to be loaded in specific order. BUSNUM specifies  
the PCI bus number, and DEVNUM specifies the device number given to the board when configured  
by the PCI BIOS. The values are in decimal.  
LineSpeed—This decimal parameter, 10 or 100, specifies the speed of the network connection.  
According to IEEE specifications, line speed of 1000 cannot be forced and its only achievable by  
auto negotiation.  
Duplex—This string parameter, HALF or FULL, specifies duplex mode on the Ethernet controller. The  
LineSpeed parameter must be set when this keyword is used. If neither the Duplex nor the LineSpeed  
parameters are specified the Ethernet controller will default to autonegotiate mode.  
NodeAddress—This string parameter specifies the network address used by the Ethernet controller. If  
Multicast Address or Broadcast Address was specified, the default MAC Address will be used.  
Adapter configurations 62  
 
       
FixCheckSumOff—This string parameter turns off the driver work-around for the TCP/IP stack to  
recognize the 1's complimented version of the checksum.  
The following are examples of the use of these keywords 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 load 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 for the B06 BNDIS 2 driver (bxnd20x.DOS).  
The following keywords are supported through the PROTOCOL.INI file. If a keyword is not present in the  
PROTOCOL.INI, the default value is chosen.  
Adapter configurations 63  
 
   
DRIVERNAME  
o
o
Possible value: BXND20X$  
This field is case sensitive  
LINESPEED  
o
o
Possible values: 10, 100, or 1000  
Default: AUTONEG  
BUSNUM  
o
o
Possible values: the bus slot number the NC37xx/NC380x device is on  
Default: the bus slot number occupied by one of NC37xx/NC380x devices in the system  
DEVNUM  
o
o
Possible value: the device number the NC37xx/NC380x device is on  
Default: the device number occupied by one of NC37xx/NC380x devices in the system  
FUNCNUM  
o
o
Possible values: the function number the NC37x/NC380x device is on  
Default: the function number occupied by one of NC37x/NC380x devices in the system  
DUPLEX  
o
o
Possible values: half or full (LineSpeed is required when this keyword is used)  
Default: full  
NODEADDRESS  
o
o
Possible values: 6 hex bytes looking like "0A0B0C0D0E0F"  
Default: the MAC address for the selected NC37xx/NC380x device  
Known limitations:  
For Windows Server 2003 or Window Server 2003 x64 self-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 adapters 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 servers 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 monolithic  
and split binary images. This feature provides flexibility to users in different environments in which the  
motherboard may or may not have built-in base-code.  
Adapter configurations 64  
 
   
HP PXE implementation is supported in the following 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 universal  
driver binds seamlessly with the HP UNDI (Universal Network Driver Interface) and provides a  
network interface in the Linux remote-booted client environment.  
Microsoft® Windows® 2000, 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 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 connect to network resources in a DOS environment.  
3Com® Boot Server.  
Client setup  
Setting up PXE client for NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server  
Adapters involves enabling/disabling 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 instead.  
However, for some machines equipped with an old BIOS, 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 when 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 instructions for your operating system.  
Windows® 2000  
The current version of Windows® 2000 does not include a network driver for the 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 Adapters to RIS  
Installations."  
DOS UNDI/APITEST  
Adapter configurations 65  
 
To boot in DOS and connect to a network for the DOS 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-  
Red Hat Linux  
Red Hat Linux 7.2 (or higher) distribution has PXE server support. It allows users to perform a complete  
Linux installation over the network. Distribution also comes with boot images—boot 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-network.img  
Refer to Red Hat documentation for instructions on installing PXE server on Linux.  
Initrd.img, distributed with Red Hat 7.2, does not have a Linux network driver for the NC320T, NC320i,  
NC324i, NC325i, NC325m, NC326i, NC326m, NC1020, NC67xx, and NC77xx Gigabit Server  
Adapters.  
/images/pxeboot/vmlinux  
/images/pxeboot/initrd.img  
Intel® has also created patches to the Linux PXE server code to allow clients to perform a remote boot.  
You can download these patches from the Intel website (http://downloadfinder.intel.com/scripts-df-  
A remote boot does not require a standard Linux network driver for the HP NC320T, NC320i, NC324i,  
NC325i, NC325m, NC326i, NC326m, NC1020, NC67xx, and NC77xx Gigabit Server Adapters.  
After the PXE client downloads the Linux kernel and initial ram disk, the Linux universal driver, 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 connections. Using the iSCSI protocol,  
systems can connect to remote 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 enabled 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 Adapter the CPU utilization is reduced, which provides  
additional processing capacity for applications.  
iSCSI is enabled and network settings are configured for the iSCSI device through the Network  
Configuration Utility (NCU). Configuration is limited to IP configuration (including DHCP support), user-  
specified MAC address, and VLAN ID (only one allowed for the iSCSI device).  
Prerequisites  
To use the Multifunction Server Adapter iSCSI accelerating feature, Microsoft® iSCSI Software Initiator  
version 2.0 or later must be installed on your system. This software is required to configure the iSCSI  
protocol, including discovery portals, iSCSI targets, authentication, and so on. The command line utility  
iscsicli.exe, which is part of Microsoft iSCSI Software Initiator can also be used to configure iSCSI  
Adapter configurations 66  
 
   
protocol. See the Microsoft website (http://www.microsoft.com) to download the Microsoft iSCSI  
Software Initiator installation wizard software and to obtain installation and usage information.  
The accelerating iSCSI feature comes standard with standup HP Multifunction Server Adapters.  
Accelerating iSCSI capabilities are available for embedded HP Multifunction Gigabit Server Adapters  
through the purchase and installation of an HP ProLiant Essentials Accelerated iSCSI Pack license  
Configuring advanced settings parameters  
If iSCSI is enabled on a Multifunction adapter port, you can adjust the SRB (iSCSI Request Block) Timeout  
Delta setting and create an iSCSI VLAN for the adapter. 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
o
o
o
o
o
o
o
o
Current MAC Address. The current MAC address for the iSCSI device.  
iSCSI Driver Name. The name of the driver for the iSCSI device  
iSCSI Driver Version. The version of the driver for the iSCSI device.  
Default Gateway. The default route for the iSCSI device.  
Subnet Mask. The subnet mask (dotted decimal) for the iSCSI device.  
iSCSI VLAN Id. The identification number (if any) for the virtual iSCSI VLAN device.  
iSCSI Packet Priority. The packet priority for the iSCSI device.  
SRB Timeout Delta. The increase to the timeout value for I/O operations.  
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
o
DHCP (Dynamic Host Configuration Protocol). Enabled indicates that DHCP is used to acquire IP  
configuration; Disabled indicates that the statically configured IP configuration is used.  
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 systems only.  
The HP iSCSI boot feature allows you to boot 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 implementations, iSCSI 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 adapters has been upgraded to allow PXE functions to  
work with the HP ProLiant servers.  
Adapter configurations 67  
 
     
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 servers 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 monolithic  
and split binary images. This feature provides flexibility to users in different environments in which the  
motherboard may or may not have built-in base-code.  
HP PXE implementation is supported in the following 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 universal  
driver binds seamlessly with the HP UNDI (Universal Network Driver Interface) and provides a  
network interface in the Linux remote-booted client environment.  
Microsoft® Windows® 2000, 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 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 connect to network resources in a DOS environment.  
3Com® Boot Server.  
Client setup  
Setting up PXE client for NC32xx, NC37xx, NC380x, NC1020, NC6770, and NC77xx Gigabit Server  
Adapters involves enabling/disabling 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 instead.  
However, for some machines equipped with an old BIOS, 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 when 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 instructions for your operating system.  
Windows® 2000  
Adapter configurations 68  
 
The current version of Windows® 2000 does not include a network driver for the 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 Adapters to RIS  
Installations."  
DOS UNDI/APITEST  
To boot in DOS and connect to a network for the DOS 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-  
Red Hat Linux  
Red Hat Linux 7.2 (or higher) distribution has PXE server support. It allows users to perform a complete  
Linux installation over the network. Distribution also comes with boot images—boot 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-network.img  
Refer to Red Hat documentation for instructions on installing PXE server on Linux.  
Initrd.img, distributed with Red Hat 7.2, does not have a Linux network driver for the NC320T, NC320i,  
NC324i, NC325i, NC325m, NC326i, NC326m, NC1020, NC67xx, and NC77xx Gigabit Server  
Adapters.  
/images/pxeboot/vmlinux  
/images/pxeboot/initrd.img  
Intel® has also created patches to the Linux PXE server code to allow clients to perform a remote boot.  
You can download these patches from the Intel website (http://downloadfinder.intel.com/scripts-df-  
A remote boot does not require a standard Linux network driver for the HP NC320T, NC320i, NC324i,  
NC325i, NC325m, NC326i, NC326m, NC1020, NC67xx, and NC77xx Gigabit Server Adapters.  
After the PXE client downloads the Linux kernel and initial ram disk, the Linux universal driver, 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 allows your networked client computer to boot using a  
program code image supplied by a remote server.  
Implementation support for the Boot Agent includes:  
Multi-protocol boot ROM code designed for remote booting from Gigabit Ethernet Adapters  
Compliance with the Pre-boot eXecution Environment (PXE) Version 2.1 Specification  
Remote Program Load (RPL) runtime and loader software included with some versions of the Boot  
Agent  
Intel® Boot Agent offers:  
Adapter configurations 69  
 
     
Compliance with the Wired for Management Baseline (WfM) 2.0 specification  
Compatibility with legacy boot agent environments that use BOOTP protocol  
Customization in pre-boot, Windows®, and DOS environments  
Operating environment  
The Boot Agent operates in a client/server environment. Often, in this environment, one or more servers  
provide remote boot services to a large number of client computers through a common network. The  
computer system where the Intel® Boot 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 system has finished booting.  
Configuration options  
The Intel® Boot Agent software provides configuration options that allow you to customize 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 Boot 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 through the boot process. The boot  
process is triggered whenever any of the following boot events occur:  
Power on  
Hard reset (Reset button 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 its Power On Self-Test (POST)  
sequence. Shortly after completion of the POST, the Boot Agent software stored in flash ROM executes.  
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 manufacturer's  
documentation for details.  
To customize the behavior of the Boot Agent software in a pre-boot PXE or RPL environment complete the  
following steps:  
1.  
Press the Ctrl+S keys immediately after the initialization message appears. A configuration setup  
menu appears allowing you to set configuration values for the Boot Agent. The configuration setup  
menu is the same for both kinds of adapters.  
If you do not press the Ctrl+S keys, the Boot Agent software will proceed with the boot process  
eventually bringing up the operating system. If you miss your opportunity to press the Ctrl+S keys  
within the allowed number of seconds, reboot the computer to try again. If you select a Setup Menu  
Wait Time setting of zero or a Show Setup Prompt setting of Disabled, you will not be prompted to  
press the Ctrl+S keys even though you can still enter the configuration setup menu using this key  
combination.  
Adapter configurations 70  
 
The configuration setup menu shows a list of configuration settings on the left and their  
corresponding values on the right. Key descriptions near the bottom of the menu indicate how to  
change values for the configuration settings. For each selected setting, a brief "mini-Help"  
description of its function appears just above the key descriptions.  
2.  
3.  
Select the setting you need to change by using the arrow keys.  
After you have accessed the setting you want to change, press the Spacebar until the desired value  
appears.  
4.  
5.  
If you want to change additional settings, repeat steps 2 and 3.  
After you have completed your changes, press the F4 key to update the adapter with the new values.  
Any changed configuration values are applied as the boot process resumes.  
Boot Agent configuration settings  
Configuration  
settings  
Possible values  
Description  
Network Boot  
Protocol  
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® Management Suite, Windows 2000 RIS,  
and Linux®.  
PXE (Preboot  
eXecution  
Environment)  
RPL (Remote  
Program Load)  
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  
Sets the boot order in which devices are selected during boot up if  
the computer does not have its own control method.  
Use BIOS Setup  
Boot Order  
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 Use BIOS Setup  
Boot Order and cannot be changed. In this case, refer to the BIOS  
setup manual specific to your client computer to set up boot options.  
Try network first,  
then local drives  
Try local drives  
first, then  
network  
If your client computer does not have a BBS- or PnP-compliant BIOS,  
you can select any one of the other possible values listed for this  
setting except for Use BIOS Setup Boot Order.  
Try network only  
Try local drives  
only  
Show Setup  
Prompt  
Controls whether or not the Boot Agent displays the Ctrl+S prompt  
after POST.  
Enabled  
Disabled  
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 71  
 
Configuration  
settings  
Possible values  
Description  
Setup Menu Wait  
Time  
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.  
0 seconds  
2 seconds  
3 seconds  
5 seconds  
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. However, 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 repeatedly 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 installed 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 menu is displayed, you may press the D key to display diagnostics  
information in the lower half of the screen. This information can be helpful during interaction with HP  
Customer Support personnel or your IT team members. After you press the D key, the information  
displayed remains until you leave 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 information  
displayed, refer to the "Diagnostics information for pre-boot PXE or RPL environments" section.  
Auxiliary DOS utilities  
The IBAUtil allows you to install and configure the Boot Agent using the DOS environment. IBAUtil is a  
utility program that provides an alternative means for modifying the adapter configuration settings. Use  
IBAUtil to:  
Change the default settings of your HP NC310F, NC3123, NC6170, or NC7170 adapter.  
Enable/disable the Wake-on-LAN (WOL) and Intel® Boot Agent capabilities.  
Allow in-the-field upgrades to the image stored 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 features. To obtain this utility:  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
Click the SoftPaq file to extract the files and then open the cmponent.htm file.  
The IBAUtil file is located in the APPS\BOOTAGNT\N100X directory.  
Wake-On-LAN. When enabled, the adapter can react to special "wake up" packets and power up  
the computer without end user intervention. However, there is a higher power draw when the system  
Adapter configurations 72  
 
is in the suspended state when this is 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 booted 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 perform both operations if orthogonal  
parameters (like the commands -FLASHENABLE -FLASHDISABLE used together) are supplied. If opposing  
parameters are specified, IBAUtil exits with an error.  
The parameters are grouped with their orthogonal pairs, where applicable. -HELP, -EXITCODES and -  
IMAGEVERSION have special overriding 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 versions 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  
-BLINK  
Selects a specific adapter (1–16).  
Blinks the LED on the selected adapter for 10 seconds 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 parameter 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  
-FLASHDISABLE or -FD  
Enables Boot ROM. Saves  
Disables Boot ROM.  
Adapter configurations 73  
 
Parameter  
Description  
-SAVE  
Saves existing Intel Boot Agent 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 Intel 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  
-LWSENABLE OR -LWSE  
-LWSDISABLE OR -LWSD  
Enables WOL bit on Fast Ethernet and enables APMPME and  
APME EEPROM bits on gigabit.  
Enables changing the legacy OS Wakeup Support Option on  
10/100 adapters.  
Disables changing the legacy OS Wakeup Support Option on  
10/100 adapters.  
PXE/RPL Configuration Options  
-SETUPENABLE or -STE  
Enables Setup Menu.  
Disables Setup Menu.  
-SETUPDISABLE OR -STD  
-DEFAULTCONFIG or -DEFCFG  
Resets configuration 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 -PROE  
-PROTOCOLDISABLE OR -PROD  
-ORDERENABLE OR -ORDE  
-ORDERDISABLE OR -ORDD  
-SETWAITTIME=X or -SWT=X  
Enables changing the boot protocol option.  
Disables changing the boot protocol option.  
Enables changing the boot order option.  
Disables changing the boot order option.  
Sets the setup delay time for the adapter 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  
-PXE BOOT  
Sets the network boot protocol to RPL if an RPL image is present in  
the adapter's flash.  
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 command line when an error occurs.  
0–Success  
1–Operator termination  
2–Invalid adapter  
3–Bad command line parameter  
4–EEPROM checksum failure  
Adapter configurations 74  
 
5–EEPROM read failure  
6–EEPROM write failure  
7–EEPROM dump failure  
8–(Not used)  
9–No memory  
10–No 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 network 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 first HP Network Adapter found in your system:  
IBAUTIL -NIC=1 -FE  
Boot Agent messages  
The following error and information messages may be displayed during initialization. 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 software 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 working correctly. Try updating the BIOS.  
PMM allocation error  
PMM could not or did not allocate the requested 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 Agent was unable to find  
enough free base memory (below 640K) to install the PXE client software. The system cannot boot  
via PXE in its current configuration. The error returns 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 administrator 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 located 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 75  
 
PCI configuration space could not be read. Machine 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 Agent 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 administrator 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 determined 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 administrator 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 Agent cannot  
function in this system. The Boot Agent returns control 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 problem, contact your system administrator or HP customer  
support to resolve the problem.  
PXE-E07: PCI BIOS calls not supported.  
BIOS-level PCI services not available. Machine is probably not PCI compliant.  
PXE-E09: Unexpected UNDI loader error. Status == xx  
The UNDI loader returned an unknown error 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 memory copy error. AH == xx  
Error occurred while trying to copy the image into extended memory. xx is the BIOS failure code.  
PXE-E51: No DHCP or BOOTP offers received.  
The Boot Agent did not receive any DHCP or BOOTP responses to its initial request. Be sure that  
your DHCP server (and/or proxyDHCP server, if one is in use) is 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 configuration. If you are using  
BOOTP, be sure that the TFTP service is running and 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 the 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 flash 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 76  
 
Base code cannot be installed. An incorrect flash 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 image 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 installed 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 structure resource. An incorrect flash image is  
installed or the image has become corrupted. 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 incorrect 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 ROM 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 appear from the configuration setup menu when  
operating in a stand-alone environment. If you think 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 network boot due to a network error (such as not  
receiving a DHCP offer). The Boot Agent continues to attempt to boot from the network until  
successful or until canceled by the user. This feature 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 solutions covers a representative set of problems that you  
might encounter while using the Boot Agent. 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 booting), 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 system, contact the operating  
system manufacturer for assistance. 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 assistance.  
Cannot change boot order  
Adapter configurations 77  
 
If you are accustomed to redefining your computer's boot order using the motherboard BIOS setup  
program, the default settings of the Boot Agent setup program can override that setup. To change the  
boot sequence, you must first override the Boot Agent setup program defaults. To start the Boot Agent  
configuration setup program, press the Ctrl+S keys during the time the initialization message appears. A  
configuration setup menu appears allowing you to set configuration values for the Boot Agent. If you do  
not press the Ctrl+S keys, the Boot Agent software proceeds with the boot process, eventually bringing up  
the operating system. To change your computer's boot order setting, refer to the "Configuring the Boot  
Agent in a pre-boot PXE or RPL environment" section.  
My computer does not complete POST  
If your computer fails to boot with an adapter installed, but does boot when you remove the adapter, try  
moving the adapter to another computer and using IBAUtil to disable the flash chip.  
If this does not work, the problem may be occurring before the Boot Agent software even begins  
operating. In this case, there may be a BIOS problem 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 Configuration 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 display after POST. However, even though you are not  
prompted to press the Ctrl+S keys, you can still enter the configuration setup menu to customize  
configuration settings as follows:  
Repeatedly press the Ctrl+S keys immediately after POST until the 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 Menu, 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 any of the wait time values that exceed  
zero seconds. Refer to the list of possible 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 configuration to troubleshoot the problem. Contact HP  
Customer Support if you need further assistance.  
Diagnostics information for pre-boot PXE or RPL environments  
Anytime the Configuration Setup Menu is displayed (refer to the "Configuring 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 interaction 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 following items:  
PWA Number—The Printed Wire Assembly number of the device as stored in the EEPROM  
Adapter configurations 78  
 
MAC Address—The individual address of the device as stored in the EEPROM  
I/O—The I/O address for PCI access determined by the software. On cards without an I/O  
address, all zeros are displayed  
Memory—The memory map PCI access address determined by the software  
Slot—The slot number reported by the BIOS. The number displayed is the BIOS version of the PCI  
slot number. Therefore, actual positions of adapters within slots may not be displayed as expected.  
Slots are not always enumerated in an obvious manner, and the diagnostics will only report what is  
indicated by the BIOS.  
Boot agent glossary  
This glossary defines terms, abbreviations, and acronyms that apply directly to the Intel® Boot Agent.  
Term / Abbreviation Definition  
/ Acronym  
ACPI  
AOL  
API  
Advanced Configuration and Power Interface  
Alert on LAN  
Application Program Interface. The specific method prescribed 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  
BAR  
BBS  
BC  
BIOS Aware IPL Device. The BIOS contains all code required to IPL from the device.  
Base Address Register  
BIOS Boot Specification  
BaseCode. The PXE BaseCode, included as a core protocol 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 booting machines.  
BCV  
BEV  
Boot Connection Vector. A field in the Plug and Play (PnP) header for a device with  
an associated option ROM.  
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 you turn it on. It also manages data flow  
between the computer's OS and attached devices.  
BIS  
Boot Integrity Services  
BOOTP  
Bootstrap Protocol. A legacy remote booting protocol developed originally for use  
with UNIX. Used as Linux's server side PXE host software. Runs a Daemon once  
installed.  
DDIM  
DHCP  
Device Driver Initialization Model  
Dynamic Host Configuration Protocol. An industry standard internet protocol defined  
by the IETF. DHCP was defined to dynamically provide communications-related  
configuration values such as network addresses to network client computers at boot  
time. DHCP is specified by IETF RFCs 1534, 2131, and 2132.  
EEPROM  
ESD  
Electrically Erasable Programmable Read-Only Memory  
Electro-Static Discharge  
Flash  
A high-density, truly non-volatile, high-performance, read-write memory solution, also  
characterized by low power consumption, extreme ruggedness, and high reliability.  
Adapter configurations 79  
 
Term / Abbreviation Definition  
/ Acronym  
Flb  
Flash Library  
IA32  
IA64  
IAL  
Shorthand for computer platforms based on Intel 32-bit architecture CPUs.  
Shorthand for computer platforms based on Intel 64-bit architecture CPUs.  
Intel Architecture Lab  
IBA  
Intel Boot Agent  
IETF  
Internet Engineering Task Force. The IETF is a large open international community of  
network designers, operators, vendors, and researchers concerned with the evolution  
of the Internet architecture and the smooth operation of the internet. It is open to any  
interested individual. For more details, refer to the Internet Engineering Task Force  
website.  
IPL  
Initial Program Load. Also known as the bootstrap or boot process.  
LANDesk Configuration Manager  
LCM  
LOM  
LAN On Motherboard. This is a network device that is built onto the motherboard (or  
baseboard) of the machine.  
MTFTP  
NBP  
Adds multicast support to a TFTP server.  
Network Bootstrap Program. The first program that is downloaded into a machine  
that has selected a PXE capable device for remote boot services. A typical NBP  
examines the machine it is running on to try to determine if the 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 the EFI boot manager and the next boot  
device is selected. If the machine is capable, the next layer is downloaded and  
control can then be passed to the downloaded program. Though most NBPs are OS  
loaders, NBPs can be written to be stand-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 network connection (including LOMs and  
network devices on external buses, such as USB 1394, and so on).  
OS  
Operating System  
PMM  
POST Memory Manager. 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 provides additional DHCP information that is  
needed by PXE clients and boot servers without making changes to existing DHCP  
servers.  
PXE  
Preboot Execution Environment. PXE provides a way for a system to initiate a  
network connection to various servers prior to loading an OS. This network  
connection supports a number of standard IP protocols such as DHCP and TFTP, and  
can be used for purposes such as software installation and system inventory  
maintenance.  
PXE 16  
RFC  
PXE Specification v2.1 or earlier. Also known as 16-bit PXE.  
Request For Comments. Used to denote that a new standard or modification to  
standard is be introduced by IETF or other standards bodies.  
Adapter configurations 80  
 
Term / Abbreviation Definition  
/ Acronym  
RIS  
Remote Installation Services. A Microsoft® service that uses PXE to deliver the  
Network Bootstrap Program (NBP) to start the boot process. Ships with Windows®  
2000 Advanced Server.  
ROM  
RPL  
Read-Only Memory. When used in this guide, ROM refers to a non-volatile memory  
storage device on an adapter  
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 and 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 the ownership.  
TFTP  
Trivial File Transfer Protocol. An industry standard internet protocol defined by the  
IETF to enable the transmission of files across the internet. Trivial File Transfer  
Protocol (TFTP, Revision 2) to support NBP download is specified by IETF RFC 1350.  
UNDI  
Universal Network Driver Interface. UNDI is an architectural interface to adapters.  
Traditionally adapters have had custom interfaces and custom drivers (each adapter  
had a driver for each OS on each platform architecture). UNDI provides a  
hardware-dependent mechanism 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 Plug and Play (PnP) interface between a computer and add-  
on devices.  
WfM  
WOL  
Wired for Management  
Wake on LAN  
Keywords for N100 NDIS2 driver  
The following how to configure the parameters for the NDIS 2 driver version 4.56 (N100 DOS).  
General configuration parameters  
The following configuration parameters are supported through the PROTOCOL.INI file. The "Possible  
Errors" section describes the driver's behavior when error 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
o
o
o
Syntax: DRIVERNAME=[N100$ | N1002$ | etc]  
Example: DRIVERNAME=N100$  
Default: None. This is a required parameter.  
Normal Behavior: The driver finds its section in PROTOCOL.INI by matching its instance ID to the  
value for this parameter.  
o
Possible Errors: The device driver uses a DOS and OS/2 function to display the name of the  
driver it is expecting. This function cannot display a "$" character. For this reason, the user may  
Adapter configurations 81  
 
   
see a message referring to this value with the "$." The user must remember to enter the "$"  
character as part of the parameter's value.  
SPEED  
o
o
o
o
Syntax: SPEED=[Auto | 10 | 100]  
Example: SPEED=10  
Default: Auto-Speed-Detect as a function of PHY detection  
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 using the value specified. If the parameter is not  
provided at all, or if Auto is provided, Auto-Speed-Sensing is enabled.  
o
Possible Errors:  
If the SPEED parameter is set to an invalid value, the parameter is ignored and the default value  
is used, and a message indicates a "Parameter 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
o
o
o
Syntax: FORCEDUPLEX=[Auto | 1 | 2]  
Example: FORCEDUPLEX=1  
Default: Duplex mode configuration detected during PHY initialization.  
Normal Behavior: The parameter value of 1 will force the duplex mode to half-duplex. A  
parameter value of 2 will force the 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 speed and full-duplex modes are set as specified.  
The N-Way PHY will not be able to auto-detect the mode until the driver is reloaded.  
o
Possible Errors:  
If the FORCEDUPLEX parameter is set to an invalid value, the parameter is ignored and the  
default value is used, and a message indicates a "Parameter value out of range" error.  
If the FORCEDUPLEX mode does not match the actual 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 FORCEDUPLEX set to FULL.  
PHYADDRESS  
o
o
o
o
Syntax: PHYADDRESS=[0..32]  
Example: PHYADDRESS=1  
Default: 1, implying logical PHY 1 at address 1.  
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 82  
 
o
o
Normal Behavior: The value will be used by the PHY detection routine to locate and initialize the  
PHY. If a PHY is not found at the address specified, the driver will load but not function properly,  
because PHY detection occurs after the initial load has occurred.  
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 indicates a "Parameter value out of range" error.  
SLOT  
o
Syntax: SLOT=[0x0..0x1FFF]  
Optional unless you want to specify a particular adapter or module that is not loaded by default.  
(By default, the driver loads on the first adapter it finds.)  
o
o
o
o
Example: SLOT=0x1C  
Default: The driver will Auto-Configure  
Normal Behavior: The driver uses the value of the parameter to decide which adapter to control.  
Possible 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 the actual configuration, and the driver finds the  
adapter and uses it.  
If more than one adapter is installed and the value 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 hexadecimal digits, enclosed in double quotes  
The value cannot be all zeros. The value cannot have the Multicast bit set (LSB of 2nd digit = 1).  
Example: NODE="00AA00123456"  
o
o
o
Default: Value from EEPROM installed on adapter  
Normal Behavior: The Current Station Address in the NDIS MAC Service-Specific Characteristics  
(MSSC) table is assigned the value of this parameter. The adapter hardware is programmed to  
receive frames with the destination 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
o
o
o
Syntax: CONGENB=[0 | 1]  
Example: CONGENB=1  
Default: Congestion Control is disabled.  
Normal Behavior: The absence of this parameter or the value 0 results in no Congestion Control.  
The value of 1 enables the Congestion Control feature.  
o
Possible Errors: If the CONGENB parameter is set to an invalid value, the parameter is ignored  
and the default value is used, and a message indicates a "Parameter value out of range" error.  
RXFIFO  
o
o
Syntax: RXFIFO=[0..15 | 0x0..0xF]  
Example: RXFIFO=8  
Adapter configurations 83  
 
o
o
Default: 8, corresponding to 32 bytes or 8 dwords in the receive FIFO.  
Normal Behavior: No message will be displayed. The driver will use the specified value to  
configure.  
o
Possible Errors: If the RXFIFO parameter is set to an invalid value, the parameter is ignored and  
the default value is used, and a message indicates a "Parameter value out of range" error.  
TXFIFO  
o
o
o
o
Syntax: TXFIFO=[0..15 | 0x0..0xF]  
Example: TXFIFO=8  
Default: 8, corresponding to 32 bytes or 8 dwords in the transmit FIFO.  
Normal Behavior: No message will be displayed. The driver will use the specified value to  
configure.  
o
Possible Errors: If the TXFIFO parameter is set to an invalid value, the parameter is ignored and  
the default value is used, and a message indicates a "Parameter value out of range" error.  
TXTHRESHOLD  
o
o
o
o
Syntax: TXTHRESHOLD=[0..254 | 0x0..0xFE]  
Example: TXTHRESHOLD=32  
Default: 32, corresponding to 32 * 8=256 bytes  
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 detected.  
Setting a low value (less than 16) in high bus latency systems could cause excessive TX DMA  
underruns, reducing throughput until the driver reacts. Setting a value of more than 200  
completely eliminates the possibility of TX DMA underruns.  
o
Possible Errors:  
If the TXTHRESHOLD parameter is set to an invalid value, the parameter is ignored and the  
default value is used, and a message indicates a "Parameter value out of range" error.  
IFS  
o
Syntax: IFS=[0..255 | 0x0..0xFF]  
Example: IFS=16  
o
o
Default: 0 (minimum spacing between frames)  
o
Normal Behavior: The value will be used to configure the 82557 at initialization 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 clients, but will have a negative impact on transmit  
throughput in a point-to-point single-client packet burst.  
o
Possible Errors: If the IFS parameter is set to an invalid value, the parameter is ignored and the  
default value is used, and a message indicates a "Parameter value out of range" error.  
RXDMACOUNT  
o
o
o
Syntax: RXDMACOUNT=[0..127 | 0x0..0x7F]  
Example: RXDMACOUNT=16  
Default: 0 implying that the Rx DMA will continue until it is done. No other DMA request can  
preempt it.  
Adapter configurations 84  
 
o
o
Normal Behavior: The value will be used to configure the 82557 at initialization time. Setting it  
to a low value will force frequent internal arbitration, which will affect overall performance. The  
value can be tuned to obtain optimum balance between the TX and Rx DMA machines.  
Possible Errors:  
If the RXDMACOUNT parameter is set to an invalid value, the parameter is ignored and the  
default value is used, and a message indicates a "Parameter value out of range" error.  
TXDMACOUNT  
o
o
o
Syntax: TXDMACOUNT=[0..127 | 0x0..0x7F]  
Example: TXDMACOUNT=16  
Default: 0 implying that the Tx DMA will continue until it is done. No other DMA request can  
preempt it.  
o
o
Normal Behavior: The value will be used to configure the 82557 at initialization time. Setting it  
to a low value will force frequent internal arbitration, which will affect overall performance. The  
value can be tuned to obtain optimum balance between the Tx and Rx DMA machines.  
Possible Errors:  
If the TXDMACOUNT parameter is set to an invalid value, the parameter is ignored and the  
default value is used, and a message is indicates a "Parameter value out of range" error.  
TRANSMITBUFFERS  
o
o
o
o
o
Syntax: TRANSMITBUFFERS=[4 .. 30]  
Example: TRANSMITBUFFERS=14  
Default: 4 for DOS driver, 10 for OS/2 driver  
Normal Behavior: The driver allocates the number of buffers indicated by the parameter value.  
Possible Errors:  
If the value provided is out of range, the parameter 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 causes the driver’s size to exceed MAXIMUM,  
the driver adjusts the number of buffers downward, by reducing each of the counts of  
transmit/receive buffers by one until the total allocation will fit in the available memory. A  
message indicates the actual number of buffers allocated. Under DOS without Window for  
Workgroups (WFW), MAXIMUM is defined 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.  
RECEIVEBUFFERS  
o
o
o
o
o
Syntax: RECEIVEBUFFERS=[4 .. 30]  
Example: RECEIVEBUFFERS=14  
Default: 8 for DOS driver, 10 for OS/2 driver  
Normal Behavior: The driver allocates the number of buffers indicated by the parameter value.  
Possible Errors:  
If the value provided is out of range, the parameter is ignored, the default value is used, and a  
message indicates a "Parameter value out of range" error.  
Adapter configurations 85  
 
CAUTION: If the number of buffers allocated causes the driver’s size to exceed MAXIMUM,  
the driver adjusts the number of buffers downward, by reducing each of the counts of  
transmit/receive buffers by one until the total allocation will fit in the available memory. A  
message indicates the actual number of buffers allocated. Under DOS without Window for  
Workgroups (WFW), MAXIMUM is defined 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
o
o
o
Syntax: MCWA=[0..255 | 0x0..0xFF]  
Example: MCWA=8  
Default: 6  
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 Receive Hang workaround will be executed. When  
the workaround condition exists, the service routine 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 needed, the counter will not be added to the timer  
chain, saving system resources.  
o
Possible Errors: The value is not checked in any way, only 8 bits of value are used. The counter's  
effectiveness could be compromised if another executable disrupts the timer chain.  
FRAMEGAP  
o
o
o
o
Syntax: FRAMEGAP=[6..64 | 0x06..0x40]  
Example: FRAMEGAP=12  
Default: 6  
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 parameter unless you are experiencing dropped packets.  
o
Possible Errors: If an invalid range or syntax for this parameter occurs, the default value is used.  
The driver does not give any outward indication as to the state of this parameter.  
CACHEFLUSH  
o
o
o
o
Syntax: CACHEFLUSH=[0|1]  
Example: CACHEFLUSH=1  
Default: 0  
Normal Behavior: Use this parameter during a remote installation or "push-install" of Windows  
NT®.  
o
Possible Errors: Any non-zero value sets this parameter to 1. The driver does not give any  
outward indication as to the state of this parameter.  
Adapter configurations 86  
 
Keywords for N1000 NDIS2 driver  
The following describes how to configure the parameters for the NDIS 2 driver version 3.62  
(N1000.DOS).  
Configuration  
The following configuration parameters are supported through the PROTOCOL.INI file. The "Possible  
Errors" section describes the driver's behavior when error 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
o
o
o
Syntax: DRIVERNAME=[N1000$ | N10002$ | etc]  
Example: DRIVERNAME=N1000$  
Default: None. This is a required parameter.  
Normal Behavior: The driver finds its section in PROTOCOL.INI by matching its instance ID to the  
value for this parameter.  
o
Possible Errors: The device driver uses a DOS function to display the name of the driver it is  
expecting. This function cannot display a "$" character. For this reason, the user may see a  
message referring to this value without the "$." The user must remember to enter the "$"  
character as part of the parameter's value.  
SPEEDDUPLEX  
o
o
o
o
Syntax: SPEEDDUPLEX = [0 | 1 | 2 | 3]  
Example: SPEEDDUPLEX = 1  
Default: Speed and Duplex mode configuration detected during PHY initialization.  
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 and the duplex mode to half duplex.  
3 will force the speed to 100 Mbps and the duplex mode to full duplex.  
o
o
Auto-detection by the PHY is disabled and the speed and full duplex modes are set as specified.  
The PHY will not be able to auto-detect the mode until the driver is reloaded.  
Possible Errors:  
If the SPEEDDUPLEX parameter is set to an invalid value, the parameter is ignored and the  
default value is used, and a message indicates 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 NC71xx Server Adapters or Modules only and is  
ignored if used with HP NC61xx 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 want to specify a particular adapter or module that is not loaded by default.  
Example: SLOT=0x1C  
o
o
Default: The driver will Auto-Configure if possible.  
Adapter configurations 87  
 
   
o
o
Normal Behavior: The driver uses the value of the parameter to decide which adapter to control.  
Possible Errors:  
If only one adapter is installed and the value does 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 value 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 hexadecimal digits, enclosed in double quotes.  
The value cannot be all zeros.  
The value cannot have the Multicast bit set (LSB of 2nd digit == 1).  
Example: NODE = "02AA00123456"  
o
o
o
Default: Value from EEPROM installed on adapter  
Normal Behavior: This value sets the adapter's Individual Address, overriding the value read  
from the adapter EEPROM. The address must consist 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 second 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
o
o
o
Syntax: CACHEFLUSH= [0|1]  
Example: CACHEFLUSH=1  
Default: 0  
Normal Behavior: Use this parameter during a remote installation or "push-install" of Windows  
NT®.  
o
Possible Errors: Any non-zero value sets this parameter to 1. The driver does not give any  
outwards indication as to the state of this parameter.  
Example PROTOCOL.INI file  
PROTOCOL.INI section for the HP Gigabit Server Adapters and Modules:  
DRIVERNAME parameter  
o
o
Use this parameter to set 1st adapter = N1000$, 2nd adapter = N10002$, and so on  
DRIVERNAME = N1000$  
NODE parameter  
o
o
Use this to change the MAC address used by the adapter  
NODE="02AA00123456"  
SLOT parameter  
o
Use this to configure a particular instance of the driver designated by DRIVERNAME, to use the  
adapter in a particular PCI slot. Slots should be specified in hex notation.  
Adapter configurations 88  
 
o
SLOT=0x7  
CACHEFLUSH parameter  
o
Set this to 1 to hook the system Disk Buffer Flush interrupt. This causes the driver to properly turn  
off the adapter when the system is shut down by Windows NT® during remote installation (push-  
install).  
o
o
CACHEFLUSH=1  
Known Limitations: This driver does not support any application that uses protected mode, for  
example, EMM386. In addition, testing has demonstrated 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 adapters has been upgraded to allow PXE functions to  
work with the HP ProLiant servers.  
The Boot Agent software provides the ability to customize its behavior through its configuration 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 following  
environments:  
A pre-boot environment (before an operating system is loaded)  
A Windows® environment  
A DOS environment  
Configuring the Boot Agent in a pre-boot environment  
You can customize the behavior of the Boot Agent software through a pre-boot (operating system-  
independent) configuration program contained within the adapter's flash ROM. You can access this pre-  
boot configuration setup program each time the client 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 during the boot process to Boot Agent 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 before 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 configuration settings on the left, and their corresponding  
values on the right. Key functions are described near the bottom of the menu. The menu indicates how to  
change the values of settings.  
To access a brief Help description of a setting, select it.  
Adapter configurations 89  
 
   
To customize the Boot Agent configuration settings:  
1.  
2.  
3.  
4.  
Highlight the setting you want to change using the arrow keys. Then press Enter.  
Press the spacebar until the desired value appears.  
If you want to change additional settings, repeat steps 1 and 2.  
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 configuration settings.  
Configuration  
settings  
Possible values  
Description  
Network Boot  
Protocol  
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® Management Suite, Windows 2000 RIS,  
and Linux®.  
PXE (Preboot  
eXecution  
Environment)  
RPL (Remote  
Program Load)  
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  
Sets the boot order in which devices are selected during boot up if  
the computer does not have its own control method.  
Use BIOS Setup  
Boot Order  
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 Use BIOS Setup  
Boot Order and cannot be changed. In this case, refer to the BIOS  
setup manual specific to your client computer to set up boot options.  
Try network first,  
then local drives  
Try local drives  
first, then  
network  
If your client computer does not have a BBS- or PnP-compliant BIOS,  
you can select any one of the other possible values listed for this  
setting except for Use BIOS Setup Boot Order.  
Try network only  
Try local drives  
only  
Show Setup  
Prompt  
Controls whether or not the Boot Agent displays the Ctrl+S prompt  
after POST.  
Enabled  
Disabled  
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 90  
 
Configuration  
settings  
Possible values  
Description  
Setup Menu Wait  
Time  
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.  
0 seconds  
2 seconds  
3 seconds  
5 seconds  
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. However, 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 repeatedly 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 Network 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
o
Default = Auto/Auto  
Range = Auto/Auto; 10/Half; 10/Full; 100/Half; 100/Full  
NOTE: The NC3133 is a 100BASE-FX fiber optic upgrade module that supports 100 Mbps  
only.  
Locally Administered Address. Specifies the user-defined MAC Address of the adapter, which  
overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed.  
o
o
Default = NULL  
Range = Valid unicast MAC Address  
802.1p QoS Packet Tagging. Enables or disables IEEE 802.1p tagging to send network traffic with  
different priority levels.  
o
o
Default = Disabled  
Range = Disabled; Enabled  
Adaptive Performance Tuning: NIC Bandwidth/CPU Utilization (NC31xx adapters only). Specifies the  
number of frames the adapter receives before triggering 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 parameter causes a high rate of receive interrupts  
(that is, a setting of the lowest value will result in an interrupt for every frame received). A low value  
increases adapter bandwidth, but may reduce CPU efficiency, slowing your computer. Setting a high  
value for this parameter reduces the interrupt rate (that is, increases the number of frames the  
Adapter configurations 91  
 
     
adapter receives before generating an interrupt). A high value improves CPU efficiency, but may  
reduce adapter bandwidth.  
o
o
Default = 1536  
Range = 0–4096 (increments of 32)  
Number of Coalesce Buffers (NC31xx adapters only). Specifies the number of memory buffers  
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 packet for later transmission. The preferred method of  
transmitting data is to use map registers, because it is the most efficient method.  
o
o
Default = 8  
Range = 1–32  
Number of Receive Descriptors: Minimum/Maximum. Specifies the number of descriptors used by the  
driver when copying data to the protocol memory. In high network load situations, increasing  
receive descriptors can increase performance. The tradeoff is that this also increases the amount of  
system memory used by the driver. If too few receive descriptors are used, performance suffers. If too  
many receive descriptors are used, the driver unnecessarily consumes memory resources.  
o
o
Default = 48  
Range = 8–1024 (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 many TCBs are used, the driver  
unnecessarily consumes memory resources.  
o
o
Default = 32  
Range = 1–64  
Adaptive Transmit Threshold. Determines how many bytes are in the FIFO before transmission  
begins.  
o
o
Default = 200  
Range = 0–200  
Flow Control. The Flow Control property allows the user to enable or disable the receipt or  
transmission of PAUSE frames. PAUSE frames enable the adapter and the switch to control the  
transmit rate. The side that is receiving the PAUSE frame will momentarily stop transmitting. The  
recommended selection is Off and it is the default setting. To disable Flow Control, select Off from  
the Value list on the Advanced tab.  
o
o
o
o
Off. (Default) PAUSE frame receipt and transmission is disabled  
Generate. PAUSE frame transmission is enabled  
Respond. PAUSE frame receipt is enabled  
Both on. PAUSE frame receipt and transmission is enabled  
HP NC110x, NC310F, NC340x, NC360x, NC364x, 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: Auto/Auto  
Adapter configurations 92  
 
o
Range  
1000BASE-SX or -LX Adapters: Auto/Auto  
10/100/1000 Adapters: Auto/Auto; 10/Half; 10/Full; 100/Half; 100/Full; 1000/Full  
NOTE: This property is disabled for the NC310F and NC61xx adapters.  
Locally Administered Address. Specifies the user-defined MAC Address of the adapter, which  
overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed.  
o
o
Default = NULL  
Range = Valid unicast MAC Address  
802.1p QoS Packet Tagging. Enables or disables IEEE 802.1p tagging to send network traffic with  
different priority levels.  
o
o
Default = Disabled  
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 putting more data in each packet, thereby sending  
out fewer packets. This is equivalent to Jumbo Frames and must comply with switch limitations.  
o
o
Default = 1514 Bytes  
Range  
NC310F and NC61xx: 1514 Bytes; 4088 Bytes; 9014 Bytes; 16128 Bytes  
NC110x, NC340x, NC360x, NC364x, and NC71xx: 1514 Bytes; 4088 Bytes; 9014 Bytes  
Number of Coalesce Buffers. Specifies the number of memory buffers 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 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 NC364x not supported)  
Windows Server 2003 and Windows Server 2003 x64: 128  
Range = 16–768 (increments of 8)  
o
Number of Receive Descriptors: Minimum/Maximum. Specifies the number of descriptors used by the  
driver when copying data to the protocol memory. In high network load situations, increasing  
receive descriptors can increase performance. The tradeoff is that this also increases the amount of  
system memory used by the driver. If too few receive descriptors are used, performance suffers. If too  
many receive descriptors are used, the driver unnecessarily consumes memory resources.  
o
Default  
Windows® 2000: 256 (NC110x and NC364x not supported)  
Windows Server 2003 Windows Server 2003 x64  
NC110x, NC310F, NC340x, NC360x, NC364x, NC6170, and NC7170: 256  
All other adapters: 160  
Range  
o
Windows® 2000  
Adapter configurations 93  
 
NC6132, NC6133, and NC6134: 80–768 (increments of 8)  
NC6136, NC7131, and NC7132: 80–256 (increments of 8)  
NC310F, NC340x, NC6170, and NC7170: 80–4096 (increments of 32)  
Windows Server 2003 Windows Server 2003 x64  
NC6132, NC6133, and NC6134: 80–768 (increments of 8)  
NC6136, NC7131, and NC7132: 80–256 (increments of 8)  
NC310F, NC340x, NC6170, and NC7170: 80–4096 (increments of 32)  
NC110x, NC360x, and NC364x: 80-2048 (increments of 8)  
Number of Transmit Descriptors: Minimum/Maximum. Specifies the number of 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 allocated).  
o
o
Default = 256  
Range  
Windows® 2000  
NC6132, NC6133, and NC6134: 80–768 (increments of 8)  
NC6136, NC7131, and NC7132: 80–256 (increments of 8)  
NC310F, NC340x, NC6170, and NC7170: 80–4096 (increments of 32)  
Windows Server 2003  
NC6132, NC6133, and NC6134: 80–768 (increments of 8)  
NC6136, NC7131, and NC7132: 80–256 (increments of 8)  
NC310F, NC340x, NC6170, and NC7170: 80–4096 (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
o
Default = Enabled  
Range = Disabled; Enabled  
Tx TCP Checksum Offload. Offloads the task of computing the checksum for outgoing TCP or UDP  
packets, thereby improving performance.  
o
o
Default = Enabled  
Range = Disabled; Enabled  
Tx IP Checksum Offload. Offloads the task of computing the checksum for outgoing IP packets,  
thereby improving performance.  
o
o
Default = Enabled  
Range = Disabled; Enabled  
Rx IP Checksum Offload. (NC110x, NC310F, NC340x, NC360x, NC364x, NC6170, and  
NC7170 only). Offloads the task of computing the checksum for incoming IP packets, thereby  
improving performance.  
o
o
Default = Enabled  
Range = Disabled; Enabled  
Adapter configurations 94  
 
Enable PME. (NC110T, NC310F, NC340x, NC360x, NC364T, NC6170, and NC7170 only)  
Allows the use of power management and wake-up functions.  
o
o
Default = No Action  
Range = Disable; Enable; No Action; H/W Default  
Wake On. (NC110T, NC310F, NC340T, NC360x, NC364T, NC6170, and NC7170 only)  
Available only when Enable PME is enabled. Allows you to choose what types of packets will cause  
the adapter to wake.  
o
o
Default = Disable (NC110, NC360x, NC364x Default=OS Controlled)  
Range = Disable; Magic Packet; Direct Packet Magic and Direct Packets; OS Controlled  
Large Send Offload. Enables offloading of large TCP packets (NC110T, NC310F, NC340x,  
NC360x, NC364T, NC6170, and NC7170 only). Windows 2003 and Windows 2003 x64 only.  
o
o
Default = Enabled  
Range = Disabled; Enabled  
Gigabit Master Slave Mode. (NC110x, NC360x, NC364x only) Determines IEEE 802.3ab Master-  
Slave resolution during gigabit auto negotiation.  
o
o
Default = Auto Detect  
Range = Auto Detect, Force Master Mode, Force Slave Mode  
Smart Power Down. (NC110T, NC360x, and NC364T only) Minimizes power consumption by  
enabling the adapter to enter a deep sleep mode under certain conditions.  
o
o
Default = H/W Default  
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
o
Default = Disable  
Range = Disable, Force  
Flow Control. (Windows Server 2003 only) Determines the type of flow control that is needed.  
(NC110x, NC360x, NC364x only):  
o
o
Default = Both On  
Range = Off, Respond, Generate, Both On  
HP NC150T, NC32x, NC37xx, NC380T, NC1020, NC67xx, and NC77xx Server Adapters  
Number of Transmit Descriptors: Minimum/Maximum. Specifies the number of 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 allocated).  
o
o
Default = 200  
Range = 100–600  
NOTE: This property is not supported on HP Multifunction Server Adapters.  
Number of Receive Descriptors: Minimum/Maximum. Specifies the number of descriptors used by the  
driver when copying data to the protocol memory. In high network load situations, increasing  
receive descriptors can increase performance. The tradeoff is that this also increases the amount of  
Adapter configurations 95  
 
system memory used by the driver. If too few receive descriptors are used, performance suffers. If too  
many receive descriptors are used, the driver unnecessarily consumes memory resources.  
o
o
Default = 200  
Range = 100–512  
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/Auto  
Default = Auto/Auto  
o
NC32x, NC37xx, NC380T, NC1020, and NC77xx  
Range = Auto/Auto; 10/Half; 10/Full; 100/Half; 100/Full; 1000/Full  
Default = Auto/Auto  
Locally Administered Address. Specifies the user-defined MAC Address of the adapter, which  
overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed.  
o
o
Default = NULL  
Range = Valid unicast MAC Address  
802.1p QoS Packet Tagging. Enables or disables IEEE 802.1p tagging to send network traffic with  
different priority levels.  
o
o
Default = Disabled  
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 putting more data in each packet, thereby sending  
out fewer packets. This is equivalent to Jumbo Frames and must comply with switch limitations.  
When TCP/IP Offload Engine (TOE) is enabled, this property cannot be set to greater than 1514  
bytes. If this property is set to greater than 1514 bytes, TOE cannot be enabled.  
o
o
Default = 1514 Bytes  
Range = 1514 Bytes; 4088 Bytes; 9014 Bytes  
NOTE: This property is not supported 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
o
Default = Both  
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 96  
 
o
o
Default = Auto  
Range = Auto, 10Mb, 100Mb  
Flow Control. Specifies the type of flow control that is needed.  
o
o
Default = Auto  
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
o
Default = Enabled  
Range = Disabled; Enabled  
Tx TCP Checksum Offload. Offloads the task of computing the checksum for outgoing TCP or UDP  
packets, thereby improving performance.  
o
o
Default = Enabled  
Range = Disabled; Enabled  
Tx IP Checksum Offload. Offloads the task of computing the checksum for outgoing IP packets,  
thereby improving performance.  
o
o
Default = Enabled  
Range = Disabled; Enabled  
Rx IP Checksum Offload. Offloads the task of computing the checksum for incoming IP packets,  
thereby improving performance.  
o
o
Default = Enabled  
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
o
Default = Enabled  
Range = Disabled; Enabled  
NOTE: This property is not supported by NC7761 adapters.  
Large Send Offload. Normally, the TCP segmentation 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 from the Value list on  
the Advanced tab.  
o
o
Default = Enabled  
Range = Disabled; Enabled  
IPMI. Enables Intelligent Platform Management Interface, 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 97  
 
o
o
SRB Timeout Delta. Increases or decreases the timeout value for I/O operations through the iSCSI  
HBA. Increasing this value may help prevent premature I/O timeouts if your 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 VLAN. Allows you to create a VLAN for an iSCSI-enabled device.  
TCP/IP Offload Engine (TOE). Offloads TCP connections to the Multifunction Server Adapter.  
(Microsoft Scalable Networking Pack (SNP) for Windows 2003 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 Windows Server 2003 Scalable Networking Pack (SNP)  
and the required Microsoft hotfixes in article numbers 921136, 919948, and 923187 of the  
Microsoft Knowledge Base (KB).  
o
o
Default = Enabled  
Range = Disabled; Enabled  
Receive-Side Scaling (RSS). Allows the network load from a network 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
o
Default = Enabled  
Range = Disabled; Enabled  
Optimized Interrupt Moderation. When this configuration is set to optimize, the driver automatically  
and continuously determines the best Target DPC (deferred procedure call) Rate to use based on the  
current network conditions. When this configuration is set to manual, the user can select a Target  
DPC Rate and the driver adjusts the adapter's interrupt setting to achieve that DPC rate. Changing  
this setting may improve performance for some configurations.  
Target DPC Rate. Determines the desired DPC rate. The driver adjusts the adapter's interrupt  
moderation setting dynamically based 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 following 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.  
3.  
Select the Hardware tab and click Device Manager. The Device Manager window appears.  
Scroll down the list of hardware devices to Network Adapters and click the plus (+). This displays the  
list of adapters currently configured.  
4.  
5.  
6.  
7.  
Right-click the adapter to configure and select Properties. The Properties window appears.  
Select the Advanced tab. The list of properties and current values for the adapter appears.  
Select the property to modify and adjust the Value as needed.  
Save the settings.  
a. When the adapter property configuration is complete, click OK to accept the settings.  
Adapter configurations 98  
 
   
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 registers.  
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 default. 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 properties, the checksum can be computed by the Gigabit  
Ethernet Adapter.  
To enable one of the Checksum Offload properties select the property from the Value list on the  
Advanced tab.  
o
o
o
o
None. Disables checksum offloading  
Rx TCP/IP Checksum. Enables receive TCP, IP, and UDP checksum offloading  
Tx TCP/IP Checksum. Enables transmit TCP, IP, and UDP checksum offloading  
Tx/Rx TCP/IP Checksum. (Default) Enables transmit and receive TCP, IP, and UDP checksum  
offloading  
Ethernet@WireSpeed. This property is only available through the HP Network Configuration Utility.  
The Ethernet@WireSpeed property 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 property  
is enabled. To disable the Ethernet @ WireSpeed property, select Disable from the Value list.  
o
o
Disable. Disables Ethernet@WireSpeed  
Enable. (Default) Enables Ethernet@WireSpeed  
Flow Control. The Flow Control property allows the user to enable or disable the receipt or  
transmission of PAUSE frames. PAUSE frames enable the adapter and the switch to control the  
transmit rate. The side that is receiving the PAUSE frame will momentarily stop transmitting. The  
recommended selection is Auto and it is the default setting. To disable Flow Control, select Disable  
from the Value list on the Advanced tab.  
o
o
o
o
o
Auto. (Default) Auto detect flow control  
Disable. PAUSE frame receipt and transmission is disabled  
Tx PAUSE. PAUSE frame transmission is enabled  
Rx PAUSE. PAUSE frame receipt is enabled  
Rx/Tx PAUSE. PAUSE frame receipt and transmission is enabled  
Large Send Offload. Normally, the TCP segmentation is performed by the protocol stack. By  
enabling the Large Send Offload property, the TCP segmentation can be performed by the Gigabit  
Ethernet Adapter. To enable the Large Send Offload property, select Enable from the Value drop-  
down menu on the Advanced tab.  
o
o
Disable. Disables large send offloading  
Enable. Enables large send offloading  
Locally Administered Address. Specifies the user-defined MAC Address of the adapter, which  
overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed.  
o
o
Default = Not Present  
Range = Value; Not Present  
Adapter configurations 99  
 
Speed & Duplex. The Speed & Duplex property 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
o
o
o
o
o
1 Gb Full Auto. Sets the speed to 1 Gbps and mode to Full-Duplex  
10 Mb Full. Sets the speed to 10 Mbps and mode to Full-Duplex  
10 Mb Half. Sets the speed to 10 Mbps and mode to Half-Duplex  
100 Mb Full. Sets the speed to 100 Mbps and mode to Full-Duplex  
100 Mb Half. Sets the speed to 100 Mbps and mode to Half-Duplex  
Auto. (Default) Sets the speed and mode for optimum network connection (recommended)  
NOTE: Auto is the recommended selection. It allows the adapter to dynamically detect the line  
speed and duplex mode of the network. Whenever the network capability changes, the  
adapter will automatically detect and adjust to the new line speed 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 forces the adapter to connect to the network in Full-Duplex mode. The  
adapter may not function if the network is not configured to operate at the same mode.  
Wake Up Capabilities. This property is only available through the HP Network Configuration Utility.  
The Wake Up Capabilities property allows 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 list on the Advanced tab.  
o
o
o
o
Both. (Default) Selects both Magic Packet and Wake up Frame as wake up frames  
Magic Packet. Selects Magic Packet as the wake up frame  
None. Selects no wake up frames  
Wake Up Frame. Selects Wake up Frame as the wake up frame  
WOL Speed. This property is only available through the HP Network Configuration Utility. The WOL  
Speed property allows the user to select the speed at which the adapter connects to the network  
while in Wake-on-LAN mode. By default, the adapter is set to Auto.  
o
o
o
10 Mb. Sets the speed to 10 Mb  
100 Mb. Sets the speed to 100 Mb  
Auto. (Default) Sets the speed for optimum network connection  
Number of Transmit Descriptors: Minimum/Maximum. Specifies the number of 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 allocated).This property is only  
available through the HP Network Configuration Utility.  
o
o
Default = 200  
Range = 100–600  
Number of Receive Descriptors: Minimum/Maximum. Specifies the number of descriptors used by the  
driver when copying data to the protocol memory. In high network load situations, increasing  
receive descriptors can increase performance. The tradeoff is that this also increases the amount of  
system memory used by the driver. If too few receive descriptors are used, performance suffers. If too  
Adapter configurations 100  
 
many receive descriptors are used, the driver unnecessarily consumes memory resources. This  
property is only available through the HP Network Configuration Utility.  
o
o
Default = 200  
Range = 100–512  
Multifunction Gigabit Server adapter properties  
SRB Timeout Delta. Increases or decreases the timeout value for I/O operations through the iSCSI  
HBA. Increasing this value may help prevent premature I/O timeouts if your 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 Gigabit Server Adapters and Multifunction Server Adapters for HP  
BladeSystem running Windows 2000, Windows Server 2003, and Windows Server 2003 x64.  
Receive-Side Scaling (RSS). Allows the network load from a network 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 Multifunction Server Adapter. (Microsoft  
Scalable Networking Pack (SNP) for Windows 2003 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 Windows Server 2003 Scalable Networking Pack (SNP) and the  
required Microsoft hotfixes in article numbers 921136, 919948, and 923187 of the Microsoft  
Knowledge Base (KB).  
Windows 2000 adapter properties  
802.1p QoS. The 802.1p QoS parameter is a standard that 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 parameters, the checksum can be computed by the Gigabit  
Ethernet Adapter.  
To enable one of the Checksum Offload parameters, select the parameter.  
o
o
o
o
None. Disables checksum offloading  
Rx TCP/IP Checksum. Enables receive TCP, IP, and UDP checksum offloading  
Tx TCP/IP Checksum. Enables transmit TCP, IP, and UDP checksum offloading  
Tx/Rx TCP/IP Checksum. (Default) Enables transmit and receive TCP, IP, and UDP checksum  
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 enabled. To disable the Ethernet@WireSpeed parameter, select Disable  
from the Value list on the Advanced tab.  
o
o
Disable. Disables Ethernet@WireSpeed  
Enable. (Default) Enables Ethernet@WireSpeed  
NOTE: This property is not supported by NC1020 and NC7761 adapters.  
Flow Control. The Flow Control property allows the user to enable or disable the receipt or  
transmission of PAUSE frames. PAUSE frames enable the adapter and the switch to control the  
transmit rate. The side that is receiving the PAUSE frame will momentarily stop transmitting. The  
Adapter configurations 101  
 
recommended selection is Auto and it is the default setting. To disable Flow Control, select Disable  
from the Value list on the Advanced tab.  
o
o
o
o
o
Auto. (Default) Auto detect flow control  
Disable. PAUSE frame receipt and transmission is disabled  
Tx PAUSE. PAUSE frame transmission is enabled  
Rx PAUSE. PAUSE frame receipt is enabled  
Rx/Tx PAUSE. PAUSE frame receipt 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
o
o
o
o
10 Mb Full. Sets the speed to 10 Mbps and mode to Full-Duplex  
10 Mb Half. Sets the speed to 10 Mbps and mode to Half-Duplex  
100 Mb Full. Sets the speed to 100 Mbps and mode to Full-Duplex  
100 Mb Half. Sets the speed to 100 Mbps and mode to Half-Duplex  
Auto. (Default) Sets the speed and mode for optimum network connection (recommended)  
NOTE: 1 Gbps speed is enabled by selecting Auto.  
NOTE: Auto is the recommended selection. It allows the adapter to dynamically detect the line  
speed and duplex mode of the network. Whenever the network capability changes, the  
adapter will automatically detect and adjust to the new line speed 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 forces the adapter to connect to the network in Full-Duplex mode. The  
adapter may not function if the network is not configured to operate at the same mode.  
Wake Up Capabilities. The Wake Up Capabilities parameter allows 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
o
o
o
Both. (Default) Selects both Magic Packet and Wake up Frame as wake up frames  
Magic Packet. Selects Magic Packet as the wake up frame  
None. Selects no wake up frames  
Wake Up Frame. Selects Wake up Frame as the wake up frame  
NOTE: This property is not supported by NC150T adapters.  
WOL Speed. The WOL Speed parameter allows the user to select the speed at which the adapter  
connects to the network while in Wake-on-LAN mode. By default, the adapter is set to Auto.  
o
o
10 Mb. Sets the speed to 10 Mb  
100 Mb. Sets the speed to 100 Mb  
Adapter configurations 102  
 
o
Auto. (Default) Sets the speed for 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 timeout value for I/O operations through the iSCSI  
HBA. Increasing this value may help prevent premature I/O timeouts if your 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 cases, you may change any of the available  
options to meet the requirements of your specific system. Use the following 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.  
3.  
Select the Hardware tab and click Device Manager. The Device Manager window appears.  
Scroll down the list of hardware devices to Network Adapters and click the plus (+). This displays the  
list of adapters currently configured.  
4.  
5.  
6.  
7.  
Right-click the adapter to configure and select Properties. The Properties window appears.  
Select the Advanced tab. The list of properties and current values for the adapter appears.  
Select the property to modify and adjust the Value as needed.  
Save the settings.  
a. When the adapter property configuration 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 registers.  
Adapter properties  
802.1p QoS Packet Tagging. Enables or disables IEEE 802.1p tagging to send network traffic with  
different priority levels.  
o
o
Default = Disabled  
Range = Disabled; Enabled  
Number of Coalesce Buffers. Specifies the number of memory buffers 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 packet for later transmission. The preferred method of transmitting data is to use map registers,  
since it is the most efficient method.  
o
o
Default = 128 (for NC110x, NC360x, NC364x = 256)  
Range = 16–768 (increments of 8)  
Flow Control. The Flow Control property allows the user to enable or disable the receipt or  
transmission of PAUSE frames. PAUSE frames enable the adapter and the switch to control the  
transmit rate. The side that is receiving the PAUSE frame will momentarily stop transmitting. The  
Adapter configurations 103  
 
   
recommended selection is Auto and it is the default setting. To disable Flow Control, select Disable  
from the Value list on the Advanced tab.  
o
o
o
o
Auto. (Default) Auto detect flow control  
Disable. PAUSE frame receipt and transmission is disabled  
Tx PAUSE. PAUSE frame transmission is enabled  
Rx PAUSE. PAUSE frame receipt is enabled  
Rx/Tx PAUSE. PAUSE frame receipt 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 connection speed to the network and mode. Full-  
Duplex Mode allows the adapter to transmit and receive network data simultaneously. This property  
is not available for NC310F and NC61xx adapters through the Microsoft® User Interface.  
o
o
Default = Auto Detect  
Range = NC340x and NC71xx Adapters: Auto Detect; 10Mbps/Half-Duplex; 10Mbps/Full-  
Duplex; 100Mbps/Half-Duplex; 100Mbps/Full-Duplex; 1000 Mbps/Full-Duplex  
Locally Administered Address. Specifies the user-defined MAC Address of the adapter, which  
overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed.  
o
o
Default = Not Present  
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 load situations, increasing receive descriptors can  
increase performance. The tradeoff is that this also increases the amount of system memory used by  
the driver. If too few receive buffers are used, performance suffers. If too many receive buffers are  
used, the driver unnecessarily consumes memory resources.  
o
Default  
NC110x, NC310F, NC340x, NC360x, NC364x, NC6170, and NC7170 = 256  
All others = 160  
o
Range  
NC6132, NC6133, and NC6134 = 80–768 (increments of 8)  
NC6136, NC7131, and NC7132 = 80–256 (increments of 8)  
NC310F, NC340x, NC6170, and NC7170 = 80–4096 (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
o
Default = 256  
Range  
NC6132, NC6133, and NC6134 = 80–768 (increments of 8)  
NC6136, NC7131, and NC7132 = 80–256 (increments of 8)  
NC310F, NC340x, NC6170, and NC7170 = 80–4096 (increments of 8)  
NC110x, NC360x, and NC364x = 80-2048 (increments of 8)  
Adapter configurations 104  
 
Offload Receive TCP Checksum. Offloads the task of computing the checksum for incoming TCP or  
UDP packets, thereby improving performance.  
o
o
Default = On  
Range = Off; On  
Offload Transmit TCP Checksum. Offloads the task of computing the checksum for outgoing TCP or  
UDP packets, thereby improving performance.  
o
o
Default = On  
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
o
Default = On  
Range = Off; On  
Offload Receive IP Checksum. (NC110x, NC310F, NC340x, NC360x, NC364x, NC6170, and  
NC7170 only). Offloads the task of computing the checksum for incoming IP packets, thereby  
improving performance.  
o
o
Default = On (NC110, NC360x, NC364x = Off)  
Range = Off; On  
Enable PME (NC110x, NC310F, NC340x, NC360x, NC364x, NC6170, and NC7170 only).  
Allows the use of power management and wake-up functions.  
o
o
Default = No Action (NC110x, NC360x, NC364x = OS Controlled)  
Range = Disabled; Enabled; Hardware Default; No Action (NC110x, NC360x, NC364x = OS  
Controlled, Disabled, Enabled)  
Wake On Settings (NC110x, NC310F, NC340x, NC360x, NC364x, NC6170, and NC7170  
only). Available only when Enable PME is enabled. Allows you to choose what types of packets will  
cause the adapter to wake.  
o
o
Default = Disabled (NC110x, NC360x, NC364x = OS Controlled)  
Range = Disabled; Wake on Directed Packet; Wake on Magic & Directed; Wake on Magic  
Packet; OS Controlled  
Large Send Offload (Windows® Server 2003 only; NC110x, NC310F, NC340T, NC360x,  
NC364x, NC6170, and NC7170 only). Enables offloading of large TCP packets.  
o
o
Default = On  
Range = Off; On  
Smart Power Down. (NC110T, NC360x, and NC364T only) Minimizes power consumption by  
enabling the adapter to enter a deep sleep mode under certain conditions.  
o
o
Default = Hardware Default  
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
o
Default = Disable  
Range = Disable, Force  
Adapter configurations 105  
 
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 following 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.  
3.  
Select the Hardware tab and click Device Manager. The Device Manager window appears.  
Scroll down the list of hardware devices to Network Adapters and click the plus (+). This displays the  
list of adapters currently configured.  
4.  
5.  
6.  
7.  
Right-click the adapter to configure and select Properties. The Properties window appears.  
Select the Advanced tab. The list of properties and current values for the adapter appears.  
Select the property to modify and adjust the Value as needed.  
Save the settings.  
a. When the adapter property configuration 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 registers.  
Adapter Properties  
802.1p QoS Packet Tagging. Enables or disables IEEE 802.1p tagging to send network traffic with  
different priority levels.  
o
o
Default = Off  
Range = Off; On  
Coalesce Buffers. Specifies the number of memory buffers 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  
packet for later transmission. The preferred method of transmitting data is to use map registers, since  
it is the most efficient method.  
o
o
Default = 8  
Range = 1–32  
Flow Control Settings. The Flow Control property allows the user to enable or disable the receipt or  
transmission of PAUSE frames. PAUSE frames enable the adapter and the switch to control the  
transmit rate. The side that is receiving the PAUSE frame will momentarily stop transmitting. The  
recommended selection is Off and it is the default setting. To disable Flow Control, select Off from  
the Value list on the Advanced tab.  
o
o
Default = Off  
Range = Off, Generate, Respond, 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 receipt and transmission is enabled  
Adapter configurations 106  
 
   
Link Speed & Duplex. Allows the user to set the connection speed to the network and mode. Full-  
Duplex Mode allows the adapter to transmit and receive network data simultaneously.  
o
o
Default = Auto Detect  
Range  
100 Mbps/Full-Duplex: Sets the speed at 100 Mbps and mode to Full-Duplex  
100 Mbps/Half-Duplex: Sets the speed at 100 Mbps and mode to Half-Duplex  
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 mode for optimum network connection  
Locally Administered Address. Specifies the user-defined MAC Address of the adapter; overrides the  
burned-in MAC Address. This is a read-only field if the adapter is teamed.  
o
o
Default = Not Present  
Range = Value; Not Present  
Receive Buffers. Specifies the number of buffers used by the driver when copying data to the protocol  
memory. In high network load situations, increasing receive buffers can increase performance. The  
tradeoff is that this also increases the amount of system memory used by the driver. If too few receive  
buffers are used, performance suffers. If too many receive buffers are used, the driver unnecessarily  
consumes memory resources.  
o
o
Default = 48  
Range = 8–1024 (increments of 1)  
Transmit Control Blocks. Specifies the number of resources that are allocated to transmit packets.  
o
o
Default = 16  
Range = 1–64  
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 Copper 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
o
Default = Auto/Auto  
Range = 10/Half; 10/Full; 100/Half; 100/Full; Auto/Auto  
NOTE: The NC3133 is a 100BASE-FX fiber optic upgrade module that supports 100 Mbps  
only.  
Locally Administered Address. Specifies the user-defined MAC Address of the adapter, which  
overrides the burned-in MAC Address. This box is disabled if the selected adapter has been teamed.  
o
o
Default = NULL  
Range = Valid unicast MAC Address  
Adapter configurations 107  
 
 
Adaptive Interframe Spacing. Compensates for excessive Ethernet packet collisions on the network.  
The default setting works best for most computers and networks by dynamically adapting to the  
network traffic conditions. However, in some rare cases, you may obtain better performance by  
manually setting the spacing value. Setting a value forces a static gap between packets. Increasing  
the value increases the delay between frames being transmitted.  
o
o
Default = 1  
Range = 1–255  
Number of Coalesce Buffers. Specifies the number of memory buffers 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 packet for later transmission. The preferred method of transmitting data is to use map registers,  
since it is the most efficient method.  
o
o
Default = 8  
Range = 1–32  
Number of Receive Descriptors: Minimum/Maximum. Specifies the number of descriptors used by the  
driver when copying data to the protocol memory. In high network load situations, increasing  
receive descriptors can increase performance. The tradeoff is that this also increases the amount of  
system memory used by the driver. If too few receive descriptors are used, performance suffers. If too  
many receive descriptors are used, the driver unnecessarily consumes memory resources.  
o
o
Default = 32  
Range = 8–1024 (in increments of 8)  
Number of Transmit Descriptors: Minimum/Maximum. Specifies the number of 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 allocated).  
o
o
Default = 12  
Range = 1–16  
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 many TCBs are used, the driver  
unnecessarily consumes memory resources.  
o
o
Default = 16  
Range = 1–80  
Transmit Threshold. Specifies the number of bytes before the PCI adapter empties 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 not attempt early transmits. Although this is the  
safest setting, the best performance is achieved when the Threshold parameter is as low as possible  
without producing underruns.  
o
o
Default = 200  
Range = 0–200  
Properties for NC61xx and NC71xx Server Adapters  
Exception: HP NC6170 and NC7170 Server Adapters are not supported under Windows NT® 4.  
Adapter configurations 108  
 
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/Full  
10/100/100 Adapters: Auto/Auto  
o
Range  
1000BASE-SX or -LX Adapters: 1000/Full  
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 putting more data in each packet, thereby sending  
out fewer packets. This is equivalent to Jumbo Frames and must comply with switch limitations.  
o
o
Default = 1514 Bytes  
Range = 1514 Bytes; 4088 Bytes; 9014 Bytes; 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 setting applies for  
Fiber Gigabit Ethernet server adapters only.  
Number of Coalesce Buffers. Specifies the number of memory buffers 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 packet for later transmission. The preferred method of transmitting data is to use map registers,  
since it is the most efficient method.  
o
o
Default = 128  
Range = 16–768 (increments of 8)  
Number of Receive Descriptors: Minimum/Maximum. Specifies the number of descriptors used by the  
driver when copying data to the protocol memory. In high network load situations, increasing  
receive descriptors can increase performance. The tradeoff is that this also increases the amount of  
system memory used by the driver. If too few receive descriptors are used, performance suffers. If too  
many receive descriptors are used, the driver unnecessarily consumes memory resources.  
o
o
Default = 160  
Range  
NC6132, NC6133, and NC6134: 64–768 (increments of 8)  
NC6136, NC7131, and NC7132: 64–256 (increments of 8)  
Number of Transmit Descriptors: Minimum/Maximum. Specifies the number of 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 allocated).  
o
o
Default = 256  
Range  
NC6132, NC6133, and NC6134: 64–768 (increments of 8)  
NC6136, NC7131, and NC7132: 64–256 (increments of 8)  
Adapter configurations 109  
 
Properties for NC67xx 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
o
Default = Auto/Auto  
Range = Auto/Auto; 10/Half; 10/Full; 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 putting more data in each packet, thereby sending  
out fewer packets. This is equivalent to Jumbo Frames and must comply with switch limitations.  
o
o
Default = 1514 Bytes  
Range = 1514 Bytes; 4088 Bytes; 9014 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 setting applies for  
Fiber Gigabit Ethernet server adapters only.  
Flow Control. Specifies the type of flow control that is needed.  
o
o
Default = Disable  
Range = Disable; Rx Pause; Tx Pause; Rx/Tx Pause; Auto  
Novell Netware  
Advanced Server Program for NetWare  
The following describes how to install and configure the 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 virtual slot. The virtual slot can be viewed as a team  
number.  
NOTE: Before loading the network drivers, the recommended sequence is to load QASP.LAN.  
This allows QASP.LAN to determine the initial link 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 110  
 
     
1.  
2.  
Load the network drivers for the network adapters that 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.  
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.  
4.  
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.  
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 virtual and the physical adapters, it is only necessary  
to bind one frame type on the virtual adapter to the same frame type on the physical 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.LAN 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) enter 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-Balance 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 111  
 
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 virtual adapter, additional frame types can only be  
loaded in the virtual adapter if the corresponding frame types are also loaded in the bound adapters. For  
example, ETHERNET_802.2 can be loaded in QASP VSLOT 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 physical 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 each 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 types specified in step 1.  
One or more VLAN IDs specified in step 1 do not 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 times 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 112  
 
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 using 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. Use decimal numbers to specify the VLAN ID.  
The following are examples of multiple VLAN configurations:  
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 executed, the adapter Q57_1_EII is bound to all  
three VLANs. If you are unable to log in to the server after configuring ASP, add the following  
command lines before loading ASP:  
UNLOAD SLPTCP  
LOAD SLPTCP  
Jumbo Frames  
Jumbo Frames are supported in all balance modes. The maximum frame size will be automatically set to  
the smallest maximum frame size of all adapters in the team. Use appropriate keywords to enable 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 improves performance if some or all adapters in the  
Adapter configurations 113  
 
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 interface 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 Registration 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 FRAME=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 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, add or delete parameters, or modify parameters.  
A valid AUTOEXEC.NCF file is shown below with 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 frame type Ethernet_II and three VLANs, number 2,  
3, and 4.  
Adapter configurations 114  
 
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 automatically # 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 not 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 bound 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 115  
 
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 116  
 
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 also bound to the other  
VLANs on the same VSLOT. VLAN refers to VLAN ID, and a valid 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 Trunking (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 advertised to the switch using GVRP. Notice that  
untagged 802.2 frames must be loaded to allow sending 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 117  
 
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 switch to support 802.3ad with LACP (Link  
Aggregation Control Protocol).  
Configuration parameters for Q57.LAN driver  
Configuration parameters for Q57.LAN driver include:  
TxDescriptors=  
o
o
o
o
This keyword is to initialize Descriptor resources on the adapter for transmits.  
Recommended Min = 100  
Max = 512  
Default = 200  
NOTE: Min = 1 when used with DOS Client32 and when Keyword P3=1 or 2.  
RxBuffers=  
o
This keyword is to pre-allocate receive ECBs & Receive adapter resources. This setting may be  
affected by the NetWare server maximum/minimum packet receive buffer settings.  
o
o
o
Recommended Min = 32  
Max = 512  
Default = 200  
NOTE: Min = 1 when used with DOS Client32 and when Keyword 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 forced. 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 for automatic link setup  
(auto-negotiation).  
o
o
Choices are: AUTO, 10FD,10HD, 100FD, 100HD  
Default is Auto. (Note that 1000 FD is only auto-detected 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
o
Choices are: AUTO, FORCE  
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 118  
 
o
o
o
o
o
This keyword is to enable the use of batching receives within a specific time period.  
Min = 0, disabled  
Max = 5000000, 5 seconds  
Units are in micro seconds  
Default = 360  
TxPacketsPer=  
o
o
o
o
This keyword is to enable the use of batching transmits to a specific amount of packets.  
Min = 0, disabled  
Max = 100  
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
o
Choices are: ON, OFF, TX, RX  
Default is ON  
TxFlow=  
o
o
o
This keyword allows enabling/disabling of TxFlow control.  
Choices are: ON, OFF  
Default is OFF  
RxFlow=  
o
o
o
This keyword allows enabling/disabling of RxFlow control.  
Choices are: ON, OFF  
Default value is OFF  
VLAN=  
o
This keyword allows enabling/disabling 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
o
Choices are: 0 (Disabled), 1 (Enabled)  
Default value is 0  
Jumbo=  
o
This keyword enables Jumbo frame support. When enabled, jumbo packets of up to 9000 bytes  
are supported.  
o
o
Jumbo=1514-9000  
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
o
Choices are: ON, OFF  
Default value is OFF  
Adapter configurations 119  
 
NOTE: If you modify any adapter parameters, you must reboot the system for the changes to  
take effect. If you make changes and do not reboot, you may experience configuration  
problems.  
NetWare Client 32 driver installation for HP adapters  
The NC10032.LAN driver is a high-performance server driver that meets the Open Data Link Interface  
(ODI) v1.11 C specification. It can be used on Client 32 workstations with NetWare v3.X servers and  
greater.  
Installing Client 32 on DOS  
1.  
2.  
3.  
Install the driver and its support files for DOS.  
From the root directory, run MAKEDISK.BAT to create driver diskettes.  
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 following CD directory: \PRODUCTS\DOSWIN32.  
5.  
6.  
7.  
Follow the Novell prompts.  
For TCP/IP, select TCP/IP protocol stack at the Options/Product screen.  
At the 32-bit LAN Driver screen, scroll to the end of listed drivers and select User Specified 32-bit  
driver.  
8.  
9.  
Insert the HP driver diskette.  
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 adapter auto-negotiates the duplex with the  
hub/switch. If the adapter and hub/switch are unable to auto-negotiate, HP adapters default to half-  
duplex. Full-duplex allows for simultaneous transmit and receive data transfer. Full-duplex can be  
used only when the adapter is connected to a hub that supports and is enabled both for full-duplex  
mode and for the specified speed.  
Adapter configurations 120  
 
   
o
Syntax: FORCEDUPLEX=n  
Where n=  
0–auto-negotiate  
1–half-duplex  
2–full-duplex  
NOTE: To set the adapter to half- or full-duplex, you must set the SPEED parameter to either 10  
or 100.  
o
o
Default: auto-negotiation  
Examples:  
100 Mbps full-duplex: FORCEDUPLEX=2 SPEED=100  
10 Mbps full-duplex: FORCEDUPLEX=2 SPEED=10  
SPEED. If no speed setting is specified, the adapter auto-negotiates the speed with the hub/switch,  
either 10 or 100 Mbps. If the adapter and hub/switch are unable to auto-negotiate, the adapter  
defaults to 10 Mbps.  
o
o
Syntax: SPEED=n (n = 10 or 100)  
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 121  
 
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 parameters in this file.  
The parameters include the following.  
ForceSpeedDuplex. Configures link (or instance) to a certain Speed and Duplex. By default,  
AutoNegotiate (0) is set. The setup is based on the following values:  
0—AutoNegotiate  
1—10 Mbps speed and Half-Duplex mode  
2—10 Mbps speed and Full-Duplex mode  
3—100 Mbps speed and Half-Duplex mode  
4—100 Mbps speed and Full-Duplex mode  
5—Force 1000 Mbps Full-Duplex mode (Fiber adapter only)  
6—AutoNegotiate only 1000 Mbps Full-Duplex mode  
7—AutoNegotiate only 1000 Mbps Half-Duplex mode  
8—AutoNegotiate only 100 Mbps Full-Duplex mode  
9—AutoNegotiate only 100 Mbps Half-Duplex mode  
10—AutoNegotiate only 10 Mbps Full-Duplex mode  
11—AutoNegotiate only 10 Mbps Half-Duplex mode  
Example: Configure adapters of instance#0 and instance#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 parameters of a link. The setup is based on the following  
values:  
0—Both Tx and Rx flow control are disabled.  
1—Tx flow control is enabled. Pause frames will be sent if resource is low, but the device will not  
process Rx Pause Frame.  
2—Rx flow control is enabled. If the device receives Pause Frame, it will stop sending, but the  
device will not send Pause Frame if resource is low.  
3—Both Rx and TX flow control are enabled. Pause frames will be sent if resource is low. If the  
device receives Pause Frame, it will stop sending.  
Adapter configurations 122  
 
 
4—Advertise both Rx and TX flow control being enable and negotiate with link partner. If link  
AutoNegotiate is not enabled, then both Tx & Rx Flow Control are disabled.  
Example: FlowControl=0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0;  
MaxJumboFrameSize. Configures 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 feature is  
disabled. BCM5705 adapters do not support jumbo frame 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 descriptors. The valid range is 32 to 512.  
More system memory resources are used for larger number of Tx Packet 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 resources are used for larger number of Rx Packet Descriptors. The  
default value is 500.  
Example: RxStdDescCnt=500;  
RxJumboDescCnt. Configures the number of Rx Jumbo packet descriptors. The valid value is 32 to  
256. More system memory resources are used for larger 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 Coalescing Ticks in microseconds. This  
determines the upper-boundary of the time interval that the device will generate interrupts if one or  
more frames are received. The default value is 150.  
Example: RxCoalescingTicks=150;  
RxMaxCoalescedFrames. Configures the number of Rx Maximum Coalesced Frames 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 Coalescing Ticks in microseconds. This  
determines the upper-boundary of the time interval that 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 Maximum 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 of Rx Host Coalescing Ticks in microseconds  
during interrupt. This determines the upper-boundary of the time interval that the device will generate  
interrupts if one or more frames are received during interrupt handling. The default value is 75.  
Example: RxCoalescingTicksDuringInt=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 generate  
interrupts if one or more frames are received during interrupt handling. The default value is 75.  
Adapter configurations 123  
 
Example: TxCoalescingTicksDuringInt=75;  
RxMaxCoalescedFramesDuringInt. Configures the number of Rx Maximum Coalesced Frames  
parameters during interrupt handling. This determines 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 number of Tx Maximum Coalesced Frames  
parameters during interrupt handling. This determines the upper-boundary of maximum number of Tx  
buffer descriptors that device processes before it will generate an interrupt during interrupt handling.  
The default value is 10.  
Example: TxMaxCoalescedFramesDuringInt=10;  
StatsCoalescingTicks. Configures how often adapter statistics 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 reboot, modify bcme.conf instead.  
To display parameters that are configurable via ndd:  
ndd /dev/bcme '?'  
The system should return 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 parameter must be set to the instance associated with  
a particular adapter. For example, to force adapter of instance 1 to 100 Mbps Full-Duplex:  
ndd -set /dev/bcme Instance 1  
ndd -set /dev/bcme ForceSpeedDuplex 3  
Adapter configurations 124  
 
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  
Adapter configurations 125  
 
Team properties  
In this section  
Adapter teaming.................................................................................................................................. 126  
Windows 2000 and Windows NT4 teaming .......................................................................................... 130  
NetWare teaming................................................................................................................................ 139  
Adapter teaming  
Adapters must all be on the same Layer 2 (Broadcast) network to be teamed.  
The maximum configuration is eight teams with eight adapters in each 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 Ethernet and Gigabit Server Adapters  
NOTE: Dual Channel Network Fault Tolerance teaming capabilities are not supported on  
systems running Windows® 2000.  
Automatic (Recommended). Default setting. The system automatically selects the best team type or  
combination of team features based on the configuration 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 Load 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 load balanced among the remaining adapters.  
Currently, Automatic is supported in Windows® 2000, Windows Server 2003, and Windows  
Server 2003 x64.  
802.3ad Dynamic Dual Channel Load Balancing (INP).  
ProLiant Essentials Intelligent Networking 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 Control Protocol (LACP) agreement with the switch. A  
failure of a team member to synchronize joining a port-trunk or channel with the switch is treated as  
an error and that team member is considered failed. All transmit packets are load balanced among  
all team members based on a Load 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 load balanced among the remaining adapters.  
802.3ad Dynamic Dual Channel Network Fault Tolerance (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. All team members within each group are  
dynamically placed into a port-trunk/channel by dynamic LACP agreement with the switch. A failure  
of a team member to synchronize joining a port-trunk or channel with the switch is treated as an  
error and that team member is considered failed. Dynamic Dual Channel Network Fault Tolerance  
Team properties 126  
 
     
(NFT) provides the safety of additional backup links between the server and hub/switch. Dynamic  
Dual Channel NFT is implemented with one channel team providing a primary adapter and another  
channel team providing a secondary, backup adapter(s). During normal operations, if the adapters  
in the primary team fail, a link to one of the adapters 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. Provides the same options 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 group to a second switch.  
All transmit packets are load balanced among all team members based on a Load 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 load balanced  
among the remaining adapters. Only Gigabit adapters can be used for Switch-assisted Dual  
Channel Load Balancing. When you select Switch-assisted Dual Channel Load Balancing, the Team  
Members window becomes Team Members Grouping and displays two windows labeled 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 INP  
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 between 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 ProLiant Essentials  
INP only. Provides the same options as Switch-assisted Dual Channel Network Fault Tolerance  
except it offers the additional option 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 members 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 the switch is treated as an error and that team  
member is considered failed. All transmit packets are load balanced among all team members  
based on a Load 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 load balanced among the remaining adapters. If a team contains an iSCSI-enabled  
adapter, you cannot use the 802.3ad Dynamic 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 load balances all traffic. Unlike TLB, SLB can be  
configured to increase both transmitting and receiving channels between your server 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 and an aggregate maximum  
receive rate of 400 Mbps, resulting in a total bandwidth of 800 Mbps.  
Team properties 127  
 
SLB works only with SLB-enabled switches, such as the Catalyst 6000 series. The SLB software  
continuously analyzes loading on each adapter and balances network traffic across the adapters as  
needed. Adapter teams configured for SLB also provide the benefits of NFT.  
To use SLB, you must have at least two HP Fast Ethernet 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 works with any  
industry-standard 10BASE-T, 100BASE-TX, or Gigabit 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-Primary 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 packets are load balanced among all remaining team  
members.  
To use TLB, you must have at least two NCxxxx adapters (ports) capable of the same speed and  
linked to the same Layer 2 (Broadcast) switch installed 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 single 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 Mbps and a 100  
Mbps receive rate resulting in a total bandwidth of 500 Mbps.  
TLB load balances IP traffic under Windows 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 transmits 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 adapters, the packets are load balanced among all  
remaining team members. 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.  
Network Fault Tolerance Only (NFT). NFT provides the safety of an additional backup link 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.  
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. 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. 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® 2000, Windows Server 2003, and Windows Server  
2003 x64, Windows NT® 4.0, NetWare 6.x, and UnixWare 7.x.  
Transmit Balancing Method  
Team properties 128  
 
Select or change the Transmit Balancing Method for the team. If the team type is Network Fault Tolerance  
Only with Preference Order or Network Fault Tolerance Only (NFT), the Transmit Balancing Algorithm list  
disabled and Automatic is selected.  
o
Automatic. (Recommended) Default setting. Teaming driver selects the load balancing mechanism  
based on the packet type.  
o
o
TCP Connection. Load balances transmit TCP packets using the TCP connection.  
Destination IP Address. Load balances transmit IP packets using the last four bits of the  
destination IP Address.  
o
o
Destination MAC Address. Load balances transmit IP packets using the last four bits of the  
destination MAC Address.  
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 rotation begins again.  
UnixWare 7.x and OpenServer 6.x teaming notes  
UnixWare 7.x and OpenServer 6.x operating systems support Network Fault Tolerance (NFT) but not  
Transmit Load Balancing (TLB) or Switch-Assisted 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-enabled 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 something else other 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 for 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, the team is created with TOE disabled.  
If TOE is disabled on a team of TOE-capable adapters, enabling TOE 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 129  
 
   
If the new team includes at least one TOE-enabled adapter and at least one non-TOE capable  
adapter, the team is created with TOE disabled and TOE is disabled on all of the individual teamed  
adapters.  
If a non-TOE capable adapter is added to a TOE-enabled team, the team is changed to TOE  
disabled. TOE functionality for each teamed adapter is disabled.  
Disabling or dissolving a TOE-enabled team:  
If TOE teaming is disabled on an existing TOE-enabled team, TOE is disabled on all of the teamed  
adapters.  
If a team is dissolved, the TOE functionality for each 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, the 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-capable and non-RSS capable adapters:  
If the new team includes at least one RSS-enabled adapter and at least one non-RSS capable  
adapter, the team is created with RSS-disabled and RSS functionality for each adapter that is in the  
team is disabled.  
If a non-RSS capable adapter is added to an RSS-enabled team, the team is changed to RSS-  
disabled and RSS functionality for each adapter 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 functionality 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 other 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 drivers.  
Upgrading teams to Windows® 2000  
The Windows® 2000 Primer Utility from HP lets you remove unsupported Windows® 2000 utilities  
before or after upgrading your OS. These unsupported utilities were designed to run on Windows NT®  
4.0 and may have been installed 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 Upgrade Utility installed  
from HP SmartStart  
Team properties 130  
 
 
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 operating systems:  
Microsoft® Windows NT® 4.0  
Microsoft® Windows® 2000  
For PRIMER.EXE, refer to the HP Softpaq #10652 at the HP support website  
Before running HP Primer Utility  
Before running the HP Primer Utility, you should be aware of the following:  
Retaining Teaming Configuration Settings in Windows NT® 4.0  
HP recommends that you record all Windows NT® 4.0 teaming configuration settings. After you run  
Primer and upgrade the OS to Windows® 2000, you can install the new HP Network Configuration  
Utility for Windows® 2000 and reconfigure the adapter teams using the saved settings.  
Windows NT® 4.0 components removed by Primer  
The Primer utility removes the installed files, the program links, and the registry entries for the  
previously installed components. In addition, general cleanup is performed to remove program links  
and common files that may not belong to any specific component.  
Adapter configuration components removed by Primer  
o
o
o
HP Advanced Network Control Utility  
HP Network Configuration Utility  
HP TLAN Network Teaming and Configuration Utility  
When Primer removes these network utilities, 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
o
o
o
o
o
o
o
o
o
o
o
o
o
o
HP Array Configuration Utility  
HP Cluster Verification Utility  
HP Insight Management Foundation Agents  
HP Insight Management Server Agents  
HP Insight Management Storage Agents  
HP Integrated Management Display Utility  
HP Integrated Management Log Viewer Utility  
HP Integration Maintenance Utility  
HP NT Management  
HP Online Storage Controller Recovery Utility  
HP PCI Hot Plug Utility  
HP Power Down Manager  
HP Power Supply Viewer  
HP Remote Monitor Service  
HP System Management Services  
Team properties 131  
 
o
HP System Partition Upgrade Utility  
Using HP Primer Utility from the graphical user interface  
To remove the Windows NT® components either before or after upgrading to Windows® 2000, follow  
these steps:  
1.  
2.  
3.  
4.  
From the location where you have downloaded Primer, double-click PRIMER.EXE.  
Click CLEANUP. The utility will now remove the previously-installed Windows NT® components.  
Exit the Windows® 2000 Primer Utility from HP.  
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 the cleanup without using the graphical user interface by  
following these steps:  
1.  
2.  
3.  
Open the MS-DOS command prompt.  
Go to the directory where the Windows® 2000 Primer Utility from HP, PRIMER.EXE, is located.  
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 install using "primer s," a log file is  
generated that allows you to view the uninstalled components 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 the current working directory (where the utility is run). However, if the  
media from which you are running the utility is write-protected, the log file will reside in the %systemdir%  
directory.  
NOTE: You must have Administrator privileges to use the Log.  
Teaming considerations when upgrading to Windows® 2000  
Under Windows NT® 4.0, adapter teaming may be managed using CPQNTAC (HP Network  
Configuration Utility) for HP NCxxxx Ethernet, Fast Ethernet, and Gigabit adapters  
Because different utilities manage teaming under Windows 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 Intermediate drivers that do not work with  
Windows® 2000. For a Windows® 2000 upgrade, you 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 removes the installed files, program links, registry entries, and  
generally cleans files that may not belong to any specific component. For further information about  
Primer, refer to HP Softpaq #10652 provided at HP support website  
2.  
3.  
Upgrade the OS to Windows® 2000.  
Install the new HP Network Configuration Utility (NCU) for Windows® 2000.  
Team properties 132  
 
Method 2: Manual Removal  
1.  
2.  
3.  
Dissolve all teams and controller-pairs with their respective utilities.  
Upgrade the OS to Windows® 2000.  
Install the new HP Network Configuration Utility (NCU) for Windows® 2000.  
Installing the teaming driver  
NOTE: Before installing the Teaming protocol (Teaming driver), you may want to back up the  
Windows® 2000 Registry.  
To install the teaming driver:  
1.  
2.  
3.  
4.  
5.  
6.  
Click Software & Driver Downloads from the left menu bar.  
Type the product name in the For product box and press Enter. For example, type NC370T.  
Select an operating system.  
Click HP ProLiant Networking Software.  
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.  
8.  
9.  
Click the SoftPaq file to extract the files and then open the cmponent.htm file.  
Select the win2000, win2003 or win2003x64 folder that is appropriate for your operating system.  
Run the HP ProLiant Network Configuration Utility Component Package (cpxxxxxx.exe file). The  
component is a self-extracting executable with a file name based on the component 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 SoftPaq SP24889 available at the HP support website  
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
o
Double-click the HP Network tray icon.  
Double-click the HP Network icon in the Control Panel window.  
2.  
3.  
4.  
5.  
Select the adapters to be teamed in the NCU Main window.  
Click Team.  
Configure the team as needed and click OK.  
Click Yes to confirm the configuration changes.  
Team properties 133  
 
6.  
7.  
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.  
Restart the HP NCU and verify the adapter protocols were configured properly. Select the adapter  
and click Properties.  
Uninstalling teamed adapters  
If you need to completely uninstall a teamed adapter (not simply replace it), perform these steps to ensure  
proper operations:  
1.  
2.  
3.  
Select the Team and click Dissolve.  
Click OK on the NCU Main window to accept the changes.  
Physically remove the adapter from the server.  
Configuring HP adapters  
The Properties dialog for the HP Network Team Local Area Connection contains the following tabs and  
buttons.  
General Tab  
Connect using—HP Network Team. Components that are selected are used by this connection. These  
include:  
o
o
o
o
o
o
o
Client for Microsoft® Networks  
HP Network Configuration Utility  
Network Load Balancing  
File and Printer Sharing for Microsoft® Networks  
Network Monitor Driver  
Net BEUI Protocol  
Internet Protocol (TCP/IP)  
NOTE: Other protocols may apply.  
Description—Describes the selected component  
Show icon in taskbar when connected—Enables/Disables the HP connection tray icon  
Buttons  
Configure—Displays Windows® 2000 device properties  
o
o
General (Troubleshooter)—Disable Device option  
Driver (Name, Version number, Driver Details; Install/Uninstall/Update driver buttons)  
Install—Installs the selected protocol  
Uninstall—Uninstalls the selected protocol  
Properties—Accesses dialogs on the protocol for the selected device.  
Sharing Tab  
Sharing allows other servers or your local network to access external resources through this connection.  
The options on the Sharing tab are:  
Team properties 134  
 
Shared Access (Enable)—If selected, it allows Internet connection sharing. You can share  
Applications and Protocols.  
Applications tab—Displays network applications to be enabled for servers sharing this connection.  
Services tab—Displays services to be provided 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—Name of the adapter or module connection. Components that are checked are used  
by this connection. These include:  
o
o
o
o
o
o
o
Client for Microsoft® Networks  
HP Network Configuration Utility  
Network Load Balancing  
File and Printer Sharing for Microsoft® Networks  
Network Monitor Driver  
Net BEUI Protocol  
Internet Protocol (TCP/IP)  
NOTE: Other protocols may apply.  
Description—Describes the selected component  
Show icon in taskbar when connected—Enables/Disables the HP tray icon  
Buttons  
Configure button—Windows® 2000 device properties device.  
General Tab  
o
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 135  
 
WakeUp Capabilities  
WOL Speed  
Driver Tab  
o
o
o
o
Name, Version number, Driver Details  
Install  
Uninstall  
Update driver  
Resources Tab  
o
o
Resource settings  
Conflicting device list  
Install button—Installs the selected protocol  
Uninstall button—Uninstalls the selected protocol  
Properties button—Accesses dialogs on the protocol for the selected device.  
Sharing Tab  
Sharing allows other servers or your local network to access external resources through this connection.  
The options on the Sharing tab are:  
Shared Access (Enable)—If selected it allows Internet connection sharing. You can share applications  
and protocols.  
o
o
Applications—Network applications to be enabled for servers sharing this connection  
Services—Services to be provided to the remote network  
Windows NT Teaming  
The Fast Ethernet and Gigabit Ethernet adapters provide the following options for increasing throughput  
and fault tolerance when running Windows NT® 4.0.  
Network Fault Tolerance (NFT). NFT provides the safety of an additional backup link 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 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-Primary 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 packets are load balanced among all remaining team members.  
To use TLB, you must have at least two NCxxxx adapters (ports) capable of the same speed and  
linked to the same Layer 2 (Broadcast) switch installed 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 single 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 Mbps and a 100  
Mbps receive rate resulting in a total bandwidth of 500 Mbps.  
Team properties 136  
 
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 throughput using Fast Ethernet or  
Gigabit Ethernet ports. SLB load balances all traffic. Unlike TLB, SLB can be configured to increase  
both transmitting and receiving channels between your server 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 and an aggregate maximum receive rate of 400  
Mbps, resulting in a total bandwidth of 800 Mbps.  
SLB works only with SLB-enabled switches, such as the Catalyst 5000 series. The SLB software  
continuously analyzes loading on each adapter and balances network traffic across the adapters as  
needed. Adapter teams configured for SLB also provide the benefits of NFT.  
To use SLB, you must have at least two HP Fast Ethernet 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 requires pairs of ports (two or  
four ports). Refer to the documentation that shipped with your SLB-capable 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 the 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 described in Windows NT® 4.0  
Installation Notes.  
2.  
After installation is complete, restart Windows NT® 4.0.  
NOTE: Windows NT® 4.0 Service Pack 5 or later is required for implementing adapter  
teaming properly. Install the service pack 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.  
2.  
Double-click the Network Icon in the Control Panel. This launches the Network Control Panel applet.  
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 configuration in Windows NT® 4.0 through CPQNTAC, you  
must be logged on to the system as an Adminstrator.  
Team properties 137  
 
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 adapters 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 condition. If this message is displayed, install the  
correct Service Pack before proceeding.  
4.  
5.  
The Teaming Property Sheet appears for the newly created team. The first page displayed is the  
Teaming Controls Sheet. The default mode of operation for a team is Fault Tolerant mode. Change  
this to the desired mode of operation by selecting either Fault Tolerant or Load Balancing at the top  
of the page.  
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 occur if  
the Current Primary adapter is dropped from the 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 adapter takes on the role of Current Primary  
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 Primary 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 Primary adapter is displayed in the Preferred Primary NIC list box. A  
single Fault Tolerant team CAN contain a mixture of Fast Ethernet and Gigabit adapters.  
6.  
7.  
For a Load Balancing team, select among the following modes:  
o
o
Transmit Load Balancing  
Switch-assisted Load Balancing  
Select the Settings tab at the top of the Teaming Property Sheet. On the Settings tab, you may  
configure the following items:  
Team Network Address. Used to "Locally Administer" the team's unique receive node address on the  
network like you would locally administer a single adapter's node address.  
Enable Heartbeats. This check box is used to enable/disable teaming Heartbeats. Heartbeats  
provide an additional level of adapter failure detection above simple link failure detection.  
Heartbeat detection, if enabled, will cause Heartbeat 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 the amount of time to wait before transmitting Heartbeat  
packets between adapters in the team. Heartbeats 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 timeout period expires (the timer interval value),  
Heartbeat packets are sent. This value MUST be greater than 3000 ms (3 seconds).  
8.  
9.  
Click OK on the Teaming Property Sheet dialog to complete the configuration of the team.  
Repeat Steps 3–8 to create additional teams as desired.  
10. Click Close at the bottom of the Main CPQNTAC dialog to finish configuration of teams.  
Team properties 138  
 
11. Click Close at the bottom of the Network Control Panel applet to finish configuration of network  
devices.  
12. Binding analysis will be done to bind newly installed adapters and teams to all installed  
communication protocols. You may be prompted to enter protocol specific information (such 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 NC77xx 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 Packs  
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 Adapters  
HP NC360T PCI Express Dual Port Gigabit Server Adapter  
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 BUFFERS = 2000 (or larger)  
SET MAXIMUM PACKET RECEIVE BUFFERS = 10000 (or larger)  
SET MAXIMUM PHYSICAL RECEIVE PACKET 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 throughput and fault tolerance for NetWare 5.1, 6.0, and 6.5.  
Network Fault Tolerance (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 Primary adapter fails and is later replaced, or if a  
cable is reattached to fix a failure on the primary 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 information.  
Team properties 139  
 
   
o
o
Syntax: LOAD CPQANS MODE=XXX  
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 adapters 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 other team members transmit. ALB works with any  
Ethernet switch. No performance gain is made with a hub or repeater.  
o
o
Syntax: LOAD CPQANS MODE=XXX  
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 EtherChannel technology. As with NFT, FEC provides  
redundancy. An FEC team uses the switch FEC technology to perform the load balancing rather than  
the server. The switch algorithm determines which ports are used for transmits and receives. All  
adapters are capable of receiving and transmitting at the same time.  
o
o
Syntax: LOAD CPQANS MODE=XXX  
Example: LOAD CPQANS MODE=FEC  
NOTE: FEC teaming requires a switch with Fast 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 Gigabit EtherChannel technology. 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 determines which ports are used for transmits and  
receives. All adapters are capable of receiving and transmitting at the same time.  
o
o
Syntax: LOAD CPQANS MODE=XXX  
Example: LOAD CPQANS MODE=GEC  
NOTE: GEC teaming requires a switch with Gigabit 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 – LACP). Creates a team of HP Gigabit server adapters  
to increase receive and transmit throughput using a switch that supports 802.3ad with LACP  
protocol.  
o
o
Syntax : LOAD CPQANS MODE=xxx  
Example : LOAD CPQANS MODE=AD  
Important teaming notes  
Configuring Fast EtherChannel (FEC) or Gigabit EtherChannel (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 your switch is correctly configured, you can configure your GEC  
or FEC team in accordance with your switch. You should always configure your switch first then your  
server for GEC or FEC teams.  
Team properties 140  
 
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), Network 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 Gigabit Server Adapters, NC71xx Gigabit Server  
Adapters, NC31xx Fast Ethernet Server Adapters and 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 adapter (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 helpful in configuring or troubleshooting teaming.  
CPQANS STATUS  
o
o
o
Description: This command displays the current status for the CPQANS team.  
Syntax: LOAD CPQANS STATUS  
Example: LOAD CPQANS STATUS  
If you have more than one team, then you can issue the same command with the team's ID number.  
o
o
Syntax: LOAD CPQANS STATUS TEAM=<IDnumber>  
Example: LOAD CPQANS STATUS TEAM=2  
CPQANS PRIMARY  
o
o
o
Description: This command creates a preferred primary.  
Syntax: BIND CPQANS <adapter name> PRIMARY  
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 automatically 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 Server 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 141  
 
Error conditions  
If you receive the error message "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 been issued for each adapter in the team.  
All adapters in a team are connected 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 been successfully 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 meet your special configuration needs. Each of these  
variables is described below.  
You can use the following formula to determine the 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
o
o
Default = 180  
Range = 18–1080  
Description: Changes the team's balance interval. 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 balance_set_default [team=nnn]  
Description: Restores the team's balance interval to the factory default setting.  
load CPQANS probes=[on | off] [team=nnn]  
o
o
Default = Probes enabled  
Description: Enables and disables probe packet generation. Probes are not supported in FEC,  
GEC, and IEEE 803.3ad modes. This can also be used after the COMMIT command.  
load CPQANS probes=[broadcast|multicast] [team=nnn]  
o
o
Default = Probes sent in broadcast  
Description: Hangs probe addressing to broadcast or multicast. This cannot be used after the  
COMMIT command.  
load CPQANS probe_retry_count=xxx [team=nnn]  
o
o
o
Default = 7  
Range = Relative to other parameters  
Description: Changes the team's probe retry count 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_burst_size=xxx [team=nnn]  
Team properties 142  
 
o
o
o
Default = 3  
Minimum = 1  
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_interval=xxx [team=nnn]  
o
o
o
Default = 18  
Range = Relative to other parameters  
Description: Changes the time interval sending of regular probes. 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_check_interval=xxx [team=nnn]  
o
o
o
Default = 2  
Range = Relative to other parameters  
Description: Changes the time interval between 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
o
o
Default = 2  
Range = Relative to other parameters  
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 probe 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
o
o
Default = 64  
Minimum = 0  
Description: Number of Tx ECBs to allocate when the Virtual Adapter is created.  
load CPQANS RESET LBN=xxx [team=nnn]  
Description: Resets a bound adapter. This accesses the adapter by its Logical Board Number.  
Supported keywords  
For a list of supported keywords, see Advanced Server Program for NetWare (on page 110).  
Examples of NetWare teaming  
The following are examples of load and bind statements 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 consists 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 143  
 
Note that ONLY an NFT mode Team can combine an HP Gigabit server adapter (server adapters that use  
N1000.LAN) and an HP Fast Ethernet server 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 slot 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 Team. 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.  
5.  
Finalize the Team with the "COMMIT" command. Since 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 that the Team is now  
in the "COMMIT-DONE" state. No changes to the Team can be made when the Team is in the  
"COMMIT-DONE" state.  
LOAD CPQANS STATUS  
Now that the Team is in the "COMMIT-DONE" state, the CPQANS Team appears as a virtual  
adapter. You may now bind a protocol to CPQANS. ALWAYS ensure that your adapters and  
CPQANS are using the same frame type.  
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 144  
 
This example also differs from the previous example in that the Team will be in the Transmit Load  
Balancing (TLB) mode.  
This example assumes a server with two HP Fast Ethernet 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 Ethernet server adapter (server adapters that use  
N100.LAN). See Overview of NetWare teaming ("NetWare teaming" on page 139) for more  
information about mixing adapters. Use the "LOAD CPQANS 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 groundwork for the Team. Since 802.2 is the default frame type we  
need to specify the frame type Ethernet_II. In addition, TLB should be specified, since NFT is the  
default Team "MODE."  
LOAD CPQANS FRAME=ETHERNET_II  
3.  
4.  
Load the base drivers for each adapter in the Team. 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 assigned 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  
Bind CPQANS.LAN to each physical adapter using the names you defined when the base drivers  
were loaded.  
BIND CPQANS PRIMARY_N100_II  
BIND CPQANS SECONDARY_N100_II  
5.  
6.  
If you have not specified your Team as an Transmit Load Balance (TLB) Team, you may do it now.  
LOAD CPQANS MODE=ALB  
If you forgot to choose the frame type or Team mode during the original "LOAD CPQANS"  
command, you can specify the desired frame type and Team mode any time before the "COMMIT"  
command.  
7.  
8.  
Finalize the Team with the "COMMIT" command.  
LOAD CPQANS COMMIT  
Bind both IPX and IP as if CPQANS 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 adapter (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 aggregation. This example uses IPX on Ethernet 802.2 and IP with the Ethernet  
II frame type.  
See Overview of NetWare teaming ("NetWare teaming" on page 139) for more information about SLB  
Team implementation.  
Use the "LOAD CPQANS STATUS" command to view 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 you do not need to include the first Team's ID  
number on the command line. However, you will need to include it for every Team after the first. If  
Team properties 145  
 
you do not include "TEAM=1" CPQANS defaults to Team 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.  
4.  
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  
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.  
6.  
Finalize the Team with the "COMMIT" command.  
LOAD CPQANS COMMIT TEAM=1  
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.  
8.  
Load CPQANS with the second team. Name each frame type with the "NAME" keyword and  
specify the second Team with the "TEAM=2" keyword. 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  
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 driver 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 146  
 
For CPQANS.LAN to work in VLAN mode, the adapter (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 supported by all HP server adapters that support  
CPQANS.LAN. (Exception: The NC6134, NC6132, and NC6133 adapters do not support VLANs.) This  
example uses an HP Fast Ethernet server adapter and 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 frame type and each VLAN the server is supposed  
to be part of. With a single Team configuration, "Team=1" is optional, since that is the default  
setting.  
a. First, load CPQANS for VLAN 5, assigning 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, assigning 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.  
4.  
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  
After loading CPQANS for all the desired VLANs, bind CPQANS to the base driver using the name  
you defined when the base drivers were originally loaded.  
BIND CPQANS E82 TEAM=1  
BIND CPQANS E83 TEAM=1  
BIND CPQANS ESP TEAM=1  
BIND CPQANS EII TEAM=1  
5.  
6.  
Finalize the Team with the "COMMIT" command.  
LOAD CPQANS COMMIT  
Bind the protocols to each instance (VLAN and frame 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 147  
 
NOTE: VLAN is not supported on heterogeneous teaming.  
You can implement Teaming and VLAN at the same time. Because the script gets large quickly, this  
example only uses one frame type. To have both IP and IPX bound to it, the Ethernet_II frame type will be  
used instead of Novell's default 802.2. A homogeneous 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 keyword once for 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.  
4.  
5.  
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  
Bind CPQANS to each physical adapter.  
BIND CPQANS PRIMARY_II  
BIND CPQANS SECONDARY_II  
Finalize the Team with the "COMMIT" command.  
LOAD CPQANS COMMIT  
6.  
7.  
Bind both IPX and IP to CPQANS for each VLAN.  
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 Fast Ethernet server adapter, loads two frame types  
(Ethernet_802.2 and Ethernet_II) on each VLAN, binds 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 148  
 
2.  
Load CPQANS twice for each frame type, once for each VLAN. Assign 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.  
4.  
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  
Bind CPQANS to each frame type on the physical adapter.  
BIND CPQANS PRI_E82 TEAM=1  
BIND CPQANS PRI_EII TEAM=1  
5.  
6.  
Finalize the Team with the "COMMIT" command.  
LOAD CPQANS COMMIT  
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  
Team properties 149  
 
VLAN properties  
In this section  
VLAN information  
A Virtual LAN (VLAN) is a collection of network nodes 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 information and resources as though located on the same  
LAN.  
The HP Network Configuration Utility supports VLANs in Windows Server 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 range 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 that are members of a team. VLAN definition must  
be done at the Team level.  
If an adapter with VLANs is added to a team, the VLANs defined for the adapter will be deleted.  
When a VLAN is defined for an adapter, the properties of that adapter will be modified 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 added to a team that has VLANs defined, then that  
adapter will have 802.1p QoS enabled.  
The VLAN defined as the ‘Default VLAN’ cannot be deleted if more than one VLAN is defined for a  
team. For example, if VLAN 1 needs to be deleted and it is defined as the ‘Default VLAN’, change  
the ‘Default VLAN’ to a different VLAN. 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 disable 802.1p QoS on the adapter or allow it to remain  
enabled.  
For more information:  
Removing a VLAN  
VLAN properties 150  
 
   
ProLiant Essentials  
In this section  
Networking pack installation ................................................................................................................. 151  
Networking pack installation  
Included with the networking software that shipped with your HP server adapter are a number of  
advanced features that extend the basic networking capabilities of your equipment. You can activate  
these advanced features through the purchase and installation of the following ProLiant Essentials  
networking pack licenses:  
Intelligent Networking Pack—Windows Edition  
Intelligent Networking Pack—Linux Edition  
Accelerated iSCSI Pack for Embedded Multifunction Server Adapters  
Installing a networking pack license from the NCU  
The HP ProLiant Essentials Network License Manager allows you to install and track network licenses on  
your system. To access the License Manager dialog:  
1.  
2.  
From the NCU Main window, click License Manager. The License Manager dialog appears.  
Click Help to access the License Manager help file.  
NOTE: The HP Accelerated iSCSI feature comes standard with stand-alone HP Multifunction  
Server Adapters and Multifunction Server 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)–Windows Edition includes:  
Virus Throttle, a network packet-filtering feature that helps slow down the spread of viruses on your  
system. Virus Throttle monitors all outbound connection requests and counts the number of unique  
connections. It detects abnormal (“virus-like”) behavior in the requests, and slows down excessive  
connection requests to new hosts until you can determine if they are viral in nature and take action.  
Advanced networking capabilities that allow ProLiant servers to adapt and change the network path  
to achieve maximum reliability and performance. These features detect and analyze network  
ProLiant Essentials 151  
 
       
bottlenecks or broken network linkages, and steer traffic to the optimum route based on network  
conditions. Advanced networking features include:  
o
o
o
Fast Path Failover—Allows a ProLiant server to use the quickest available path to the core  
network for all server traffic.  
Active Path Failover—Allows a ProLiant server to detect blocked paths and to redirect data along  
an unblocked path to the core network.  
Router Path Failover—Provides redundancy by grouping routers into one router group. One  
router serves as the active router. If that router fails, a second router assumes router functionality.  
Dual Channel Load Balancing, a special type of teaming 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 switch fault tolerance. Full inbound and outbound load  
balancing is provided across both groups. Should any group completely fail, caused by a failure of  
all teamed ports in the group or by a failure of the group’s switch, the team remains available  
through the other group. These types of Dual Channel Load Balancing teams can be configured:  
o
o
o
o
o
o
Switch-assisted Dual Channel Load Balancing  
Switch-assisted Dual Channel Network Fault Tolerance  
Switch-assisted Dual Channel NFT and Preference Order  
802.3ad Dynamic Dual Channel Load Balancing  
802.3ad Dynamic Dual Channel Network Fault Tolerance  
802.3ad Dynamic Dual Channel NFT and Preference Order  
NOTE: Dual Channel Network Fault Tolerance teaming capabilities are not supported on  
systems running Windows® 2000.  
Cisco Discovery Protocol (CDP) and Link Layer Discovery Protocol (LLDP) are both layer two protocols  
that are neighbor device discovery protocols that run on supported networking devices such as  
routers, access servers, bridges, and switches. Using CDP or LLDP, an HP adapter team can  
passively discover configuration information about directly connected network devices that have CDP  
or LLDP enabled.  
NOTE: LLDP is not supported on systems running Windows® 2000.  
For additional information about INP features and how to install a license on your system, refer to the HP  
ProLiant Essentials Intelligent Networking Pack LicenseWindows Edition User Guide located on the HP  
website (http://www.hp.com).  
For information on how to purchase an HP ProLiant Essentials Intelligent Networking Pack–Windows  
Edition license, go to the HP website (http://www.hp.com) or contact your HP reseller.  
Intelligent Networking Pack-Linux Edition  
The Intelligent Networking Pack–Linux Edition includes Virus Throttle, a network packet-filtering feature that  
helps slow down the spread of viruses on your system. Virus Throttle monitors all outbound connection  
requests and counts the number of unique connections. It detects abnormal (“virus-like”) behavior in the  
requests, and slows down excessive connection requests to new hosts until you can determine if they are  
viral in nature and take action.  
ProLiant Essentials 152  
 
 
To enable Virus Throttle for Linux, an Intelligent Networking Pack–Linux Edition license 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 Intelligent Networking Pack LicenseLinux Edition User  
Guide located on the HP website (http://www.hp.com).  
For information on how to purchase an HP ProLiant Essentials Intelligent Networking Pack–Linux Edition  
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 Windows® 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 Accelerated 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 Essentials Accelerated iSCSI Pack license. The ProLiant  
Essentials Accelerated iSCSI Pack license is part of the ProLiant Essentials software family and is  
integrated with HP Systems Insight Manager (SIM), SmartStart, and HP Management agents. It is included  
with the networking software (release 8.35 and higher) found on the HP website (http://www.hp.com).  
NOTE: The HP Accelerated iSCSI feature comes standard with stand-alone HP Multifunction  
Server Adapters and Multifunction Server Adapters for HP BladeSystem.  
The Accelerated iSCSI feature runs under Windows® 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 Multifunction adapter  
For additional information about the HP Accelerated 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 licenses for networking, go to the  
HP website (http://www.hp.com/servers/proliantessentials) or contact your HP reseller.  
ProLiant Essentials 153  
 
   
Diagnostics  
In this section  
HP NC-Series Intel adapter diagnostics................................................................................................... 154  
HP NC-Series Broadcom Multifunction adapter diagnostics ....................................................................... 167  
PCI shared interrupts  
Shared interrupts are normal in PCI servers, but not all PCI device drivers 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 server to change the interrupt assigned 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 interrupt for a new  
driver.  
HP NC-Series Intel adapter diagnostics  
Overview  
The diagnostic software lets you test 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 install 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 products. This utility is designed to test hardware operation and  
confirm the adapter's ability to communicate with another 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 conduct 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 tests  
NOTE: If there is a DOS network driver present, such as NDIS2 or DOS-ODI, the test utility  
and/or the network driver could become unstable. You should reboot and ensure that there are  
no network drivers loaded.  
Diagnostics 154  
 
     
NOTE: The DOS diagnostic utility, DIAGS.EXE, will always use auto-negotiation to determine  
the link type listed under the "View Adapter Configuration" 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 duplex 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.BATto create driver diskettes.  
2.  
3.  
4.  
Boot to DOS or the EFI shell.  
From the DOS prompt navigate to the \APPS\DIAGS\N100X directory.  
Type DIAGS and press the Enter key. The test utility program automatically scans the hardware and  
lists all NC110X, NC310F, NC340x, NC360x, 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 the PCI slot occupied by the adapter. This is  
primarily used for troubleshooting 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 adapter menu. The test menu allows the user  
to begin the adapter tests, change test options, display diagnostic log files, and access the networking  
submenu.  
Beginning adapter tests  
Selecting Begin Adapter Tests brings up the test screen. While tests are being performed, a rotating  
spinner is shown letting the user know that the application is still running. The results 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 containing zeros means that all tests have passed.  
If there is no responder on the network, then the Network 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 configure the specific tests desired. Toggle between 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—This option determines how many times a test runs. While the test is running,  
the information on the screen is continuously updated. When you press the Esc key, the test is  
cancelled and control is returned to the Test Adapter menu.  
Diagnostic Log—This test is disabled by default. If enabled, the program will ask for a directory for  
the log file. The file it places here is named DIAGS.LOG.  
Diagnostics 155  
 
 
Device Registers—Test patterns are written, read, and verified through the adapter's device registers  
to check proper functionality.  
FIFO—This option writes test bit patterns to the adapter's FIFO buffers to check the FIFOs proper  
functionality.  
EEPROM—This option tests both the readability of the 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 reports  
failure.  
Interrupt—This option tests the adapter's ability to generate an interrupt and have it propagated  
through the system to the Programmable Interrupt Controller (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 register and reads the  
interrupt set register, which verifies the card internally registered an interrupt.  
Loopback Tests—These options are internal loopback tests. These tests set the adapter in the  
appropriate loopback mode and send packets back through the adapter's receive circuitry and  
logic.  
Link—This option checks to see whether or not the adapter has link.  
Network Test—This option tests network communication. It looks for a responder and then sends  
packets. If no responder is found, then the test reports failure. If packets are received back from the  
responder, the test reports success.  
Displaying Diagnostic Log  
The Display Diagnostic Log option allows you to view a detailed report of the tests you just ran. When  
Diagnostic Log is enabled, test results are recorded in a log file named DIAGS.LOG. If it does not already  
exist, the test utility creates it. If it already exists, 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 file is time  
stamped. The test run banner identifies the tested 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 detect a spanning tree on  
the network.  
Setting up as a responder—This allows the user to set up the adapter as a responder so another  
system can perform the continuous network test. Selecting this option displays the transmit/receive  
screen. This test will fail if the adapter does not 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 cancelled and control is immediately  
returned to the Test Adapter menu.  
Detecting spanning tree—This 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 test 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 install an  
adapter.  
Diagnostics 156  
 
 
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 products. This utility is designed to test hardware operation and  
confirm the adapter's ability to communicate with another 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 conduct 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 should be found in the same location as the q57diag.exe file.  
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
5701.bin  
5703.bin  
config.sys  
cpu05.bin  
cpu.bin  
cpu14a.bin  
cpu14b.bin  
cpudg05.bin  
cpudiag.bin  
cpusj.bin  
cpumem.bin  
diagcfg.bin  
flshdg05.bin  
flshdg5x.bin  
flshdiag.bin  
flshd14a.bin  
flshd14b.bin  
flshdgsj.bin  
q57diag.exe  
ump14a.bin  
ump14b.bin  
Do not use emm386.exe. Himem.sys 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 groups: Register tests, Memory tests, Miscellaneous tests, and Data tests.  
They are identified as group A, B, C, and D.  
Group A: Register tests  
A1. Indirect register test  
Diagnostics 157  
 
 
Using indirect addressing method, this test writes increment data into the MAC Hash Register table  
and reads back for verification. The memory read/write is done 100 times while incrementing test  
data.  
A2. Control register test  
Each register specified in the configuration 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 error if it could not generate interrupts.  
A4. Built-In-Self test  
This test initiates Hardware Built-In-Self-Test (BIST) 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 integrity of the PCI config registers.  
Group B: Memory tests  
B1. Scratch pad test  
This tests the scratch pad SRAM on board. The following tests are performed:  
o
Address test: Writes each address with unique increment data. Reads back data to ensure data  
is correct. After filling the entire data with the 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 data. 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 the test, the program reads back one more time to  
ensure data stays correct.  
B2. BD SRAM test  
This tests the BD SRAM by performing the tests as described in test B1. The Scratch pad test.  
B3. DMA SRAM test  
This tests DMA SRAM by performing the tests described in test B1. The Scratch pad test.  
B4. MBUF SRAM test  
This tests DMA SRAM by performing 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 0x1000 sized data buffer is  
used for this test. Before each pattern test, the buffer 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 repeats the DMA for the entire MBUF buffer. Then it  
performs receive DMA from adapter to host. The 0x1000-byte test buffer is cleared to zero before  
each receive-DMA. It verifies the data integrity and the test is repeated for the entire MBUF SRAM  
range.  
Test Pattern Description:  
Diagnostics 158  
 
"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 with 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 with 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 buffer with all FF's.  
"AA55AA55's" Fill the entire host DMA buffer with data 0xAA55AA55.  
"55AA55AA's" Fill the entire host DMA buffer with data 0x55AA55AA.  
Group C: Miscellaneous tests  
C1. NVRAM test  
An increment test data is used in the EEPROM test. It fills the test data into the 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 priorities 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 configuration  
contents is 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.  
C5. VPD test  
The content of VPD is saved first before performing the test. After it is saved, the test writes one of the  
five pattern test data, 0xff, 0xaa, 0x55, increment 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
o
Reset test. Sets the reset bit and polls for self-clearing. Verifies the reset value of the registers.  
Event Mapping Test. Sets SMB_ATTN bit by changing ASF_ATTN LOC bits. Verifies the mapping  
bits in TX_CPU or RX_CPU event bits.  
o
Counter Test. Clears WG_TO, HB_TO, PA_TO, PL_TO, RT_TO bits by setting those bits. Makes  
sure the bits are clear. Clears the Timestamp Counter. 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 to 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 the Timestamp Counter is incremented.  
C7. Expansion ROM test  
This tests the ability to enable/disable/access the expansion ROM on the device.  
Group D: Driver associated tests  
D1. Mac loopback test  
Diagnostics 159  
 
This is an internal loopback data transmit/receive test. It initializes MAC into internal loopback  
mode, and transmits 100 packets. The data should be routed back to the receive channel and is  
received by the receive routine, which verifies the integrity of data. One Gigabit rate is used for this  
test.  
D2. Phy loopback test  
This test is the same as D1. Mac Loopback Test except 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. These are the functionalities of the  
phy.  
D6. MSI test  
This tests the Testing Message Interrupt Function to see if it handles this 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.BATto create a driver disk.  
2.  
3.  
4.  
Boot to DOS or the EFI shell.  
From the DOS prompt or shell navigate to the \APPS\DIAGS\Q57 directory.  
Type Q57DIAG.exe and press the Enter key. The diagnostic tests 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 numbered starting from zero. For  
example, if there are three devices detected, the devices are numbered as 0, 1, and 2. In this  
case, entering the parameter -c 2 will select the last found device as default UUT. In manufacture  
testing mode, by default, all devices are tested; however, if this option is used, only that selected  
device is tested.  
o
Example: q57diag.exe -c 2  
-smbaddr <num>-  
o
Specifies the ASF SMBus Address  
-h  
o
Specifies high resolution (80x50) video mode  
Example: q57diag.exe -h  
o
-l <file>  
o
o
Specifies the log file in which all diagnostic output is saved. The default is no log file.  
Example: q57diag.exe -l mylogfile.txt  
Specifies Print on Error  
-p  
o
Diagnostics 160  
 
o
Example: q57diag.exe -p  
-npol  
o
o
Selects negative link polarity in TBI test  
Example: q57diag.exe -npol  
-q  
Selects quick diagnostic mode  
w <value>  
o
Enables WOL programming in manufacture mode  
After a successful manufacturing testing, the program 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
o
Example: q57diag.exe -w 1  
To disable/enable WOL for LOM designs, Lan On Motherboard, use the Rom Bios Setup Utility,  
RBSU.  
-x <value>  
o
Enables PXE in manufacture mode  
After a successful manufacturing testing, the program 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
o
Example: q57diag.exe -x 0  
To disable/enable PXE for LOM designs, Lan On Motherboard, use the Rom Bios Setup Utility,  
RBSU.  
-asf <value>  
o
o
Enables asf in manufacture mode  
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 number represents the sub-test number. 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 not specified, 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 161  
 
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 are to be run. The default is to  
run one time. A number zero indicates loop forever. A control-C or control-break key can be  
used to break the loop. Any error detected will also stop testing after reporting the error.  
o
Example: q57diag.exe -I 5  
Runs tests five times  
-ver  
o
o
Displays the version of the current test software/eeprom.bin  
If this option is entered, it displays the software version number/silkscreen revision and then exits  
the program.  
o
Example: q57diag.exe -ver  
-bus <num>  
o
o
Tests UUT location  
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 162  
 
o
Example: q57diag.exe -bus 5  
Program tests UUTs on bus 5  
-dpmi  
o
o
Specifies the use of DPMI memory allocation  
Uses DPMI memory allocation method to allocate memory instead of malloc() or free().  
-dir  
o
Displays the version of the IPMI (Intelligent Platform Management 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
o
Sets external loopback packets  
Format: <1000Mbps:100Mbps:10Mbps>  
-mba <value>  
o
o
Enables/disables MBA  
Value = 0 to disable; Value = 1 to enable  
-mbap <value>  
o
o
Selects MBA protocol  
Value = 0 for PXE; Value = 1 for RPL; Value = 2 for BootP  
-mbas <value>  
o
o
Selects MBA speed  
Value = 0 for auto; Value = 1 for 10HD; Value = 2 for 10FD; Value = 3 for 100HD; Value = 4  
for 100FD; Value = 6 for 1000FD (fiber)  
-firm <file>  
Updates device eeprom based on <file> image match  
Diagnostics 163  
 
-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
o
Programs ASF firmware  
<file> is a text file that contains ASF bin file names  
-pxes <value>  
o
o
Sets PXE speed  
Value = 0 for auto; Value = 1 for 10HD; Value = 2 for 10FD; Value = 3 for 100HD; Value = 4  
for 100FD; Value = 6 for 1000FD (fiber)  
-hlb <spd>  
o
o
o
Sets device host loopback mode  
spd = 10/100/1000  
dflt = 1000  
-tr <pkts> <spd>  
o
o
o
Sets Tx and Rx pkts packets  
dflt Tx pkts = 1e6  
spd = 10/100/1000  
-lbsped <spd>  
o
o
Runs D3 loopback test at <spd>  
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 displays packet count  
-errlimit <lim>  
Runs D4 loopback test. Fail after 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 164  
 
o
dflt = 1  
-rstpl <x> <y>  
o
o
<x> = post GRC rst poll interval, dly in ms; dflt = 10  
<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 -firmall options to hide warnings  
-exe  
Executes diags commands without 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 printout  
-logfl  
Log file buffer flush per line  
-nosz  
Do not write the NVRAM with the TPM/NVRAM size  
Error codes and messages  
1.  
2.  
3.  
4.  
Got 0x%08x @ 0x%08x. Expected 0x%08x  
Cannot run test while chip is running  
Invalid adapter device  
Read only bit %s got changed after writing zero at offset 0x%X  
Diagnostics 165  
 
5.  
6.  
7.  
8.  
9.  
Read only bit %s got changed after writing 1's at offset 0x%X  
Read/Write bit %s did not get cleared after writing zero at offset 0x%X  
Read/Write bit %s did not get set after writing 1's at offset 0x%X  
BIST failed.  
Could not generate interrupt  
10. Test aborted by user  
11. Tx DMA:Got 0x%08x @ 0x%08x. Expected 0x%08x  
12. Rx DMA:Got 0x%08x @ 0x%08x. Expected 0x%08x  
13. Tx DMA failed  
14. Rx DMA failed  
15. Data error, got 0x%08X at 0x%08X, expected 0x%08X  
16. Second read error, got 0x%08X at 0x%08X, expected 0x%08X  
17. Failed writing EEPROM at 0x%04X  
18. Failed reading EEPROM at 0x%04X  
19. EEPROM data error, got 0x08X at 0x04X, expected 0x%08X  
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 0x%08x-0x%08x and exclude 0x%08x-0x%08x  
28. DMA:Got 0x%08x @ 0x%08x. Expected 0x%08x  
29. Unsupported PhyId %04X:%04X  
30. Too many registers specified in the file, max is %d  
31. Cannot write to VPD memory  
32. VPD data error, got %08X @ 0x04X, expected %08X  
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 166  
 
41. No external memory detected  
42. DMA buffer %04X is large, size must be less than %04X  
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 mapped 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 reset 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 %08x but found %08x  
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 %02X-%02X-%02X-%02X-%02X-%02X  
69. Slot error, expected an UUT to be found at location %02X:%02X:00  
70. Adjacent memory has been corrupted while testing block 0x%08x-0x%08x  
Got 0x%08x @ address 0x%08x. Expected 0x%08x  
*1 Internal Use. Program will not generate this error.  
HP NC-Series Broadcom Multifunction adapter  
diagnostics  
Overview  
Diagnostics 167  
 
 
The diagnostic software lets you test 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 install 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 products. This utility is designed to test hardware operation and  
confirm the adapter's ability to communicate with another 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 conduct 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 executable.  
About the XDIAG.exe diagnostic tests  
The xdiag.exe diagnostic tests are divided 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 chip registers, accessible through PCI/PCI-E bus, for their read-only and read/write  
attributes. Some critical registers 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) receives 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 make sure no message is generated when interrupt is  
masked off.  
A6. Memory BIST  
This tests all memory modules inside the chip using Built-In-Self-Test (BIST).  
A7. Network link test  
This tests the external link connection. For the fiber medium, this is simply another external loopback  
test. For the copper medium, this is not applicable.  
Group B: Memory tests  
Various patterns (0x55aa55aa, 0xaa55aa55, & address) are used to test each of the memory blocks.  
B1. TXP scratchpad  
B2. TPAT scratchpad  
Diagnostics 168  
 
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 each of the on-chip CPUs. The tests also cover the  
DMA interface exposed to the CPUs. These tests require the presence of a test firmware file inside  
the “diagfw” 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 offload logic. This test  
requires the presence of a test firmware file inside the “diagfw” 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 of the test firmware files in the “diagfw” directory.  
C6. NVM access and reset-corruption tests  
This tests the non-volatile memory access (both read 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 “diagfw” directory.  
C7. Core-reset integrity test  
Diagnostics 169  
 
This test issues multiple chip resets and constant driver load/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 requires the presence of  
test firmware files inside the “diagfw” directory.  
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 CPU’s Fast IO interface, making sure appropriate bits are  
triggered when a particular event occurs (GPIO 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 transmits 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 them 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 requires an external loopback terminator to the  
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 smaller TCP packets (based on the MSS). The packets  
are returned in the MAC loopback mode and their integrity is checked 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 sending packets to a different Tx chain. The packets will  
go through the RPC logic and return to the receive buffers as Rx packets. The integrity of each packet  
is checked to ensure no data corruption.  
Starting the XDIAG.exe tests  
Diagnostics 170  
 
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.  
3.  
4.  
Boot to DOS or the EFI shell.  
From the DOS prompt navigate to the \APPS\DIAGS\MFDIAG directory.  
Type XDIAG and press the Enter key. The diagnostic tests run 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/groups (for example: a1c2)  
-T <grps/tests>  
Enables certain tests/groups (for example: a1c2)  
-c <card#>  
Selects device number on which tests will be running  
-I <iteration#>  
Specifies how many iterations 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 tests 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  
Diagnostics 171  
 
Technical support  
In this section  
Free automated customer support services............................................................................................... 172  
Before you contact HP  
Be sure to have the following information available before you call HP:  
Technical support registration 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 authorized reseller:  
In the United States, see the HP US service locator webpage (http://www.hp.com/service_locator).  
In other locations, see the Contact HP worldwide (in English) webpage  
For HP technical support:  
In the United States, for contact options see the Contact HP United States webpage  
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 improvement, calls may be recorded or monitored.  
o
If you have purchased a Care Pack (service 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 Contact HP worldwide (in English) webpage  
Free automated customer support services  
Free Automated Customer Support Services  
Technical support 172  
 
       
The following sites offer troubleshooting information, compatibility notes, and software 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 website (http://h18007.www1.hp.com/support/files/server).  
For downloadable support software for HP Digital Networking Products, Hubs, Integrated Access  
Devices, Modems and ISDN, Adapters, Remote Access Concentrators/Servers, Software, and  
Switches, go to the HP software and drivers website  
All SoftPaqs sorted by SoftPaq number can be found at the HP ftp support website  
An ASCII version of a SoftPaq can be found by selecting a SoftPaq at the HP ftp support website.  
An index of available software sorted by product can be found at the HP software and drivers  
HP FTP Server  
Navigate to a specific product, and then look for support information from the list of support resources at  
the HP ftp support website (ftp://ftp.compaq.com/pub/softpaq/).  
Technical support 173  
 
Acronyms and abbreviations  
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/Internet Protocol  
TOE  
TCP/IP Offload Engine  
Acronyms and abbreviations 174  
 
 
UNDI  
Universal Network Driver Interface  
VLAN  
virtual local-area network  
Acronyms and abbreviations 175  
 
Index  
A
P
Adapter teaming 129, 139  
Advanced Server Program for NetWare 110  
PCI Hot Plug 9  
PCI shared interrupts 155  
PXE configuration  
B
Boot Agent 69  
D
S
Solaris driver information guide 23  
supported operating systems 5  
Diver and software information 7  
DXE configuration  
NC37xx and NC380x 67  
T
F
Firmware upgrade 48  
U
Unix driver software  
H
Gigabit Ethernet Adapter user's guide 21  
HP ProLiant Essentials value packs for  
networking 151  
W
Windows 2000 driver software  
I
NC150, NC32x, NC37x, NC380, NC1020,  
NC6770, and NC77xx Gigabit Ethernet  
Adapter user's guide 98  
K
Windows 2000 installation notes 15  
Windows Server 2003 NC-series Intel adapter  
properties 103  
Keywords for B06 NDIS2 driver 63  
Keywords for N100 NDIS2 driver 81  
Keywords for N1000 NDIS2 driver 87  
Keywords for Q57 NDIS2 driver 62  
Windows Server 2003 and Windows Server 2003  
x64 installation notes 13  
Windows Server 2003 NC31xx Fast Ethernet  
Adapter user's guide 106  
Windows Server 2003 NC-series Broadcom adapter  
properties 98  
L
Linux installation 20  
N
NCU properties 91  
NetWare Client 32 driver installation for HP  
NetWare installation 28  
NetWare teaming 139  
Index 176  
 
 

Greenheck Fan Ventilation Hood SSFSD 211 User Manual
Grizzly Home Safety Product GS702 User Manual
Harman Kardon Stereo System MAS 101 User Manual
Hearth and Home Technologies Fan GFK21B User Manual
Honeywell Air Cleaner HHT 090 User Manual
HP Hewlett Packard Printer 1160 Series User Manual
HP Hewlett Packard Server DL180 G User Manual
HP Hewlett Packard Stereo System 618C User Manual
Hypertec Mouse LIGHTNING 02 User Manual
IBM Laptop Power 285 User Manual