Problems Configuring HP-UX 11i v1

Kernel Build Warnings for krm Module During Cold-Install or Update

During a cold-install or update, you may encounter a warning message in either /var/opt/ignite/local/install.log or /var/adm/sw/update-ux.log indicating that debug information may be corrupted. The warning may appear as follows:

Generating module: krm...
/usr/lbin/sw/bin/mkdir -p /stand/build
Compiling /stand/build/conf.c...
Loading the kernel...
(Warning) Optimizing ADDILs in the presence of the debug
information. Debug information may be corrupted (1594)
Generating kernel symbol table...

You may safely disregard this warning.

cdrecord Application Requires Patch

The cdrecord application delivered with the XCDROAST-RUN fileset in the TC-OpenSource product requires kernel patch PHKL_24283 or one of its successors (including PHKL_26410 and PHKL_27225).

Enabling AutoFS on High Availability Servers Can Cause NFS Packages to Be Dropped

You are affected if you enable AutoFS on an HA server. Restarting NFS clients (with nfs.client stop/start) or NFS servers (with nfs.server stop/start) while NFS packages are running on the High Available (HA) server will cause the exported file systems that tie to NFS packages to be dropped.

This is because the SG-NFS exported file systems are handled interactively and they are not in the /etc/exports file. When stopping NFS clients or NFS servers, it will unexport all the file systems on the system through the /usr/sbin/exportfs -au command. The file systems will be re-exported if they have the entries in /etc/exports file at the starting phase again, but not the file systems exported interactively by the NFS packages.

Enabling AutoFS on the HA NFS server needs to be done with caution. However, this kind of configuration is fully supported if one of these two actions is taken before stopping the AutoFS daemon:

The Auto FS daemon can be stopped either with the kill -9 command, the nfs.client script, or the nfs.server script.

Once the AutoFS daemon is restarted, the NFS packages can be restarted or manually moved back to the server node. AutoFS can be restarted either with the /usr/sbin/automountd command or the nfs.client script.

Hardware Information Not Captured on Systems with Floppy Drive

On systems with a floppy-disk drive attached, hardware information will not be captured in the System Configuration Repository (SCR) under HP-UX 11i.

After updating a system with a floppy drive from 10.20 to 11i, you will find a core file in /usr/dmi/bin, and hpuxci (HP-UX DMI Component Instrumentation) will not be running.

This problem affects you if there is a floppy drive attached to (or installed in) your system, and you use the System Configuration Repository (SCR) to store hardware information. Since the hpuxci is not running, hardware information will not be captured. The problem occurs if /dev/floppy exists.

A fix for the problem will be made available in the earliest follow-up patch release for 11i. In the meantime, if you depend on the SCR, you should either not update to 11i, or remove the floppy drive.

	NOTE: If /dev/floppy exists, hpuxci sets default attributes for the device. The code that does this does not null-terminate the string passed to atoi() which results in a segfault (segmentation fault). The fix for this will terminate the string correctly.

Auto-Port Aggregation Will Not Work on rp7410 Systems

The igelan driver for core rp2710 cards does not support Auto-Port Aggregation (APA).

To fix this problem install patch PHNE_25833 (or any patch that supersedes it).

EISA Cards Not Supported on 64-bit OS

EISA I/O drivers for TermIO, FDDI, 100BT and Token Ring cards are supported on HP-UX 32-bit OS, but not 64-bit OS. You will not see an error or warning message stating that these EISA products will not work after installing or updating to 64-bit OS.

100BT HSC-FX Cards Do Not Work on HP-UX 11i v1

The100BT HSC-FX cards, J3514A#002 and J3516A#002, do not work under 11i. The card is claimed, but the link does not come up.

This card is supported on K-Class, D-Class and A-180 systems. If you are not sure whether or not the card is installed in your system, use the appropriate command:

ioscan -kfClan | grep J3514A#002 (on K-Class)

ioscan -kfClan | grep J3516A#002 (on A-180 or D-Class)

If the card is installed, you will see the hardware path, interface number and other details displayed on the screen.

To fix the problem, either install patch PHNE_22727 or replace the HSC-FX card with an HSC-TX card, J3514A#001 (K-class) or J3516A#001 (A-180 and D-Class). If you replace the card, you will also need an FX-to-TX converter. The converter is not an HP product but has been qualified for use with HP cards; it is made by Transition Networks, part number C/E-100BTX-FX-04(SC).

SCSI PCI Restrictions for OLA/R

Online addition and replacement (OLA/R) of PCI HBAs is available in HP-UX 11i v1 on the HP 9000 L-Class, N-Class, and Superdome systems. This feature allows you to add or replace a failed HBA without having to reboot the system.

Some restrictions in the ability to perform online addition (OLA) and/or online replacement (OLR) apply with these SCSI HBAs:

You can find more information about SCSI HBA OLA/R restrictions by going to the HP-IT Resource Center (ITRC), at the following URL:

http://itrc.hp.com