This chapter describes known issues and restrictions pertaining to the NHD-5 kit, and includes notes on the following areas:
Documentation (Section 2.1)
Patch kit compatibility (Section 2.2)
Full Installation only on DS25 (Section 2.3)
DS25 startup with no IDE disks (Section 2.4)
Shell access during Full Installation (Section 2.5)
Smart Array 5300A (Section 2.6 )
Installation messages (Section 2.7)
The files and their layout in the NHD-5 kit are updated before production. Subsequently, the listings and messages documented in this manual may differ slightly from what you see on your system. This should have no effect on your installation.
Some listings and messages may have a backslash (\
2.2 Patch Kit Compatibility
When you install NHD-5, you also must install the most current Version 5.1A patch kit before you return your system to production. It does not matter which one you install first.
Tru64 UNIX Version 5.1A patch kits are available on the World Wide Web at the following URL:
http://ftp.support.compaq.com/public/unix/v5.1a/
Section 2.7.5
describes messages you may
see if you install the current Version 5.1A patch kit before you install NHD-5.
2.3 Full Installation Only on DS25
You must install the NHD-5 kit during a Full Installation of
the operating system on a DS25 system.
The kernel modules to support the DS25
are included in the NHD-5 kit.
2.4 DS25 Startup with No IDE Disks
You may encounter the following behavior when starting up a DS25 system configured with SCSI disks but no IDE disks:
If you power on a DS25 system and it has no IDE disks installed, the IDE controller driver may detect and report a stray interrupt. Subsequent boots without cycling power do not exhibit this behavior. This is a known issue and can be ignored.
If you boot a DS25 system from a SCSI disk and it has no IDE disks installed, the system may report IDE probe errors during the boot process. Boot time may be increased as the system issues bus resets. This is a known issue.
2.5 Shell Access During Full Installation
If you are installing NHD-5 from the New Hardware Delivery CD during a Full Installation and you try to launch a shell window, it will fail. You see the following message in the console window:
Error acquiring pty slave as controlling terminal: Not a typewriter.
Perform the following steps to access the shell:
Exit the Full Installation process to access the shell from the console device.
At the shell prompt, perform the intended actions.
Enter
restart
and press
[Return]
to restart the Full Installation process.
2.6 Setting Up SA5300A Controllers
This section discusses issues related to installing NHD-5 onto a system that includes SA5300A series RAID controllers.
If you want to use a disk managed by an SA5300A series RAID
controller as a boot device, you must install the NHD-5 kit during
a full installation of the operating system.
If you install NHD-5 onto
an existing system with the
nhd_install
script, disks managed
by SA5300A series RAID controllers cannot be used as boot devices.
If your system includes SA5300A series RAID controllers, you
must perform certain actions from the SRM console before you install NHD-5
during a Full Installation of the operating system.
If you are
installing NHD-5 onto a system already running Version 5.1A of the
operating system, you must perform the same actions from the SRM console and
may need to take additional steps to rebuild the kernel before you run the
nhd_install
script.
If you are installing NHD-5 during a Full Installation, follow the instructions in Section 2.6.1 before you install the NHD-5 kit. This applies whether you are installing to an SA5300A-backed disk or to a non-SA5300A disk.
If you are installing NHD-5 onto a system already running Version 5.1A
of the operating system and you have changed the value of the
heap_expand
console variable, follow the instructions in both
Section 2.6.1
and
Section 2.6.2
before you install the NHD-5
kit.
2.6.1 Performing SRM Console Actions
Before you install NHD-5 onto a system that includes SA5300A series RAID controllers, follow these steps:
At the console prompt, use the following command to determine the console device name of the SA5300A controller:
>>> show config | more
Look for devices with a
py
prefix; you see output similar to the following:
CPQ SmartArray 5300 pya.0.0.10.0
In this example, the console device name of SA5300A controller is
pya0
.
SA5300A controllers are assigned console device name
pya0
,
pyb0
,
pyc0
, and so on,
in the order that they are scanned by the system upon powering up.
Determine the value of the
heap_expand
console variable.
For example:
>>> show heap_expand heap_expand = nMB
If the value is
2MB
, go to the next step.
If the value is anything else, set the value to
2MB
:
>>> set heap_expand 2MB
Caution
If your system is already running Version 5.1A of the operating system and you change the value of the
heap_expand
console variable, you may not be able to reboot your system unless you follow the kernel rebuilding instructions in Section 2.6.2 first.
Initialize the system:
>>> init
Set the value of the
bootbios
console variable
to the SA5300A console device name.
For example:
>>> set bootbios pya0
Initialize the system again:
>>> init
If you are installing NHD-5 to an SA5300A-backed disk, you must create at least one logical volume first. See the SA5300A series RAID controller Installation and Configuration Guide for the procedure to create logical volumes with the Option ROM Configuration for Arrays (ORCA) offline configuration utility.
The target SA5300A BIOS now runs whenever the console is initialized.
The console
show dev
command will show SA5300A-backed logical
disks such as
DYA0
,
DYA1
, and so on
when
bootbios
is set to
pya0
.
If you have more than one SA5300A series RAID controller, you will only
see the devices backed by the controller specified by the
bootbios
setting.
If you are installing NHD-5 during a Full Installation, you are ready to follow the instructions in Chapter 3.
If you are installing NHD-5 onto a system that includes
SA5300A series RAID controllers and is already running Version 5.1A of the
operating system and you have changed the value of the
heap_expand
console variable, follow the steps in
Section 2.6.2
before you install NHD-5.
2.6.2 Rebuilding the Kernel After Changing HEAP_EXPAND
If you are installing NHD-5 onto a system that includes SA5300A
series RAID controllers and is already running Version 5.1A of the operating
system, make sure you have performed all of the SRM console actions in
Section 2.6.1.
If you have changed the value of the
heap_expand
console variable, follow these steps:
From the console prompt, boot the generic kernel. For example:
>>> boot -fi /genvmunix
If you do not have a generic kernel on your boot disk, boot the generic kernel from the operating system software distribution CD.
At the root shell prompt, use the
sizer
command to create a temporary configuration file in the
/tmp
directory.
For example:
# sizer -n filename
Make sure that
filename
or
filename.dev
does not already exist in the
/tmp
directory.
Search the temporary configuration file for the embedded value
of the
LOADADDR
parameter.
For example:
# grep LOADADDR /tmp/myfile makeoptions LOADADDR="fffffc0000230000"
Edit the current system configuration file
/sys/conf/SYSNAME
and change the existing value of the
LOADADDR
parameter to the value in the temporary configuration file.
In this example, the value from the previous step is
fffffc0000230000
.
Use the
doconfig
command to build a new
custom kernel.
For example:
# doconfig -c /sys/conf/SYSNAME
Copy the newly built kernel to
/vmunix
.
For example:
# cp /sys/conf/SYSNAME/vmunix /vmunix
Reboot your system.
The following sections describe messages related to installing the NHD-5
kit and tells you whether to ignore them or take corrective action.
2.7.1 DS25 Boot Messages
During a DS25 system boot, you may see the following PCI table error message:
PCI device at bus 0, slot 3, function 0 could not be configured: Vendor ID 0x1000, Device ID 0x21, Base class 0x1, Sub class 0x0 Sub-VID 0x14d9 Sub-DID 0x8002 has no matching entry in the PCI option table
You also may see the following unrelated driver error message:
<FDI: FATAL ERROR #174 sra:000000ff dor:000000ff msr:000000ff> fd internal driver error: FDI PROBE FAIL (A,11).
Both of these messages reflect known errors and can be ignored.
2.7.2 DS25 daemon.log Error Messages
After booting a DS25 system, you may see error messages similar to the
following in the
daemon.log
files:
Sep 24 11:28:37 hpsq16 [616]: Initializing the threshold structure Sep 24 11:28:38 hpsq16 [616]: **ERROR svrsys_fru_parse.c line 1264: \ FRU Table TLV tag of 0 is not ISOLATIN1
These messages are generated when there is no FRU table available on
the system, and can be ignored.
2.7.3 Tag File Creation Message
During the Setup Stage of a Rolling Upgrade, you may see a message similar to the following during tag file creation:
clubase: Entry not found in /cluster/admin/tmp/stanza.stdin.530756
This reflects a known error and can be ignored.
2.7.4 Rolling Upgrade Status Messages
If you check the cluster upgrade status during a Rolling Upgrade, you see output similar to the following example, done after the Setup Stage is complete:
# clu_upgrade -v Retrieving cluster upgrade status. Upgrade Status Stage Status Date setup started: Thu Jun 20 08:59:48 EDT 2002 lead member: 1 patch kit source: /mnt completed: Thu Jun 20 09:05:29 EDT 2002 Member Status Tagged File Status ID Hostname State Rolled Running with On Next Boot 1 member01.site.place.net UP No No No 10 member10.site.place.net UP No Yes Yes 20 member20.site.place.net UP No Yes Yes
This output incorrectly indicates
patch kit source
rather than
nhd kit source
.
The source information (in
this example,
/mnt
) is correct, the label is in error.
2.7.5 Older File Installation Error Messages
If you install the current Version 5.1A patch kit before you install the NHD-5 kit, you may see multiple error messages similar to the following when you install the patch kit.
OSHHWBASE521 version ./path/file is older than the previous, restoring previous version
These messages indicate that the files in the NHD-5 kit have
been superseded by the files in the patch kit and that the patch kit's version
of the files will be used instead.
You can ignore these messages.
2.7.6 RIS Installation Error Messages
If you are installing NHD-5 from a RIS server, you may see messages similar to the following before the system is configured:
The /usr/sbin/versw -setnew command failed. This error is not fatal, and the operating system installation will continue. The following message was received from /usr/sbin/versw -setnew: Failure to set the new version identifier
The kernel that is running at this point includes NHD bits, but NHD
support has not yet been loaded or configured on your system.
This reflects
a known issue and can be ignored.
2.7.7 mknod Installation Error Message
When you install NHD-5 either during a Full Installation or with
the
nhd_install
utility from either a CD or CD image, you
may see messages similar to the following when the installation process is
loading the NHD base subset:
Copying from /instkit1//520/kit (disk) /usr/sbin/setld: mknod: not found /usr/sbin/setld: /var/tmp/pipe5255: cannot open Working....Thu May 23 14:23:11 EDT 2002 Verifying mkfifo: /var/tmp/pipe5255: File exists
These reflect known issues and can be ignored.