NVRAM Configuration mismatch

[Helen Nanong]

After a shutdown and reboot of our Digital Alpha Server 1000A 5/400, the following error appeared: NVRAM Configuration does not match.
In issuing the swxcrmgr command, the GUI displays a FLD drive, however, I am not able to rebuild that drive.Warning message tells me that the drive cannot be selected.
This error has no adverse effect on the server, as users are logging on and using the databases on it. I just want to get rid of the FLD drive indicated, and the error message that appears in the console.

[Steven E. Protter]

Shalom,

This is a very,very serious problem that can lead to a total loss of data. I recommend shutting down all services on the server, backing up all data an configuration on the thing and geting this problem repaired prior to using the system again.

An NVRAM fault could lead to your data on disk being damaged.

SEP

[Hein van den Heuvel]

Helen,

Welcome to the HPUX forum, but you appear to have a TRU64 question which may get better attention in:
http://forums1.itrc.hp.com/service/forums/familyhome.do?familyId=280

Some moderator might move this topic there.)

Actually, it is not rally an TRU64 question either, more like a an ALPHA hardware question.

Anyway... the NVRAM holds config params to be used by hardware, notably during boot.
You set them on the console, without OS, with stuff like P00>> SET ...

You may need to init the system, or otherwise reconfig but i don't think it will be as dramatic as portrait. Still, go slow and ask for help when in doubt (call support!)

Was the NVRAM message always there, or did it appear (recently) with the raid error message?
It coudl be a firmware mismatch between the Alpha itself and the Interfaces. Update your KZPSA's or NICs as appropriate.

If you need to ask for further help here be sure to include the EXACT error texts and/or cut&paste the var/adm/message from a screen shot.

If TRU64 is running, then check /var/adm/messages for clarification on the NVRAM message.
Which interface appears to be involved?

The SWXCR (KZPSC?) error message with the FLD drive may be just that. A FAILED drive which refused to commumincate with the system and thus you can not control. The RAID-1 or RAID-5 config may have saved the day, but you may be on thin ice.
Typically I would re-seat the drive for a final chance on perhaps reboot, but that drive may well be a total loss and the message will stay untill it is gone.
Read SWXCR doc and/or call support!

hth,
Hein.

[Helen Nanong]

Thanks Steven and Hein,

I appreciate your comments. I would like to run an SCU, RA200RCU.exe for now, before I do anthing more. This a first for me, and I do not want to do anything that is not reversible.

[Helen Nanong]

Attached is the exact content of the var/adm/message for the 19th of August, which is the date the server was shutdown and rebooted. btw, the server was shutdown to allow for a UPS maintenance, I am unsure whether that has anything to do with this error.

[Hein van den Heuvel]

This log confirms a disk failure:

"vmunix: re0 at xcr0 unit 0 (unit status = CRITICAL, raid
level = 1)"

Follow the process in your operations book to replaced the failed drive asap.
You can attempt a re-seat, but the data will eb out-of-sync and replaced whether you re-use the failed drive, or put in a fresh one.

The log does not appear to mention 'nvram'.
Does that message appread only on the 'real' console? Then you'll have to deal with it there.ehind bus 0,

The log also indicates trouble with a scsi device behind the Qlogic (aka Digital KZPDA - Fast Wide SCSI-2 PIC adapter).
Check connectors/terminators on cable?

You may want to try boot the generic kernel and rebuild a kernel to match your system with doconfig?


hth,
Hein.

[Helen Nanong]

Thanks Hein.
The NVRAM error appears on the console itself, I did check for it in the /var/adm/message but as you pointed out it is not there. I will use what you have given and see how I go.

[Helen Nanong]

The NVRAM error is no longer displayed after I checked the connectors/terminators at the back of the server. However in the swxcrmgr the FLD drive is still showing - with the following warning: A device is configured but not physically present at this location, when I select to rebuild drive.

[Hein van den Heuvel]

[ha, this topic was moved from the hpux forum to here. That's better]

" A device is configured but not physically present at this location, when I select to rebuild drive. "

Like I said, a bad drive which refuses to play.

Is it physically present and not talking, or did someone actually remove it?

The prudent action is to replace it completely, and while you are there, add a spare drive to protect against future failure.

You may be able to bring it back to life by unplugging is and re-seating it, but the 'work' involved making it real again, making it catch up is by now pretty much the same as just replacing it outright.

Hein.

[Helen Nanong]

Hein,

In answer to your question, the drive is physically present but not talking. Even so there is not orange light indicating a FLD drive, and I just lost the GUI (don;t know what happened) and cannot get into the swxcrmgr. I am stuck as I am not sure what to do next.

[Vladimir Fabecic]

Hello Helen
I think that Hein said enough.
So you have one degraded RAID set because of failed disk.
Replace the disk. I have seen many failed disks without the amber light on.
If having problem with OS (you should not have) you can do it from console (>>>arc , then "run a program" then a:swxcrmgr) running program fron diskette.

[Helen Nanong]

Thanks for the help. I will take it on from here