1831647 Members
2085 Online
110029 Solutions
New Discussion

Help with Crashes

 
Vasu_1
Occasional Advisor

Help with Crashes

Hi,

Two questions:
1. We have a J6700 running 11.11 which has been experiencing crashes every few hours. Attached is the /etc/shutdownlog file. Does the pattern point to a hardware error? I am running q4 on the crash file as I write this so that should give me some info.
2. I am also trying to figure out what patch bundles to apply to the machine, in hopes that this is a software error (the machine has been experiencing heavy loads lately, so maybe that's a factor). Does anybody know what the hardware enablement and the GOLDBASE/GOLDAPPS patch bundles are for? Are they necessary?

Thanks in advance.

Vasu
10 REPLIES 10
Marco Santerre
Honored Contributor

Re: Help with Crashes

2. Hardware enablements will give you the latest patches for drivers which will enable your server to recognize the newest hardware. GOLDBASE/GOLDAPPS, I would recommend strongly as they are the general release patch bundles for 11i.

Cooperation is doing with a smile what you have to do anyhow.
Robert-Jan Goossens
Honored Contributor

Re: Help with Crashes

Hi Vasu,

http://software.hp.com/SUPPORT_PLUS/qpk.html

You can read more about these defect updaters at above page.

Kind regards,
Robert-Jan
Helen French
Honored Contributor

Re: Help with Crashes

Yes, Hardware enablement and GOLDBASE/GOLDAPPS are required patch bundles and you need to install it on the system. HWE patches will resolve hardware issues on the system where as GOLD* patches will resolve software issues.

I think installing these latest bundles and the latest pacthes will resolve your issue.
Life is a promise, fulfill it!
Ashwani Kashyap
Honored Contributor

Re: Help with Crashes

1. It could be a hardware or software error . The hardware error could be on the memory . Run stm and see if you have multiple single bit errors on the memory board . If it is then its time to replace the memory .

It could also be due to a faulty code , which could be corrupting the memory . In that case you might need to patch the system . Send the q4 analysis results to HP immediately and they would know exactly what happened .

2. GOLDBASE patches are software patches for the base operating system like core os , lvm etc.
GOLDAPPS patches are for other applications like MCSG etc .

3. Hardware enablement patches are obviously patches for the hardware .

YOu should have a combination of these patches installed periodiaclly after careful review .
Vasu_1
Occasional Advisor

Re: Help with Crashes

1. So the presence of a lot of "Data page fault" and "Instruction page fault" and
some of the other lines in the shutdown log don't necessarily imply a hardware error?

2. q4 has been hanging at the "Run Analyze AU" command for a long time now -- still waiting.

3. When I run stm on a different, good machine here, I see a LOT of single-bit errors in memory, so it seems to me that that may not indicate a serious problem. The machine with the problem is at a remote location and I need to ask someone to run stm on it and send me the results.
Helen French
Honored Contributor

Re: Help with Crashes

Data page fault and Instruction page fault messages on shutdownlog.log could be a software error. I 've seen lot of these situations where the latest patches are the best solution.
Life is a promise, fulfill it!
Steven E. Protter
Exalted Contributor

Re: Help with Crashes

A careful read of this thread would lead me to to the following in your shoes:

Install The Gold Base, HWE and Bundle Patch sets if not installed.

Install the September 2003 Patch set(I'm a little queasy about December 2003 right now).

Then see if you get crashes.

It is also a good idea to send your q4 output to HP. They are really good an analyzing the output and figuring out if you are missing a specific patch.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
Vasu_1
Occasional Advisor

Re: Help with Crashes

One more question: When I click on the links from the September 2003 patch bundle, I see 3 depots and 4 patches. Does anyone know how to combine these into one depot so that only one reboot is required? Any pointers to documentation? The man pages for swcopy and swinstall don't talk about this. Any help appreciated and many thanks to all who responded previously.

Vasu
Vasu_1
Occasional Advisor

Re: Help with Crashes

Never mind the above. Got it from the online docs. Thanks.

Vasu
melvyn burnard
Honored Contributor

Re: Help with Crashes

Having looked at the shutdown log, I strongly suspect you have a hardware problem, with the cpu being most likely to be the culprit.
I would suggest you get these crashes analysed by your local HP Response Centre
My house is the bank's, my money the wife's, But my opinions belong to me, not HP!