Operating System - HP-UX
1832396 Members
3176 Online
110041 Solutions
New Discussion

crash dump on R380 server

 
CSP_ALGERIA
Frequent Advisor

crash dump on R380 server

Hello
Our customer has a crash on his server
an HP9000 R380 HPUX 11.0 FLT B800 and WARN E000 that means an OS panic .
we have changed the disk and memory but it still doing a crash and hang
system , i send you also the
/var/tombstones/ts99 file .
thanks
Nothing in the world can take the place of persistence.
5 REPLIES 5
Steven E. Protter
Exalted Contributor

Re: crash dump on R380 server

I would reccomend crash dump analysis with the q4 tool.

Attaching my version of this process. There is a much better document if you have a software contract.

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
Mohanasundaram_1
Honored Contributor

Re: crash dump on R380 server

Hi,

The ts99 file indicates an HPMC on 20th May 2004. This indicates there is a hardware issue (99.99% of the time).
HPMC Chassis Codes = 0xcbf0 0x5008 0x5408 0x5508 0xcbfb

Would suggest something about I/O giving you the problem. If your crash is very frequent, then you can try starting the system with minimum hardware. Remove all the I/O cards from the system. Remove any external tape/arrays and check if the system is stable. Then add the I/O cards one-by-one, each time ensuring that the problem is not appearing. during this process if the system crashes, the last component you Added is the culprit.

An easier way is to get HP engineers decode this HPMC (TS99) to pinpoint the failed component (if you have a contract).

Note: before starting, ensure you have got a good ignite and data backup for the system.

Hope this helps.

Rgds,
Mohan.
Attitude, Not aptitude, determines your altitude
Tonya Underwood
Regular Advisor

Re: crash dump on R380 server

I would probably lean towards a CPU on this one but it's hard to say. What exactly was the panic string? (tail /var/adm/shutdownlog)

Also if you could post the following as well:

grep vmunix /var/adm/syslog/OLDsyslog.log

Thanks,
Tonya Underwood
Kent Ostby
Honored Contributor

Re: crash dump on R380 server

Its an HPMC.

Its most likely NOT a CPU which would have a 2xxx code.

Its likely either Memory, I/O card, or possibly something fixable by patches (you have essentially the most generic type of HW failure based solely on Chassis Codes).

HP support can determine the proper route for you on this case.

Best regards,

Kent M. Ostby
"Well, actually, she is a rocket scientist" -- Steve Martin in "Roxanne"
Mohanasundaram_1
Honored Contributor

Re: crash dump on R380 server

HI,

Did you fix the problem ? DO you mind posting the resolution?

Cheers,
Mohan.
Attitude, Not aptitude, determines your altitude