Crash analysis!

[S. Karunanidhi]

Hi!
Can anyone brief me how to do the initial Crash analysis using either adb or q4? I have a K450 which crashes frequently? No info form Syslogs.

Thanks in advacne,
Karunanidhi.S

[Antoanetta Naghiu]

q4 per HP recommandation:
#ls -al /usr/contrib/bin/q4
if not exist: swinstall from OS-Core cdrom (OS-Core.Q4)
Supposelly you've got a crach in /var/adm/crash/crash.n, n=the number of the last crash, cd to that directory:
#cd /var/adm/crash/crash.0 (or 1 or... whatever numer is your crash)
If there the files are ziped,
#gunzip vmunix.gz
After that:
#q4prep -p
#q4 -p . (q4_space_minus_p_space_dot_enter)
q4>trace
at prompt type: "include analyze.pl" and next, type "run Analyze au >> ana.out
at the next prompt: exit
grep -v HPMC ana.out
if HPMC exists, is a hardware failure.
Check as well /var/tombstone, /etc/shatdownlog
A q4 procedure you can get from Response Centre as well.
Good luck

[Bill Hassell]

Brief crash dump analysis procedure? That's not really possible as determining the 'real' reason for a crash requires access to source code (plus all the patch sources), a huge database of stack trace signatures and a lot of experience. The reason for a crash might be found in shutdownlog: "data segmentation violation" which means the kernel tried to use an invalid address in a memory reference (ie, an odd numbered address for an integer). The message doesn't provide a clue as to why the kernel has become unstable.

Some basics: look at the stack trace to see what modules in the kernel are involved (ie, filesystem, lan code, virtual memory, etc). Then use the modules near the crash event to search through the patch database.

Or, apply the Support Plus patch bundles. There are dozens of crash fixes in the Critical and General release bundles. If your system is unpatched, this is likely the problem. If managing patches is difficult, come to HP World 2000 in Philadelphia next month and attend several papers and seminars on patch management.

Or, call the HP Response Center and they can decode the crash dump quite quickly.

[S. Karunanidhi]

Hi Guys!
Thanks for the response.
From the /etc/shutdownlog, I understand 2 things,

1. The machine has been crashing since febrauray.
2. There is a HPMC behind all os these crahses.

I have reseated all the add-on cards once.
I have attached the /etc/shutdownlog.

Any inputs ??

Karunanidhi.S

[melvyn burnard]

Get a Hardware call logged with HP to fix this hardware problem.

[Stefan Farrelly]

Wow, your server has been crashing a lot! The problem is hardware related, the only real thing to do is call HP support for help as an engineer will be required to fix it.

[Alan Riggs]

Do you have online diagnostics installed? If so you should have a /var/tombstones directory which will have a record of teh chasis codes associated with each HPMC. You can log a hardware card and have these codes (one set for each CPU) analyzed. If not, then a technician can come onsite (or talk you through) examining the pims on the console. This will require additional server downtime, but it looks like you are experiencing much of that anyway.

[Cheryl Griffin]

HPMC's would be reported to hardware support for resolution but for any other non-hardware related crashes, the ITRC Knowledge Base contains many articles on how to pre-process the information for HP using Q4

For instance:
OZBEKBRC00000611 How do I pre-process my crash dump so HP can troubleshoot it?

Search keywords: Q4 and crash