topic Re: Server Rebooted in Operating System - HP-UX

Server Rebooted

John Ferro — Sat, 15 Aug 2009 08:07:58 GMT

Hello,

I have HP 9000/800/SD64000, unfortunately server rebooted. Which log files i should look into to know the real cause of reboot. Pls. mention any docs clarify how to analyze reboot log files.

Regards,
John

Re: Server Rebooted

SoorajCleris — Sat, 15 Aug 2009 08:34:52 GMT

Hi,

You may check for the Crash dumps, There are different types of Crashes,

1.Panic, toc, SG toc, and HPMC. If your server was crashed because of HPMC it will give you

/var/tombstones/ts99

check for the crash files in /var/adm/crash (default )

Analysis these dumps is not that easy ( user) you may please contact HP.

Regards,
Sooraj

Re: Server Rebooted

sen_ux — Sat, 15 Aug 2009 08:56:43 GMT

Check /etc/shutdownlog
#cat /etc/shutdownlog

this will give a hint on what caused the server reboot.

Check for extstance of ts99 and crash with a valid timestamp.

For a detailed analysis, you may open a case with HP and they will help you further.

Re: Server Rebooted

Shailendran V Naidu — Sat, 15 Aug 2009 09:25:36 GMT

Server reboots if :

There is a Hardware failure
Software failure like a Panic.

If you look at /etc/shutdownlog and there is a line like :

timestamp Panic ------- something ---

in that file it is a Panic condition can be triggered by s/w.

If there is no entry as such then it is a Hardware issue. You may want to check tombstones for a ts99 file and verify the timestamp and timestamp the server has rebooted.

If both are not the case - Mostly the issue can be related to power problems at site.

Re: Server Rebooted

Raj D. — Sat, 15 Aug 2009 13:17:46 GMT

John,

You can also check GSP log :
SL --> SEL ->
Also check by alert level , you can choose "7" fatal error logs.

-First of all you can check the /etc/shutdownlog to see the cause of the reboot, if it is "reboot badnews" , check for anyother information there.

- Also to check /var/adm/crash/crash.x
if there is any crash.0 present with the timestamp , the the crash dump can be uploaded to HP FTP site for analysis. And they will provide the cause of the crash. Or they will tell if there is any defective hardware need to replace .

- also to check /var/tombstone/ if there is any file with valid cpu timestamp. May be you got some faulty cpu.

Cheers,
Raj.

Re: Server Rebooted

Raj D. — Sat, 15 Aug 2009 13:22:02 GMT

Also if the reboot is due to a power outage, there will be no entry in the log in /etc/shutdownlog .
Cool!.
Then it needs to be check from GSP log, and it will report power input issue.

Once you verified the above you can come to a conclusion the cause of the reboot.

Cheers,
Raj.

Re: Server Rebooted

Dennis Handly — Sat, 15 Aug 2009 23:07:52 GMT

Strictly speaking your problem isn't that your server automatically rebooted, it's that it crashed.

Re: Server Rebooted

Sudeesh — Sun, 16 Aug 2009 15:16:40 GMT

Verify /etc/shutdownlog ->
if you see an entry with latest timestamp, check if someone rebooted the server or it is a panic(you can identify from the shutdownlog entry)

If panic -> check /var/tombstones/ts99 and /var/adm/crash/crash.#(Usually only HP can analyze these files)

No entry in shutdownlog -> login to MP -> SL -> and check if there are any errors logged with latest timestamp. HP can decode the logs for you.

Thanks

Sudeesh

Re: Server Rebooted

John Ferro — Sun, 16 Aug 2009 15:31:11 GMT

Hi

Thanks to all for your time, i checked all log files and HP Mission critical team found oracle related issue...

Regards,

Re: Server Rebooted

John Ferro — Sun, 16 Aug 2009 15:31:31 GMT

Thanks....