sudden reboot

[CowBoy]

Dear All
My yesterday i have lost the ping of my OS(hp-ux ) and when i connected by MP Port i saw that the server is rebooting but there is not any unusual log in any log directory....
I also installe oracle 10gr2 and the alert log is the same ,no sign of stopping oracle,there is just "starting instance normal".
in syslog also there is nothing important to notice....
How can i find what is really going on my OS???
The MP log is also empty!!!!!!!!!

[Torsten.]

How about a sudden power loss?

There must be anything in the MP logs - at least for the last boot.

[Ganesan R]

Hi,

There are lot of reasons for server reboot. Could be due to hardware issue(LPMC, HPMC, etc..), software panic, etc.

Check if you have any dump files in /var/adm/crash directory with yesterday timestamp. If you have crash directory involve hp to analyse.

check the ts99 file in /var/tombstone directory. if it is created yesterday then it could be due to hardware issue.

[Suraj K Sankari]

Hi,

Did you checked the syslog.log file.

Suraj

[R.K. #]

Hello,

Please check following:

1. /etc/shutdownlog
2. /var/adm/syslog/OLDsyslog.log
3. Also see if the crash dump is created in /var/adm/crash with latest timstamp

Regds,
R.K.

[DeafFrog]

Hi ,

Did u check the more /var/opt/resmon/log/event.log ...check if there is some issue with H/W.

[UVK]

#cat /etc/shutdownlog
#dmesg

Check for the above files to see if you find anything reg the reboot.

Check for the MP logs there should be something in it.

Then check for /var/tombstones/ts99 file and see if there was any HPMC or LPMC.

[S. Ney · Accepted solution]

I've attached a doc from HP about what to check when the system crashes.

[CowBoy]

Thanks alot buddies
I have checked all the logs that you said to me...but there is no unusual log that give me a clue to find the reason ...
I am really getting mad ....

[Nido]

Hi,

As Torsten said, sudden power loss may be the cause, which is not leaving any system logs.
Power-supply issue should have been reported in EMS logs.

Thanks,

[CowBoy]

In power loss The MP show the sth to lose power but there is nothing there...
but in /var/tombstones/ts99 there is a log from that date bu i dont know how to read the file????

[Torsten.]

What server model is it?

Try to run "pdcinfo" if this is a RISC system and attach. Let's see if this helps.

[Nido]

Hi,

Open a Hardware call with HP and have the ts99 and other files analysed by the HPCE.
I prefer openning a call with HP rather than checking the possibility of the causes.

Thanks,

[CowBoy]

Hi torsten
It is rp8420....

[Torsten.]

Another point to look at is firmware ... if you read all the release notes you can find somethings like

"... in rare cases ... panic ..."

in the fixes sections.

How about your firmware?

MP:CM> sysrev

Did the "pdcinfo" return something?

[S. Ney]

Depending on your OS system what was the output of the txt file from one of the tools you ran? After running either q4, crashlite, or crashinfo did you generate a crash.txt file? Was there any useful information in the txt file that you could decipher?

The /var/tombstone file is a binary file that the HP CE's will have to diagnose for you.(From the previously attached document) Email the file per instructions below (find it at: /var/tombstones/ts99 --binary file). -it is hardware.

After having gone through a similar scenario HP was able to determine for me what hardware sent the panic on my system after I sent them the ts99 file.

[CowBoy]

There is no file in crach dump directory.
I think i have to send the log to HP.
Thanks alot

[Sp4admin]

Did you check for any files in the /var/tombstone directory. There should be a ts## file which would point to a hardware problem.

sp,

[CowBoy]

yes "ts99" but it is not undrestandable....

[R.K. #]

Hi CowBoy,

Just check the timestamp of ts99. It should be latest

And you can open ts99 in..say wordpad.

If it is full of zeroes....then possibly not related to hardware.

Else issue is with hardware.

-R.K.

[CowBoy]

hi R.K
the time of ts99 is for the time of disaster and its content is full of zero...

[R.K. #]

Hi,

Probably we need to get the crash dump analysed to find out the exact issue.

R.K.