HPE 9000 and HPE e3000 Servers
1754885 Members
3753 Online
108827 Solutions
New Discussion юеВ

Re: System panic caused by CPU

 
Nick_104
Occasional Advisor

System panic caused by CPU

I have a 11.11 64 bit system that panics and reboots every so often. From the ts99 file I can see it's one of the CPU's that is causing this. I checked the firmware revisons on the CPU's were up to date, which they were, it's a D380 and the PDC Firmware Revision is 42.11, however I notice from cstm that the timestamp on the 'good' CPU is the time it last crashed whereas the time on the CPU that is causing the crash is sometime in May 2007. Is this a complete red herring or is this indicative of the problem?
I'll attach the output I got from CSTM. Any help will be appreciated.
7 REPLIES 7
V. Nyga
Honored Contributor

Re: System panic caused by CPU

Hi,

well for me it looks like it's the date when the problem occured the first time ....

Volkmar
*** Say 'Thanks' with Kudos ***
Nick_104
Occasional Advisor

Re: System panic caused by CPU

Thanks for the reply, but I don't think that's the case, the system has only started panic rebooting in the last couple of weeks. Since I updated the CPU firmware and installed 11.11 64-bit.
V. Nyga
Honored Contributor

Re: System panic caused by CPU

It's crashing after firmware upgrade?
Is it under support of HP?
So have you opened a call?

V.
*** Say 'Thanks' with Kudos ***
Nick_104
Occasional Advisor

Re: System panic caused by CPU

Unfortunatly it's not under cover.
Bill Hassell
Honored Contributor

Re: System panic caused by CPU

Processor time stamps in the ts99 are a clear indication of a CPU failure. Normally, these are identical and when a processor doesn't have the same time, it's time (;-)) to replace the CPU. This is true for a D380 and 2 processors or an rp8420 with 32 processors.


Bill Hassell, sysadmin
Michael Steele_2
Honored Contributor

Re: System panic caused by CPU

I can't tell for sure. Paste in /etc/shutdownlog and /var/tombstones/ts99. What I see is an HPMC for processor 0. But is you have a panic string in /etc/shutdownlog or a core dump in /var/adm/crash then I can tell you more.

HPMC stands for High Priority Machine Check and shouldn't be ignored. So regarding the red herring question, no, this is not a red herring. You never ignore HPMC's.
Support Fatherhood - Stop Family Law
Grayh
Trusted Contributor

Re: System panic caused by CPU

Hi,

Could you provide the output of the attached script for further investigation.

Procedure for script execution.

Use the binary mode to transfer the files. In case you use ftp mode , Copy the script in the working directory and execute the following command.

# sh getsysinfo.sh -t

The output of this script would be 'sysinfo.tgz' in '/tmp' directory.