Mysterious reboots in J5000!!

[S. Karunanidhi]

Hi !
I have a J5000 which reboots intermittently. It works for 1 month and after that reboots thrice in one day and again works fine for one month. There is no entry in /etc/shutdownlog for these reboots. There is no crash dump available form /var/adm/crash. I think the machine is not crashing. No info. form syslog.(But syslog has info.about the previous reboot only. How do I keep all the logs of previous days also? Once the machine reboots, all the syslog is lost except for booting messages.)
I have reseated all the cards once in the machine.
Any inputs ?

Thanks in advacne
S.Karunanidhi.

[Andy Monks]

sounds like it might be an HPMC then. (Hardware problem). Reboot the machine and interrupt the boot process. Then at the admin menu, look at the pim information and see if it reports any problems. If so, call the HP support people to decode what it means.

[Cheryl Griffin]

Could be a power loss problem.

[Alan Riggs]

The previous syslog is kept in /var/adm/syslog/OLDsyslog.log
If you wish to preserve more than one generation just place a command in a startup script to copy this file to a date/time stamped archive locaation.

It does sound like you are having an HPMC, though. Look for entries in .var.tombstones. If you have online diagnostics, then you can get the chasis codes from the last crash from the appropriate entry. Call HP hardware support and have the chasis codes interpreted for you. That should identify the hardware error.

[Stefan Farrelly]

It certainly sounds like an intermittant hardware problem is causing HPMC's - ie. resetting the machine. Have you installed OnlineDiag and then run XSTM and go to tools -> logtool and see if any hardware errors get loggged ? if so send them to HP. Thats about the only way, without a dump, to determine whats causing it.

[Mark Mitchell]

I have had something like this happen to me once. HP came out and worked on my server's main system board. They happened to leave it set in a debug mode from the factory. This would cause the server to panic and reboot over anything as small as a single bit memory error that it would otherwise self-correct. The next time that it boots, try and halt it and see if there are debug options on the console menu. Its a shot in the dark but the symtoms are the same.