Reason for shutdown of HP-UX server

[HP-UX 11.31]

Hi,

We have HP-UX 11.31 itanium server (rx2660).
I am not able to find the server shutdown reason. Log files are attached please help.

Thanks,
Ranjeet

[Stephan. · Accepted solution]

Aug 9 10:31:03 docdev /usr/sbin/envd[1577]: ***** OVERTEMP_CRIT WARNING *****
Aug 9 10:31:03 docdev /usr/sbin/envd[1577]: Temperature exceeded the normal operation threshold. Correct the over-temperature condition.


It seems it's a little bit hot today ;-)

[Michal Kapalka (mikap)]

hi,

Aug 9 10:31:03 docdev /usr/sbin/envd[1577]: ***** OVERTEMP_CRIT WARNING *****

check thos web site :

http://www.informatik.uni-frankfurt.de/doc/man/hpux/envd.1m.html

or man envd

There are examples, how will the daemon escalating in this situations.

mikap

[Jamie A Dennis]

I don't see evidence in your log that OVERTEMP_CRIT escalated to OVERTEMP_EMERG, or that OVERTEMP_CRIT repeated.

How is /etc/envd.conf configured (what events are configured for these conditions)?

Also, what is logged in /etc/rc.log.old?

Best regards,


Jamie

[Steven E. Protter]

Shalom,

Run xstm cstm or mstm and check the hardware for faults.

Connect to iLo and runs sl show logs and look for a fan failure.

SEP

[Michael Steele_2]

I doubt very much that a critical over temp warning would knock down a server. Additionally, there would be fan fault lights that would be hard to miss if you were looking at your boxes every day, or checking for them in the EFI or BCH, where they will be listed as failed if failed.

I think it more likely that you've reset a tape drive or have a bad tape drive event, however, this is also unlikely to knock down you box and should only end up with a disabled tape drive unit that you could easily check for on console or inspection.

I think you need more information gathering, check /etc/shutdownlog, /etc/resmon/logs, MCA under /var/tombstones as well as alert events in the GSP.

[Bill Hassell]

I would be concerned about the OVERTEMP. High temperatures are lethal to big servers and disk arrays and wide temperature variations will cause major tape drive issues, especially unreadable data.

syslog is almost useless for finding the reason for a reboot. If the server crashes or otherwise goes down, there isn't enough time to wwrite anything in syslog. The first most useful log is /etc/shutdownlog. The next are the hardware logs as mentioned before.

I would also find out what facility and priority is being used by the Networker program. Ideally it uses a one of the local facilities so you can redirect all the messages into another logfile.