server hanged per hour

[Robin Laing]

Rilawich Ango wrote:
> The hourly reboot is abnormal.  I have no rebooting job in
> /etc/cron.hourly.  The reboot seems caused by something (hardward,
> kernel) that I don't know.  The server runs for several months without
> problem.  Hourly reboot is found this week but it keeps rebooting.  We
> have hard reset the server but reboot still alive.
> Below is the complete log message when it rebooted.  You can see
> 09:36:07 the server rebooted itself.  Next reboot will be 10:3x.
> Anyone can help to find out what problem it is?
> 
> Fedora 8
> kernel: 2.6.23.1-49.fc8
> Intel(R) Core(TM)2 Duo CPU E4500 @2.20GHz
> 
> Jun 12 08:35:51 localhost ntpd[1710]: frequency initialized 148.719
> PPM from /var/lib/ntp/drift
> Jun 12 08:36:19 localhost gpm[1841]: *** info [startup.c(95)]:
> Jun 12 08:36:19 localhost gpm[1841]: Started gpm successfully. Entered
> daemon mode.
> J

If this is how your logs go, I suspect it is related to the time setting.

Two issues that I would look at.

1.  How far out is your time?  Manually synchronize your time to see if 
the realtime clock is causing the issue.  If you can turn the server off 
and see how accurate the time stays without power.  May only require a 
new battery on the motherboard.  Also turn off automatic time update and 
see if the reboots stop.


2.  Check each application in cron.hourly and see if one of them is 
crashing our system.



-- 
Robin Laing