I am having a problem with Centos 4.4.67-4 . Periodically the system goes into an undefined state. The network services and console services both go down, although the box is powered up. In consequence it is not possible to see what, if anything, is going on.

The message log would indicate the box is not running anything (there are regular 10-minute cron jobs which do not appear in the message log aftter a certain time). I suppose its possible that because I have to hard reset the things may have still been running but the output was not flushed to disk - personally, I doubt this.

The logs in /var/log are unhelpful. Anybody got any ideas ? Where should I look for clues as to whats going on ?

Your question is very vague. The system stops responding.... What's the last thing that happens in the logs prior to things hanging? Does the system hard lock or is just not accessible remotely?

Something has to be shown in the log files...

Nothing, other than the running of a cron job (sa1) . No alerts, rien.


I do however have a good idea of what it might be. It appears to be always doing a tar when this happens. Only this is a tar of an NFS mounted filesystem, the fileserver being a Windows environment. But no, alas the logs are quite useless.