I have a PC with RedHat Linux installed.
(The PC acts as server for a few third party applications.)

It runs without problems far more than a year.

But now, at regular times, weekly exactly the same time!, the PC 'freezes'?
2 Lights at the keyboard (NumLock and Scrollock?) and are flashing then.

There's nothing I can do at that moment besides a hardware reboot.

After rebooting I see in the messages files there has been no logging for a certain time, far before the reboot.
So the OS is 'freezing' in such a way there's even no logging possible.

The PC is part of a network.


How can I find out what the problem is?

Hi.

The 2 flashing keyboard lights means the kernel has panicked. I'd suspect either bad RAM, a bad power supply, or a heat problem. My money's on the RAM going bad.

Grab a copy of memtest http://www.memtest.org/ and boot from that to check your RAM's OK. Also have a look at the output from 'sensors' to see if it's running hot.

Dave

Ok, I'll try that.

How is bad RAM related to 'freezing' at the same second every week? (Same active job/process what asks for resources not used until then?)

Don't know. Is there a cron job that runs between the last message and the crash?

Dave

I don't know the time between a job and the crash.
The only thing I see, after rebooting in the messages logfile that both machines have at exactly the same time their last logged message. Both machines are 'frozen'.


I see I've forgotten to mention that it's a combination of 2 PC's, with a third party application in clusterconfiguration.

Does the clustering software include kernel modules?

Yes they do.
But it has worked for more than one year without problems. Nothing has changed in software/kernel versions.

Whetever the problem is, it's going to be in the kernel modules for the cluster software. Userland issues shouldn't panic a system.

Dave