unexplained crashes - how can I log them?
 

I have a linux box running debian 5.0 which has been up and running for a couple months now. It's mainly setup as a fileserver with 6 harddrives. 5 PATA and 1 SATA running on an ASUS motherboard. I have a promise controller with the extra PATA drives connected and a silicon image controller with the SATA drive connected.

It's crashed a couple times with no error messages on the screen, just froze. When it boots back up I check the syslog, dmesg log and other log files and can't find anything indicating the issue.

If anyone has any suggestions how to log this and find out what's wrong the next time it freezes I'd appreciate it. As of now I can't find a single clue in any of the logs once the machine has rebooted.

If you need any other info about the hardware please let me know.

I'd suggest trying netdump. If there's a kernel panic it should be able to log the information to a remote server. I think that with kexec and the like there may be more advanced tools, but netdump is what I know. You can find a little article about it here.