I mentioned this at the Ipswich meeting, and I’ve not yet had a chance to do anything about it, so I thought I’d see if anyone here had any ideas too :-)

My server crashes roughly once a month.

It's not consistent on the day that it crashes, sometimes it can be weeks apart. But it's usually just after 7am (although today I had my first crash at 13:30).

The locking up is a bit odd, as the PC still responds to pings, but I can't log in (I can type the username, but it never requests the password), and Apache stops serving files, but samba doesn’t.

I’ve got two scripts that run, an rsync backup script that runs at 7am, and one that gathers system info, which runs every 15 minutes.

I thought the problem might be something clashing between the two scripts, so I tried stopping the system info script running between 7am and 8am, this obviously hasn’t helped.

At the Ipswich meeting it was mentioned that it could be the syslog failing, so I’m going to look into redirecting the syslog to another server.

Is there anything else that anyone can think of?

System:
Debian Stable (Sarge, 3.1), 2.6.8-2-686-smp kernel
Dual 1Ghz P3
ASUS CUR-DLS motherboard
1Gb of ECC PC100 RAM
2x Silicon Image 3114 SATA cards
6x 250Gb SATA drives in RAID 5 (mdadm)
2x 40Gb PATA drives
HP 7170 Dual Port Gigabit card (64-bit PCI)

Cheers,

Matt