Server freezes sporadically

[Adrian Kirchner <adrian.kirchner@gmail.com>]

Hi,

I have a small backup Server which runs on Debian 5.0 Lenny with all 
available stable updates.

For one month now this Server wasn't running for more than 24 hours 
because of problems which passes me by. I checked the syslog, messages 
and kernel log and the only anomaly I could find was a segfault entry, 
normaly immediately before the restart entry like this:

> Aug 10 06:33:16 LXL-CGN-01 kernel: [2041002.927632] perl[11972]: 
segfault at 157040 ip b7d563a2 sp bf803470 error 6 in 
libc-2.7.so[b7d40000+155000]

I thought it is a problem with libc and probably with the main memory. 
But the last few days, the server freezes without that messages. Now the 
last log entries look like this:

> Sep  2 16:44:35 LXL-CGN-01 -- MARK --
> Sep  2 17:04:35 LXL-CGN-01 -- MARK --
> Sep  2 17:24:35 LXL-CGN-01 -- MARK --
> Sep  2 17:30:01 LXL-CGN-01 rsnapshot[17080]: /usr/bin/rsnapshot 
daily: completed successfully
> Sep  2 17:44:35 LXL-CGN-01 -- MARK --
> Sep  2 18:04:34 LXL-CGN-01 -- MARK --
> Sep  2 18:24:34 LXL-CGN-01 -- MARK --
> Sep  5 12:53:24 LXL-CGN-01 syslogd 1.5.0#5: restart.
> Sep  5 12:53:24 LXL-CGN-01 kernel: klogd 1.5.0#5, log source = 
/proc/kmsg started.
> Sep  5 12:53:24 LXL-CGN-01 kernel: [    0.000000] Initializing cgroup 
subsys cpuset
> Sep  5 12:53:24 LXL-CGN-01 kernel: [    0.000000] Initializing cgroup 
subsys cpu

This looks to me like: "Nothing to do and then freeze without a loggable 
reason". Does anybody has an idea what I can do to determine the problem?

The whole /var/log/messages can be found here: http://pastebin.com/f194c41e2
The whole /var/log/syslog can be found bzipped here: 
http://www.box.net/shared/azgde0yqfg


Thanks in advance

Adrian