critical: services going down with daily cron

I have installed ISPConfig on several Vservers using Falko's perfect server for Debian Squeeze.

I got a big problem:

at server time 01:14, on *some* sundays, the following services go down:
sshd
pureftpd
courier
syslog (no more logging)
maybe more services go down, but these are the ones I noticed

at server time 01:14, it does the daily cron as well as log rotation.

here is what happens:

rsyslogd received a HUP and "exists on signal 15". Logs stop with that line. after that, nothing gets logged anymore.

in the general server apache2 error.log, I find:
[notice] Graceful restart requested, doing restart
and then some lines:
DBD::mysql::db do failed: MySQL server has gone away at /usr/local/ispconfig/server/scripts/vlogger 528
DBD::mysql::db st execute failed: MySQL server has gone away at /usr/local/ispconfig/server/scripts/vlogger 521
DBD::mysql::db do failed: MySQL server has gone away at /usr/local/ispconfig/server/scripts/vlogger 528
then the log ends

when this happens I have no sshd, no ftp and no mail,
BUT: apache and mysql are still running.

can someone help me debug this?

Thank you!

 

thank you Till

it is ISPConfig 3.0.4.5

I will ask the hoster about the virtualization technology - I'm sure it is not cloud tech, if that helps.

Should I upgrade to ISPC 3.0.5 ?

 

I forgot to list some errors from the logs:

in var/log/ispconfig/cron.log I found the following errors:

FATAL: could not load /lib/modules/2.6.32-5-vserver-amd64/modules.dep: no such file or directory
and I also got a message on ip6tables initialization failed

 

I upgraded the server to ispc 3.0.5.2

 

I don't know if this is related, but the fork rate at 01:00 am rises to 600 forks/second for about 10 to 20 minutes, and CPU usage triples, load jumps from 0.4 to 1.2 average for 30 minutes.

What is the server doing at 01:00 am?
This is the logrotation and server.sh scripts, right?