[CentOS-virt] can cron.daily crash my dom0s?

Mon Aug 18 09:07:29 UTC 2008
Heiko <rupertt at gmail.com>

I have 2 servers which share 2 partitions with drbd, on each machine
runs one VM on the drbd device,
so that I have primary/secondary and secondary/primary drbd devices.
There are also some more XEN VM´s that only do aa mysql replication
and one is standalone.
In the last 4 weeks I had 2 incidents where both machines did a sudden
reboot, first one machine and 2 minutes later the other one.
I cant find anything in the logfiles, only that a few seconds before
the secondary drbd device got a timeout
(Aug 18 04:02:14 xen-A1 kernel: drbd1: PingAck did not arrive in
time.) the cron.daily started:

Aug 18 04:02:01 xen-B1 crond[25033]: (root) CMD (run-parts /etc/cron.daily)
Aug 18 04:02:01 xen-B1 anacron[25037]: Updated timestamp for job
`cron.daily' to 2008-08-18
Aug 18 04:06:06 xen-B1 crond[2963]: (CRON) STARTUP (V5.0)

the last line seems to be the reboot.
here the content of my cron.daily folder.
[root at xen-B1.blab:/etc/cron.daily]# ls
0anacron  0logwatch  cups  logrotate  makewhatis.cron  mlocate.cron
prelink  rpm  tmpwatch

I had also heartbeat running on the machines, after the last crash I
thougt it was heartbeat that rebooted the machines, so I disabled it,
but now I see it isnt heartbeat that causes these reboots.
I use centos 5.1 with no selfcompiled packages.

Why are my machines doing this, and how to fix this?

greetings and thanks