Fedora8+Zimbra reboots every hour. Why?

[SergKz]

I have mailserver on Fedora 8 and Zimbra 5.0.4, Pentium D-930, 4G RAM, two SATA HDD as software mirror.
It works well all the year from February, but last week my server reboots every hour at 57 minutes without any reason. All the logs (system and zimbra logs) also don't contain any unusual messages, no unwanted records in cron configs, simply the normal process in logs stopped to 2 minutes and after than I see messages about reboot like this (zimbra.log):
=
Oct 30 21:57:02 sayan postfix/smtpd[15327]: lost connection after DATA from unknown[93.185.179.151]
Oct 30 21:57:02 sayan postfix/smtpd[15327]: disconnect from unknown[93.185.179.151]
Oct 30 21:57:06 sayan postfix/smtpd[11214]: connect from unknown[89.123.172.76]
Oct 30 21:58:24 sayan postgrey[1931]: Process Backgrounded
Oct 30 21:58:24 sayan postgrey[1931]: 2008/10/30-21:58:24 postgrey (type Net::Server::Multiplex) starting! pid(1931)
Oct 30 21:58:24 sayan postgrey[1931]: Binding to UNIX socket file /var/spool/postfix/postgrey/socket using SOCK_STREAM#012
Oct 30 21:58:24 sayan postgrey[1931]: Setting gid to "498 498"
Oct 30 21:58:24 sayan postgrey[1931]: Setting uid to "499"
Oct 30 21:58:49 sayan zimbramon[2133]: 2133:info: Starting services
Oct 30 21:58:50 sayan slapd[2174]: @(#) $OpenLDAP: slapd 2.3.41 (Feb 27 2008 16:01:59) $#012#011build@build20.re.zimbra.com:/home/build/p4/main/ThirdParty/openldap/openld
ap-2.3.41.6z/servers/slapd
Oct 30 21:58:50 sayan slapd[2175]: bdb_db_open: unclean shutdown detected; attempting recovery.
Oct 30 21:58:51 sayan slapd[2175]: slapd starting
=
The first three lines is normal process, and the forth line already written after reboot.
I was tried to stop any additional services like ftp and acpi, but it reboots again at 58 minutes.
Any ideas?

[y@w]

Is it a hard reboot? Does the server just stop running and then starts booting?

[SergKz]

Originally Posted by y@w

Is it a hard reboot? Does the server just stop running and then starts booting?

Not any services or daemons stops at reboot, and it isn't any messages of shutdown in logs. It's like the effect of "reboot -nf" command. Also, the reboot seems not like cron job result, because it slowly shifted the time of reboot - now it is already 01 minute of each hour.

[SergKz]

The problem solved with fsck.ext3 on reboot.
It's very strange....