At one of zimbra sites we administer and support we have experienced a strange postfix errorneous behaviour on Zimbra restart
All aour sites are configured for daily automatic maintenance (yes, backups too)
and usually it goes with no errors
This is todays maintenance script output head:
everything looks OK.
Wed Apr 13 02:00:01 MSD 2011 Backup started
Wed Apr 13 02:00:02 MSD 2011 Stopping zimbra...
Wed Apr 13 02:00:39 MSD 2011 Creating snapshot
Logical volume "zmsnapshot" created
Wed Apr 13 02:00:42 MSD 2011 Starting zimbra...
Wed Apr 13 02:01:48 MSD 2011 Mounting and backing up...
Usually this is how services log their startup in zimbra.log:
But postfix actually didn't start correctly (here's zimbra.log):
Apr 8 23:25:54 mx clamd: Database correctly reloaded (941823 signatures)
Apr 8 23:25:59 mx postfix/postfix-script: starting the Postfix mail system
First, I do not understand why timezone didn't apply for postqueue (look at date and time, these 2 lines are sequential in log file)
Apr 13 02:01:51 mx clamd: Database correctly reloaded (846254 signatures)
Apr 12 22:02:00 mx postfix/postqueue: fatal: Queue report unavailable - mail system is down
Second, there was no even a try to start up postfix and the first startup related postfix-script record in log appeared when we manually restarted the whole zimbra (service zimbra restart). Look at the maintenance script log output (the first code block): zimbra's reported that MTA had been successfully started.
We could not find any related to this either in /var/log or in /opt/zimbra/log. Does anyone have any ideas on how to investigate this?