Page 1 of 1

Troubleshooting Zimbra Outage

Posted: Wed Sep 26, 2007 11:56 am
by swu
We recently had an "outage" - users unable to access Zimbra. Their browsers would get the "page not found" errors. The Zimbra admin gui was fine, ssh to the server was fine, mail was still coming in (mail.log) - so which services should we start checking to see where the issue is?
Our resolution this time was to do a shutdown -r and Zimbra came back up just fine and users were able to get their email. Next time we'd like to know where to look and if it's restarting a service or something of that sort, go that route.
Thx for any advice.

Troubleshooting Zimbra Outage

Posted: Wed Sep 26, 2007 12:51 pm
by gmsmith
Before you rebooted did you do a zmcontrol status to see what processes were running. Almost sounds like Tomcat died.
Did you sort through the mailbox.log in /opt/zimbra/logs/? Anything abnormal in there?

Troubleshooting Zimbra Outage

Posted: Wed Sep 26, 2007 1:33 pm
by mmorse
Building on that; after you dig through mailbox.log-then check out: /opt/zimbra/tomcat/logs/catalina.out
Log file list is here: /docs/ne/latest/administration_guide/9_Monitoring.12.1.html#1075561

Troubleshooting Zimbra Outage

Posted: Wed Sep 26, 2007 2:01 pm
by kirme3
SWU, it looks like you are using the NE edition. You should probably open a support case about it, if you can. I've had quite a few times where tomcat stops, hangs, doesn't respond..etc..on our Mac install. The first thing I always do when there is a no response issue is run zmcontrol status.
The typical logs I get for support are:

/opt/zimbra/log/mailbox.log

/opt/zimbra/tomcat/logs/catalina.out

/opt/zimbra/tomcat/logs/stacktrace.xxxx(if there is one)
The logs are pretty easy to decipher about what happened. I also run ps auxww to see what processes are running..sometimes helps.
If it is just tomcat stopped, a faster way to bring it up than a reboot is to issue a tomcat start as zimbra user. You may have to issue a zmmailboxctl start too.

Troubleshooting Zimbra Outage

Posted: Mon Oct 01, 2007 9:58 pm
by swu
Thanks for all the replies and tips. I wasn't quite sure where to start poking and it was prime time for email to be up. I've taken notes and should it happen again I'll take the time to run the checks mentioned.
Does the admin GUI use a different instance of Tomcat? I thought maybe Tomcat gave up, but I could still get in through the admin GUI.
Thx again!