Zimbra Randomly Stopping
I have an instance of Zimbra OS Edition version 4.5.10_GA_1575 running on Ubuntu 6.06.1. This instance is running on a VM instance that is a P3 1Ghz with 1 GB of RAM.
I migrated this instance from iron that was running an unsupported version of Ubuntu (7.04). I restored this instance from backup using the various how-tos. It was working fine for a while and then all of a sudden started randomly stopping. I am new to Zimbra, so what logs should I send to help identify what might be causing the random stops? Finally, I had a thought that if this persists I would export my users, remove Zimbra, install from scratch and import the users.
Any help would be appreciated. Thanks!
When it stopps, can you get us the logs?
List of logfiles: /docs/ne/latest/administration_guide/9_Monitoring.12.1.html#1075561
Originally Posted by pikeman94
/var/log/zimbra.log & /opt/zimbra/log/mailbox.log are the big ones
For instance, if you run a su zimbra, zmcontrol status and find that the mailbox is stopped we would love to see /opt/zimbra/tomcat/logs/catalina.out
Here the logs you asked for. Thanks again!
Hmmm, The only thing I could find was an ldap problem.
When it goes down, I assume you do a zmcontrol status, right? Does it say "Cannot determine Services"?
And if you grep for processes, are there still Zimbra procs left over? If so, I think it's ldap that's going down.
Unfortunetly, I cannot think of any ldap logs. It may be because I"m tired.
Mods, any ideas?
Interesting. One more symptom I am noting as I watch the system. I started Zimbra and already free memory is down to 25MB. That is out of 1GB of RAM! What gives with that? Should Zimbra use that much memory when there are no active users on the system? BTW, my hosts file looks like this:
127.0.0.1 localhost.localdomain localhost
192.168.0.202 host.mydomain.com host
I am also running Bind on this server in a split DNS capacity. I don't think that is my problem as this has been working rock solid for almost a year. Do you think that migrating it to a VM could be the issue? Is there any merit to the rebuild from scratch idea? I am running out of options as Zimbra keeps crashing. If this is an LDAP issue, is there a way to do an integrity check on LDAP?
Again, thanks for taking the time to help me with this.
To be fair, the min requirement of RAM is 2 GB.
Do you have many IMAP/POP3 users? What happens if you turn of pop3/imap and bounce Zimbra.
My guess would be that you have a ton of IMAP threads, resulting in low memory. Once the memory reaches a certain level, java will do a core dump and stop.
There are two users of this system. I use this as a personal email server out of my home. I almost never connect via IMAP/POP, but instead use the web client.
As of this post Zimbra is still running and there is about 50MB of RAM free. TO turn off IMAP/POP would I use the zmlocalconfig command?
You can turn them off using the webAdmin UI : "Global settings" then "IMAP" and "POP" tabs.
You should restart tomcat once done.
I turned IMAP/POP off and now I have about 250MB of free RAM. I will keep an eye on it over night to see what happens. Is there any way to run IMAP/POP, but with fewer threads? Why does that component of Zimbra suck up so much RAM and again, why is it causing me a problem now after running for so long and on only 512MB?