Hi all,
I'm trying to move my Zimbra Community Edition install from an old server to a new one. They are both the same Architecture - x86_64. The old system is running CentOS 4.5, and the new one is running CentOS 5.1. I've spent a few days trolling through the forums and the wiki to help me out with the problems I've been having. Here is the process I've come up with so far:
- Make sure zimbra on the original system is up-to-date. In this case, I updated to 5.0.6
- Install a clean version of Zimbra (same version) on the new machine.
- Shut down zimbra on both systems.
- Sync over data directories - /opt/zimbra as the root, for store, data, conf, openldap-data, db/data, logger/db/data
- Run /opt/zimbra/libexec/zmfixperms on the new system (as root)
- Run /opt/zimbra/bin/zmhostname <newhostname> on the new system (as zimbra)
- Start Zimbra on the new system.
After following this process, all of the zimbra services start up. A zmcontrol status will show that all of the systems are running.
The problem that I am encountering is with zmmailboxd. My mail clients appear to connect to it, but basically hang. The system shows a network connection between my desktop and the server. The log file for zmmailboxd shows a large number of errors, with two lines repeating until a stack overflow error occurs:
Code:
at org.mortbay.jetty.webapp.WebAppContext.isTempWorkDirectory(WebAppContext.java:756)
at org.mortbay.jetty.webapp.WebAppContext.getTempDirectory(WebAppContext.java:718) Attempts to connect to the webmail client, or the admin console are met with 404 return codes:
Code:
HTTP ERROR: 404
NOT_FOUND
RequestURI=/
Powered by jetty://
From my limited experience with Jetty, and the research I've done, it appears that Jetty is trying to find a temp directory that doesn't exist, but I can't tell where, or if, I should create one.
So, the two big questions are:
1) Is there anything missing from my process above?
2) Where should I go from here to debug / resolve this issue?
Thanks in advance for any and all help.
-Steve J.