Migrating from Zimbra 4.5.x to 5.0 beta?

[jbwiv]

Guys,

I'd like to try the 5.0 beta out.

I run the open source version at home and have for sometime...and I really enjoy it. I'm also the CIO of a company of roughly 3500 people, and am considering the possible replacement of our Exchange server, which I despise, but that's a different story (I've registered for the webinar ).

I really like what you guys are doing and hope you keep it up. Google should take lessons from you.

NOW (after the flattery), my question. I'm running 4.5.6 on Debian 3.1 currently, but as I stated above, I wish to try the beta. I've backed up my /opt/zimbra directory, installed Fedora 5 and installed the latest beta on top of that (on a different OpenVZ VM). Now, I'm curious what I need to do to use my old data from 4.5.6 in the new system. Is it enough to copy over /opt/zimbra and run "install.sh -u" from the beta untarred directory? The restore procedure on the wiki seems to indicate that install.sh -u is the wrong way to go for restore of backup, but I'm not certain in this case. If the above is not right, what are the steps to do so?

Is there a clear upgrade path here, or am I stuck where I am until 5.0 comes out?

Thanks!
jbwiv

[jbwiv]

Ok, so far, I've done this:

- install 4.5.6 on the new server.
- copy over /opt/zimbra from old server
- re-run 4.5.6 install.sh and choose upgrade
- verify functionality works after this
- install 5.0 beta, and choose upgrade

Now most seems to be working. I modified the clamd interval as described in other parts of the forum, otherwise it kept restarting every two minutes and consuming lots of CPU.

Now, only thing I'm having trouble with is that the admin interface says "stats" is not running, and the refresh of mail queue data never succeeds. Any ideas?

Thanks,
jbwiv

[ArcaneMagus]

While zimbra is running as the zimbra user run /opt/zimbra/libexec/zmlogprocess. You probably have at least one of the logger tables marked as crashed and this process will show up this issue with a line like "Table mta is marked as crashed" or similar. If this is the case then this thread has saved me from this situation a couple of times. The mta and raw_logs tables are often marked as crashed if the server got rebooted without shutting down the zimbra service.