Testing restore on OSE with missing users upon completion

[dmmincrjr]

I am a rather new user to Zimbra and am attempting to verify that the daily backup I am making will allow me to restore a system. I am following the OSE rotating backup script from this thread OSE rotating backup script and the backup is being produced. I have set up a test environment on another machine to try and perform the restore. I have been fooling around with this for awhile and am stumped as to why all the users are not being restored. I have 10 users on the system and only the first 5 are being restored. I am using Centos 4.5 and Zimbra 4.5.5 on both machines. The test machine only has 1GB of memory and I have assigned it a different IP address. I have configured the test machines DNS the same as the production machine with the exception of the different IP address of the test machine inserted in the config files in a split DNS format from the Wiki. When I do host command it resolves back to the internal IP address of the test machine. I used rsync to copy back my latest backup to /opt/zimbra. I then used ./install.sh and performed an upgrade. Attached is the log for this process. I did get a warning of port conflicts on 389 and two on 25 during the process which I accepted. The process continues on and at the end I can log on the the admin console but only have 5 of the 10 users. I have also changed permissions using /opt/zimbra/libexec/zmfixperms and chown -R zimbra:zimbra /opt/zimbra/st* thinking that might be the issue but still no joy. I have not tried sending emails because the mta service is down but at this point I am more concerned about the missing users.

I also am unsure if this has any bearing on the problem but on the initial install of Zimbra I had made the mistake of creating the domain as mailhost.example.com and not the domain of example.com. I just went through and added this domain as well as another domain that will be receiving mail on this server and then deleted the mailhost.example.com domain. I also corrected the email address of everyone that had a mailhost.example.com domain. I see in the logs that it is trying to create email addresses for mailhost.example.com for admin users. With this issue I am also concerned going forward and trying to upgrade to newer versions if this will present a problem.

Any insights would be appreciated.

Thanks.

[jholder]

Hi dmm,
Welcome to the Zimbra forums!

Forgive me, but I couldn't find the question.

Also, looks like you may have an issue with LDAP

Best,
john

[dmmincrjr]

Sorry if it was buried in the text but I was trying to provide information in my post to figure out what happened to the 5 missing users and how do I go about getting them back in Zimbra? Two of the 5 users were created during the install of Zimbra. One user was the wiki and the other was spam. The other 3 users I created and migrated mail from my old email server. I realize from the logs I probably have an LDAP problem but am not that knowledgeable to figure out how to solve. What other information can I provide to help diagnosis the issue. I have looked under /opt/zimbra/store/0/ and I have 10 folders and can look at the messages so I know they are being copied back. During the upgrade process it stops saying I have a Port conflict detected: 389 (zimbra-ldap). However before running the upgrade I shutdown Zimbra and run ps auxww | grep ldap and it is not running so I am not sure why that is occurring.

To close out a statement in my original post I did find where I had exim running and that was causing a conflict with postfix, which is why I had the port 25 errors in the upgrade. Once that service was turned off postfix started and can send and receive mail internally.

Any help with restoring the missing users would be appreciated.

Thanks.

[dmmincrjr]

To post a resolution in case anyone else is using this same backup script I discovered it was a permission issue during the rysnc procedure and not all the files were being copied. Once that was corrected I was able to restore without a problem.