Migrating to New Server

[sujithmk]

Hello,

I am running with a lot of errors while migrating the data from one zimbra server to a new server. I followed the steps in the link http://wiki.zimbra.com/index.php?tit...ine#Using_LDAP

The new server is setup with same hostname and zimbra version. But when I restart zimbra it fails to start ldap server and the error logs show the error message below

Nov 11 15:39:57 washington01 slapd[11795]: @(#) $OpenLDAP: slapd 2.3.40 (Jan 22 2008 14:44:40) $ build@build10.re.zimbra.com:/home/build/p4/main/ThirdParty/openldap/openldap-2.3.40.7z/servers/slapd
Nov 11 15:39:58 washington01 slapd[11796]: bdb_db_open: DB_CONFIG for suffix has changed. Performing database recovery to activate new settings.
Nov 11 15:39:58 washington01 slapd[11796]: bdb(): DB_ENV->log_flush: LSN of 2/1059990 past current end-of-log of 1/1492804
Nov 11 15:39:58 washington01 slapd[11796]: bdb(): Database environment corrupt; the wrong log files may have been removed or incompatible database files imported from another environment
Nov 11 15:39:58 washington01 slapd[11796]: bdb(): dn2id.bdb: unable to flush page: 0
Nov 11 15:39:58 washington01 slapd[11796]: bdb(): txn_checkpoint: failed to flush the buffer cache Invalid argument
Nov 11 15:39:58 washington01 slapd[11796]: bdb(): PANIC: Invalid argument
Nov 11 15:39:58 washington01 slapd[11796]: bdb_db_open: Database cannot be recovered, err -30978. Restore from backup!
Nov 11 15:39:58 washington01 slapd[11796]: bdb(): DB_ENV->lock_id_free interface requires an environment configured for the locking subsystem
Nov 11 15:39:58 washington01 slapd[11796]: bdb(): txn_checkpoint interface requires an environment configured for the transaction subsystem
Nov 11 15:39:58 washington01 slapd[11796]: bdb_db_close: txn_checkpoint failed: Invalid argument (22)
Nov 11 15:39:58 washington01 slapd[11796]: backend_startup_one: bi_db_open failed! (-30978)
Nov 11 15:39:58 washington01 slapd[11796]: bdb_db_close: alock_close failed
Nov 11 15:39:58 washington01 slapd[11796]: slapd stopped.
Nov 11 15:39:58 washington01 slapd[11796]: connections_destroy: nothing to destroy.

I am planning to get this restored in new server and have an upgrade to the version of zimbra to latest zcs-5.0.10_GA (OS). The current one I run is zcs-5.0.2_GA (OS).

I would appreciate any help as am working on to get this done for last 2 days.

Regards
Sujith

[phoenix]

These instructions would be (one of) the ones to follow. I'm assuming that you're moving to the same operating system and architecture?

[sujithmk]

Hello Bill,

OS on old Server : centos 5
OS New server : centos 5.2 ...

Arch is same....

Would the OS difference cause any issues ? Also, the mail server is a live one to which around 200 people are logged in . So its hard to stop the service and rsync.... Is their any other way to do it keeping the source server live ?


Regards
Sujith

[phoenix]

OK, if the architecture is the same then the change from CentOS 5 to 5.2 shouldn't cause any problems. During this migration there will be a period when the users will not have access to their mail, I'd schedule this move for a quiet period. Effectively what you need to do is stop incoming by by blocking port 25 (plus 80 and/or 443) at the firewall, shutdown Zimbra then tar the /opt/zimbra directory. Copy the tar file to the new server, install the same version of Zimbra (making sure that all your DNS records, hosts file etc are correct for the new server), stop Zimbra then untar the file over the newly installed version and then install Zimbra again. That should give you a working system, check the logs for any problems and open port 25 again and you should be set to go.

Depending on the size of your /opt/zimbra directory this procedure shouldn't take you very long, while it's possible to use rsync to perform this type of operation my preference would be for a cold backup.