Thursday, October 24, 2013

Exchange backup failing due to ESE 474, how to move mailboxes

Problem
I am using Windows Server Backup to backup my Exchange Server database and noticed that the backup job has an error saying "the application will not be available for recovery. The consistency check failed for component Microsoft Exchange Server...", and I also noticed that transaction logs were not getting flushed.

Troubleshooting and Solution
Looking in the Event Viewer I found event 474, ESE, which states that a page level verification failed due to a page checksum mismatch. Googling lead me to this: http://technet.microsoft.com/en-us/library/bb397387%28EXCHG.80%29.aspx

First, I ran eseutil /K against the mailbox database. This showed me that there were indeed consistency check failures.

Following the advice in the article I chose to try moving mailboxes to a new database, being sure to choose the option to not move mailboxes if they had errors. If they had errors I would then run eseutil /P against the mailbox database that had corrupt mailboxes in it.

Step 1 - Create a new mailbox database. This can be done in the Exchange Management Console is straightforward.

Step 2 Prep the server to speed up the mailbox migration process
a. using the advice here: http://edmckinzie.wordpress.com/2011/12/12/how-to-speed-up-a-large-exchange-2010-migration/
b. Turn on circular logging

 Step 3 - Move the mailboxes in bulk, using the advice here: http://www.msexchange.org/articles-tutorials/exchange-server-2010/management-administration/moving-mailboxes-exchange-2010-part1.html

Step 4  -   Check to see if any mailboxes failed to move.

a. If they did, run eseutil /K on the old database, then eseutil /P.
b. After repairing try to move the mailboxes that failed

Step 5 - Dismount the new DB and run eseutil /K on it
If there's problems it probably means there's a hardware failure. Please read more on here about the ESE 474 problem: http://technet.microsoft.com/en-us/library/bb397387%28EXCHG.80%29.aspx They say sometimes it's a hardware failure and sometimes it's a logical file problem.

Step 6 (if no problems in step 5) - Disable circular logging. Then remount the DB.

Step 7 - Run a full backup on the new DB and check for problems.

No comments:

Post a Comment

There was an error in this gadget