Patching Exchange 2010 DAG Clusters


exchangelogo

  • Tip: Download and manually install rollups and service packs via an administrative command line. Soooo many problems can come from updating these via WSUS or automatic updates. Run these first, then reboot if need be and install the remaining patches. This can save you many hours of heartache.

Prerequisites

Patch your servers

  1. Run Exchange Management Shell as administrator.
  2. Put the server into maintenance mode (must be run on the server being placed in maintenance mode):
    1. maintenancewrapper –server <servername> –action start
  3. Run one of the following to verify that the server is in maintenance mode and that the PAM and database have been moved:
    1. This script will give you quite a few stats for the cluster and servers:
      1. checkpam
    2. This command will list the PAM and any servers in maintenance along with the server containing the mounted database:
      1. get-databaseavailabilitygroup <dagname>-status | fl name,primaryactivemanager,serversinmaintenance;get-mailboxdatabase | fl server,name
  4. Once the server is in maintenance mode and the PAM and database are active on the other server, patch and reboot.
  5. Once patches complete run the following to remove the server from maintenance mode, then verify using the checkpam script:
  6. Stop maintenance mode:
    1. maintenancewrapper –server <servername> –action stop
  7. Run checkpam to get health and status of the DAG:
    1. checkpam
  8. If the server you want to does not take over as cluster manager, run this command:
    1. cluster <dagname> group “cluster group” /moveto:<servername>
  9. If the database does not activate on the server, follow these steps:
    1. Open the management console
    2. Expand ‘Organization Configuration’
    3. Highlight ‘Mailbox Database number’ in the top pane
    4. Right-click the same database name in the bottom pane and select ‘Activate Database Copy’
    5. Choose the default option of ‘NONE’ and click OK

Troubleshooting

The cluster file share witness will not come back online
  1. Go to Server Manager → Features → Failover Cluster Manager
  2. Highlight <dagname.fqdn>
  3. In the right hand Actions pane, click on More Actions → Configure Cluster Quorum Settings
  4. Click Next and copy the path to the share, then paste it somewhere for later.
  5. Click Previous and choose Node Majority, then Next and finish
  6. The cluster will then be in Node Majority, meaning that if either server goes down, the cluster is dead and so, too, is the Exchange database. So do not reboot.
  7. Once done, go back into Configure Cluster Quorum Settings and choose Node and File Share Majority
  8. Click Next, paste in the path to the share copied in step 4, click Next and finish
  9. Under Cluster Manager, you should now see 2 shares listed. Highlight the file share that has failed. It should have a red X next to it. Remove this resource.
  10. The cluster should now show healthy. Determine this by running checkpam on either server.
  11. Manually move the active database to another server
Mailbox database won’t move to server
  • move-activemailboxdatabase -identity ‘mailbox database number’ -activateOnServer <servername> -mountdialoverride ‘none’
  • This can be accomplished through the Exchange Management Console as well via step #9 above