| Sys Admin Extraordinaire |

2011-01-10 Updated

2008-12-11 Original post

Tested on Microsoft Cluster Server - Windows Server 2003 R2 Enterprise Edition, SP2 and Exchange Server 2003 Enterprise Edition, SP2 for both Windows and Exchange hotfixes.

These are my simplified instructions based on the instructions from How to apply Exchange service packs and hotfixes. I’ve only used this for updating hotfixes since both Windows Server 2003 and Exchange Server 2003 have not had any service packs out since SP2.

I’ve used the steps below to update a two-node active/passive cluster in production and test environments with no issues. I used http://update.microsoft.com and selected the custom option. Change the node names to suite your configuration.

Prior to any updates to a cluster you should fail over the active node to verify that failover works properly. There isn’t really a technical requirement to perform the failover, but you should perform the failover to ensure that the passive node is working properly before running any updates. This is included in the steps below.

1. Check the event logs on both nodes for errors and ensure proper system operation.
2. Make system state backups of both nodes and make a full or incremental backup of all Exchange stores.
3. In Cluster Administrator, right-click on NODE-01 --> click on Stop Cluster Service. This will automatically start the failover of all cluster groups/resources over to NODE-02.
4. Check the event logs on both nodes for errors and ensure proper failover and system operation before continuing.
5. In the services applet on NODE-01, set Cluster Service to disabled.
6. Before installing any updates, make a note of exactly which updates were installed on the first node so that only those same exact updates are later installed on the other node. You can copy and paste the list of updates from the Microsoft Update screen into a text file and then use the FC command to compare the file from the first updated node to the file from the last updated node. Here's an example command to compare two files:  C:\>FC X:\File1.txt X:\File2.txt. You can install updates directly from a file or from http://update.microsoft.com. Install Windows and Exchange updates on NODE-01 and then reboot the node as necessary. Repeat this step after the reboot until all updates are installed. That is necessary because some updates require that others be installed first. Microsoft Update will show the Exchange updates with the word “Cluster” at the end, showing that it understands that the server is part of a cluster. You’ll get a lot of prompts for the Exchange updates, so don’t walk off during the updates.
7. Check the event log on NODE-01 for errors. If you find any errors, troubleshoot them before continuing.
8. In the services applet on NODE-01, set Cluster Service to automatic.
9. Reboot NODE-01 one last time.
10. In Cluster Administrator, right-click on NODE-02 --> click on Stop Cluster Service. This will automatically start the failover of all groups/resources over to NODE-01.
11. Check the event logs on both nodes for errors and ensure proper failover and system operation before continuing.
12. In the services applet on NODE-02, set Cluster Service to disabled.
13. Install Windows and Exchange updates on NODE-02 and then reboot the node as necessary. Repeat this step after the reboot until all updates are installed.
14. Check the event log on NODE-02 for errors. If you find any errors, troubleshoot them before continuing.
15. In the services applet on NODE-02, set Cluster Service to automatic.
16. Reboot NODE-02 one last time.
17. Check the event logs on both nodes for errors and ensure proper system operation. You’re done.

This entry was posted on Monday, January 10th, 2011 at 5:26 PM and is filed under Disaster Recovery - Fault Tolerance - High Availability, Exchange Server. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.