Thursday, April 7, 2011

Clustering the SCOM Microsoft Exchange Monitoring Correlation Service Part 2

In part 1 of this guide, I explained what the Microsoft Exchange Monitoring Correlation Service was and in this part, I will demonstrate what is needed to make this service Highly Available.

Follow the steps below to begin your configuration:

  • Run the .msi installer for the Exchange 2010 MP on all cluster nodes first as this creates the ‘Microsoft Exchange Monitoring Correlation’ service
  • Go to the ‘services.msc’ snapin on each cluster node and stop the ‘Microsoft Exchange Monitoring Correlation’ service 
  • Set the startup type for this service on all nodes to ‘Manual’ 
  • Edit the CONFIG file located at 'C:\Program Files\Microsoft\Exchange Server\V14\Bin\Microsoft.Exchange.Monitoring.CorrelationEngine.exe' for the Microsoft Exchange Monitoring Correlation Service to reflect the Virtual RMS Cluster name (it is set as 'Localhost' initially and this is what needs to be changed)

  • Provision a LUN from your shared storage and present it to the cluster nodes (1GB or 2GB should be more than enough)
  • From one of the cluster nodes, open 'Failover Cluster Manager' from the 'Administrative Tools' menu, expand your cluster name, right mouse click on 'Services and Applications' and then click on 'Configure a Service or Application' to open the 'High Availability Wizard'

  • Select 'Next' and then click on 'Generic Service'

  • Click 'Next again and then select 'Microsoft Exchange Monitoring Correlation' from the 'Select Service' window

  • Click 'Next and then input in the name and IP address that you are going to assign to the newly clustered service

  • Select your shared storage that you had presented to the cluster previously

  • Don’t bother selecting any registry settings to transfer over as the service is already installed on the other nodes. 
  • Click on ‘Next’ until you come to the review page, then click on ‘Finish’ to complete the service clustering. 
  • Test service cluster failover by moving the service between nodes using ‘Failover Cluster Manager’ and keep an eye on the service state from the ‘services.msc’ snapin as this should go from started to stopped on the current node, then from stopped to started on the new node you have chosen to move the service to. 

This completes your Microsoft Exchange Server 2010 Monitoring Correlation Service clustering.

All that’s left to do now is to import the Exchange 2010 MP into your environment and then tune away those noisy alerts (if any!)



No comments:

Post a Comment