Configuration Settings for a Four-Node Cluster

  • Article

 

As shown in the following figure, the recommended configuration for a four-node Microsoft® Exchange Server 2003 cluster contains three active nodes and one passive node, where each of the active nodes contains one Exchange Virtual Server (EVS). This configuration is useful because it gives you the capacity of running three active Exchange servers, while maintaining the failover security provided by one passive server.

Recommended configuration of a four-node Exchange cluster

dffb0365-e309-4ecf-aebd-18180cd7410f

Note

All four nodes of this cluster are running Microsoft Windows Server™ 2003 Enterprise Edition and Microsoft Exchange Server 2003 Enterprise Edition.

The recommended four-node cluster can handle a single node failure at a time and maintain 100 percent availability after the failover has occurred. A second failure during this period leaves the cluster in a partially up state. To illustrate this concept, here is an example:

  • **First failure   **If Node 1 fails, Node 2 still owns EVS2, Node 3 still owns EVS3, and Node 4 takes ownership of EVS1 with all the storage groups mounted after the failover.

  • Second failure   If another node fails while Node 1 is still recovering from the failure, the Exchange Virtual Server on the second failed node tries to fail over to a node not hosting an Exchange Virtual Server. Because failover is not possible, the second Exchange Virtual Server remains in a failed state.

The following tables list the recommended configuration settings for this four-node cluster.

Exchange Virtual Server settings

Properties dialog box Tab Recommended settings

EVS1

General

Preferred Owners   Node 1

EVS2

General

Preferred Owners   Node 2

EVS3

General

Preferred Owners   Node 3

EVS1, EVS2, EVS3

Failback

Prevent Failback

This default option disables failback on each ESV. The administrator can move the server back at an appropriate time.

Exchange resource settings

Properties dialog box Tab Recommended settings

Exchange Resource

General

Possible Owners   All nodes are possible owners.

Exchange Resource

Advanced

Restart   

This default option enables Cluster Service to try to restart the resource after the initial failure of the resources.

To enable Restart, select the Affect the group check box with a threshold of 3 and a period of 900 seconds.

Pending Timeout   3 minutes (default)

As mentioned in "Setting Pending States," the Exchange store instance is not restricted by this setting when coming online.