If one node of a cluster is not functioning properly (for example, if it has failed over), as long as there is one functioning node in the cluster, you can simply replace the damaged node with a new node. For performance and fault tolerance purposes, you should repair or replace damaged nodes as soon as possible.

Because active/active clusters use load balancing, an active/active cluster that is missing a functional node results in decreased performance (since the remaining node has to assume the duties of the failed node).
Furthermore, an active/passive cluster that is missing one functional node decreases fault tolerance because, if that last node fails, you would have to rebuild the entire cluster.

It is relatively easy to build a new node and join it to the cluster. Replacement nodes can have any computer name, but they must have the same processor type and amount of RAM as the remaining nodes in the cluster. Although you can build a new node to replace a failed cluster node, it is recommended that you prepare a recovery server for your cluster in advance; such a server is called a cluster node recovery server. If you prepare the cluster node recovery server in advance, you can immediately apply the new node to the cluster after another node fails.
For information about how to prepare a recovery server, see "Preparing a Server to Replace a Failed Node"

To replace a node if you have not prepared a stand-by recovery server

1. To evict the damaged node from the cluster and remove it from the cluster's shared small computer system interface (SCSI) bus:
   1. Click Start, point to Programs, point to Administrative Tools, and then click Cluster Administrator. If Cluster Administrator does not automatically locate the cluster to which the node belongs, in the Cluster or server name box, type the name of an active node in the cluster. You can also connect to the cluster by opening Cluster Administrator on a cluster node and entering a PERIOD (.) in the Cluster or server name box.
   2. Stop the Cluster service on the node you want to evict.
   3. In Cluster Administrator, on the File menu, click Evict Node.
   4. From the Windows 2000 control panel, double-click Add or Remove Programs, and then click Add/Remove Windows Components.
   5. In Components, clear the check box next to Cluster Service, and then click Next to uninstall Cluster service.
   6. Remove the evicted node from the cluster's shared SCSI bus.
2. Install Windows 2000 Server, including the latest service pack that the server was running, software updates, and the software for your Exchange 2000 cluster nodes (such as anti-virus software). You can use any computer name that is not on the network for the new node. Do not install Exchange 2000 on the new node at this time.
3. Join this replacement node to the same domain as the other nodes in the cluster.
4. Connect the computer to the shared SCSI bus that the cluster is using.
5. Start the replacement node.
6. To add the replacement node to the cluster, from the Windows 2000 control panel, double-click Add or Remove Programs, and then click Add/Remove Windows Components.
7. In Components, select the check box next to Cluster Service, and then click Next to install Cluster service.
8. In Cluster Wizard, follow the steps to add the node to the cluster.
9. Install Exchange 2000 Server on the replacement node. Exchange automatically notifies you that the "cluster-aware" version of Exchange 2000 is being installed. (Applications that support the Cluster API are defined as "cluster-aware.") After Exchange 2000 Setup completes, install any Exchange 2000 service packs that are running on the other nodes in the cluster.

Important When you run Exchange 2000 Setup as part of rebuilding a cluster node, you do not run Exchange 2000 Setup in Disaster Recovery mode. The Disaster Recovery mode Setup option is not available when running Setup on cluster nodes. The Setup modes available for Exchange clusters are Install, Change, Uninstall, and Reinstall.

To replace a node if you have a stand-by cluster node recovery server

1. To evict the damaged node from the cluster and remove it from the cluster's shared SCSI bus:
   1. Click Start, point to Programs, point to Administrative Tools, and then click Cluster Administrator. If Cluster Administrator does not automatically locate the cluster to which the node belongs, in the Cluster or server name box, type the name of an active node in the cluster. You can also connect to the cluster by opening Cluster Administrator on a cluster node and entering a PERIOD (.) in the Cluster or server name box.
   2. Stop the Cluster service on the node you want to evict.
   3. In Cluster Administrator, on the File menu, click Evict Node.
   4. From the Windows 2000 control panel, double-click Add or Remove Programs, and then click Add/Remove Windows Components.
   5. In Components, clear the check box next to Cluster Service, and then click Next to uninstall the Cluster service.
   6. Remove the evicted node from the cluster's shared SCSI bus.
2. Start the stand-by cluster node.
3. To add the stand-by cluster node to the cluster, from the Windows 2000 control panel, double-click Add or Remove Programs, and then click Add/Remove Windows Components.
4. In Components, select the check box next to Cluster Service, and then click Next to install Cluster service.
5. From Cluster Wizard, follow the steps to add a node to the cluster.

The information in this article applies to:

• Microsoft Exchange Server 2003 Enterprise Edition
• Microsoft Exchange Server 2003 Standard Edition
• Microsoft Windows Small Business Server 2003, Premium Edition
• Microsoft Windows Small Business Server 2003, Standard Edition
• Microsoft Exchange 2000 Server