| Legal Information |
|
The disaster recovery processes for restoring Exchange 2000 clusters are similar to the processes for restoring data on stand-alone Exchange servers. However, before you begin to perform recovery processes on your clusters, it is helpful to understand how Exchange cluster resources can continue to remain on-line, even if one of the nodes experiences a disaster.
If one of the nodes in a cluster fails (known as a failover event), the Cluster service takes control of the cluster. Following the failover, one of the preferred nodes for the resource group attempts to take control of that group.
If all the resources are able to come on-line for the new node, that node continues to perform the tasks that were previously performed by the damaged node. If the resources fail to come on-line for the new node, that node will fail over to the next node.
This process continues until all preferred nodes for that group fail to come on-line.
Similarly, if one of the resources in an Exchange 2000 virtual server fails, the Exchange 2000 virtual server goes off-line, and then a preferred node attempts to start all the resources for that Exchange virtual server. If the resources fail to come on-line for the new node, that new node fails over to the next preferred node. If all nodes fail to bring the Exchange virtual server resources on-line, the resources on that virtual server will be unavailable to Exchange clients until the problem is resolved.
An important difference in disaster recovery processes for Exchange 2000 clusters is the task of identifying what caused a particular resource to fail. If a problem occurs, you should first determine if the failure is on a single node (which indicates that there are problems with the node's files) or on every node (which indicates that there are problems with the cluster's objects or the shared cluster resources). To determine the cause of the failure, search the event logs within Event Viewer.
If you are still unable to determine the cause of the failure, you can perform the repair options listed in "Repairing Windows 2000" or "Repairing Exchange 2000".
If repairing the node or entire cluster is unsuccessful, you must consider replacing the node or recovering the node, cluster, or resources (such as the quorum disk resource or mailbox and public folder stores).
The following provides the following procedural information about restoring Exchange 2000 clusters:
The information in this article applies to:
| Search Knowledge Base | Feedback |