HOW TO: Restore an Exchange 2000 Cluster After a Complete Cluster Failure

Use the following procedure to restore a Microsoft Exchange 2000 cluster following a complete cluster failure.

Restore an Exchange 2000 Cluster

NOTE: The following example assumes:

�	You are running a Microsoft Exchange 2000 cluster with two nodes, NodeA and NodeB.
�	NodeA is the active node.
�	The whole cluster suffers catastrophic failure.

To restore services on to completely new hardware, follow these steps.

1.	Install a new node named NodeAA.
2.	Join the new node to the domain, and then give it a new IP address
3.	Set up a new Q drive ready for the Quorum (where Q was the drive that previously had the Quorum resources on it).
4.	Set up a new T drive ready for the mdbdata and other folders (where T was the drive that previously had the Information Stores on it).
5.	Configure the cluster. Make sure that it has the same name (for example Cluster1) and the same IP address (for example 192.168.178.210). Give it Q as the quorum device.
6.	Install Exchange 2000 Server. You receive a message that states that it is the cluster-aware version. Binaries are laid on to the disk.
7.	Restart the node.
8.	Create a new resource group with the same name that you previously used (for example EVS1).
9.	Create an IP address resource in the newly created resource group, with the same IP that you previously used for the virtual server (for example 192.168.178.230)
10.	Create a network name resource, with the same name as before (for example EVS1). Make that dependent on the IP resource.
11.	Create a new physical disk resource, using the same disk as before (for example T) and using the same drive letter as before. You may be able to drag and drop this out of one of the other resource groups.
12.	Create a new System Attendant resource. Make that dependent on the network name and physical disk.
13.	Apply the appropriate Microsoft Exchange Server Service Pack.
14.	Reboot the node.
15.	Dismount the stores.
16.	View the properties of the Mailbox Store and the Public Folder Store. Make sure that the checkbox that says that the database can be overwritten by a restore is selected.
17.	Dismount the Mailbox Store and the Public Folder Store.
18.	Create a temporary folder on T:, which is where the log files from the backup will be placed (for example T:\templogs)
19.	Do a restore, pointing the temporary logs back to T:\templogs. The place to restore to is the virtual server name (for example, EVS1). Click to select the check box that says it is the last backup in the set.
20.	Mount the stores manually.

You can now install the second node into the cluster. You may also encounter an issue where the mailboxes appear orphaned when viewed in Exchange System Manager. If they do not reconnect after a few minutes, you have to try to run MBCONN to reconnect the mailboxes.