Recovering Multiple Cluster Nodes

Recovering Multiple Cluster Nodes

This section provides general guidelines for performing a cluster recovery in case both cluster nodes fail. In this scenario, the operating system is unusable on each node. Therefore, this recovery procedure includes reinstallation of Windows on each node, as well as NetWorker recovery of the cluster database.

Important: Because cluster configurations vary, it is not possible to provide cluster disaster recovery procedures for every situation. Depending on the particular cluster configuration and the nature of the failure, it might be necessary to vary some of the procedures described in this section.

How to Recover Multiple Cluster Nodes

To perform a complete cluster recovery in the event that both cluster nodes,

Node_A and Node_B, have failed:

1.Ensure that the system disk of each node is configured using the same partitioning scheme and drive letter that was in use prior to the loss of the node.

Note: On each cluster node, the repaired or replaced system disk must have the same partitioning scheme, disk format, and drive letter configuration that was in use prior to the node failure. However, drive letter assignments of available volumes during Windows setup may cause the system disk drive letter to change. Therefore, prior to reinstalling the Windows operating system, disconnect the SCSI bus if the shared drives are being enumerated before the local drives.

2.Format and verify the system disk on each node.

3.On each node, reinstall Windows to the partition it occupied before the node failure. During Windows setup on each node, verify that the domain controller is available and that each potential node is able to join.

4.Delete the MSCS folder on the quorum disk.

5.On the quorum disk, run the chkdsk command.

7

132

Legato NetWorker, Release 7.0

Disaster Recovery Guide