Disaster recovery of a cluster

Prepare for recovery by creating a disaster preparation plan.

See About key elements of a disaster preparation plan (DPP).

Prepare to restore SQL, Exchange, Oracle, and Lotus Domino databases in a cluster after a disaster by reading the sections on preparing for disaster recovery in the appropriate chapters.

In addition to the initial preparation instructions, further action is required to completely protect the Microsoft clusters.

If a disaster occurs, the following information is required to successfully recover the cluster:

  • General Cluster Information

    • Cluster name

    • Cluster IP address and subnet mask

    • Cluster node names

    • Node IP addresses

    • Local and shared drive letters and partition scheme

    • Disk signatures

  • Cluster Groups

    • Group name

    • Preferred nodes

    • Failover/failback policies

  • Cluster Resources

    • Resource name

    • Resource type

    • Group membership

    • Possible owners

    • Resource dependencies

    • Restart and Looks Alive/Is Alive properties

    • Resource-related parameters

    • Application-specific configuration (SQL Database Character Set)

  • If you are recovering a Microsoft Cluster Server, run Dumpcfg.exe from the Microsoft 2000 Resource Kit or and the Clusterrecovery.exe from the Microsoft 2003 Resource Kit to retrieve the disk signatures from the shared disk. The Microsoft 2000 Resource Kit allows you to replace disk signatures.

  • If you are recovering a Veritas Cluster Server, run Vmgetdrive.exe to retrieve the disk signatures, disk group, and volume information from the shared disk.

Disaster recovery of a cluster