Recover Data on Failed NEC CLUSTERPRO/ExpressCluster SE Shared Disks

If the shared disk fails, but the cluster nodes are undamaged, perform the following steps to recover data residing on the shared disks:

To recover data residing on the shared disks

  1. On each cluster node, select Control Panel, Services, and change the Startup Type of the following services to Manual:
    • NEC ExpressCluster Server
    • NEC ExpressCluster Log Collector
  2. Shut down the cluster and turn off all servers.
  3. Turn off the shared disk and replace the shared disk if necessary.
  4. Turn on the shared disk, and set the parameters for the shared disk.
  5. If RAID reconstruction or LUN configuration change is necessary, use the setting tool attached with the shared disk. See the shared disk documentation for information about the setting tool.
  6. To perform any setting or configuration from a cluster node, turn on only one server at a time.
  7. On the primary cluster node only, perform the following procedure:
    • Write a signature (identical to the original) to the disk with the operating system’s disk administrator, if one does not already exist.
    • Recreate the original partitions on the disk. If X-Call settings have been performed to HBA, you must connect the partition using the NEC ExpressCluster disk administrator before formatting.

      • Note: X-Call is a setting that enables viewing of the shared partition from both the active and passive sides. See the CLUSTERPRO/ExpressCluster products document for more information about the setting for X-Call.

    • Using the operating system’s disk administrator, specify the original drive letter to the shared disk.
    • Use Arcserve Backup to restore the backed up data to the shared disk.
    • If you have performed X-Call settings for a disk, start the NEC ExpressCluster disk administrator and specify the recovered shared disk as X-CALLDISK in X-CALL DISK configuration.
    • If you have performed X-Call settings for HBA, these settings are not changed. Go on to the next step.
    • If the disk access path has been dualized, confirm that the access path is dualized.
    • Reboot the server.
    • Confirm that the drive letter is identical to the one you set in the previous step using the operating system’s disk administrator.
    • Check the cluster letters on the CLUSTER disk partition with the NEC ExpressCluster disk administrator. If the cluster letter does not appear, set it to the original letter.
    • Shut down the cluster node.
  8. Perform the following steps on all cluster nodes:
    • Boot up the cluster node.
    • Using the operating system’s disk administrator, specify the original drive letter to the shared disk, if necessary.
    • Set the Startup type of the following services from Manual to Automatic:
      • NEC ExpressCluster Server
      • NEC ExpressCluster Log Collector
    • Shut down the server and shut down the cluster node.
  9. Start all cluster nodes and perform the Return to cluster(R) operation from the NEC ExpressCluster Manager. Recover all servers to Normal.