Interstage Shunsaku Data Manager Operator's Guide - Microsoft(R) Windows(R) 2000/ Microsoft(R) Windows Server(TM) 2003 - - UNIX -
Contents Index PreviousNext

Chapter 8 HA Functions> 8.5 Recovering Shared Disk Faults

8.5.3 Recovery Procedure for Director Data Files (PRIMECLUSTER)

If the shared disk where director data files are located fails, replace the disk, and then recover the director data files on the active server.

The following diagram shows the procedure for recovering director data files when PRIMECLUSTER is used:

Before recovering director data files, stop all applications to prevent data searches and updates from occurring.

[Figure : Recovery Procedure for Director Data Files (PRIMECLUSTER)]

  1. Check the status of the cluster application.

    Check the status of the cluster application using the Web-Based Admin View.

    If the Shunsaku cluster application is 'online', stop it from the Web-Based Admin View.

  2. Replace the shared disk.

    Replace the shared disk where the fault has occurred. Then restore the configuration status for the volume.

  3. Activate the GDS resource.

    Activate the GDS resource to restore the director data files.

    This will enable the file system on the shared disk to be recognized form the active server.

    Refer to Activating GDS Resources in 8.4.2 Activating GDS Resources for more information on the procedure for activating the GDS resource.

  4. Create an area for storing director data files.

    Create a directory for storing director data files on the file system on the shared disk if one does not exist.

    Create the directory using the following operating system command:

    mkdir  storage directory for director data files
  5. Deactivate the GDS resource.

    Deactivate the GDS resource that was activated in Step 3.

    Refer to 8.4.5 Deactivating GDS Resources for more information on the procedure for deactivating the GDS resource.

  6. Make sure that all applications are stopped, then activate the Shunsaku cluster application.

    Start the cluster application from the Web-Based Admin View.

  7. Recover the director data files.

    Recover the director data files by executing the shundrecover command from the active director server.

    1. Make a recovery start declaration by executing the shundrecover command with the -b option specified.
      shundrecover  -s director-identifier -b
    2. Copy the backup data for the director data files to the director data file folder using the copy function provided by the operating system.
      cp  backup-data director-data-file-folder
    3. Make a recovery end declaration by executing the shundrecover command with the -e option specified.
      shundrecover  -s director-identifier -e

      This will cause Shunsaku to restore the director data files to their latest state by using the operation log files.

    An example is shown below.

    shundrecover -s shund1 -b
    cp  /backup/shund1/* /share1/shunsk/shund1/
    shundrecover -s shund1 -e

    These commands will complete the restoration of the director data files, allowing jobs to recommence.


Contents Index PreviousNext

All Rights Reserved, Copyright (C) FUJITSU LIMITED 2006