Example 3 - Trial failover using the snapshot feature

Because the goal is to simulate a crash on the Primary, do not use IBC Messaging for trial failover.

  1. Pause the RLINK either on the Primary or the Secondary to maintain consistency.

    To pause the RLINK on the Primary, type:

    # vxrlink -g hrdg pause rlk_london_hr_rvg

    To pause the RLINK on the Secondary, type:

    # vxrlink -g hrdg pause rlk_seattle_hr_rvg
  2. When the RLINK is paused, take snapshots of all the data volumes in the RVG:
    # vxrvg -g hrdg -P trial snapshot hr_rvg

    where trial is the prefix for the snapshot plexes for all data volumes. Snapshot data volumes with names trial-hr_dv01 and trial-hr_dv02 are created.

  3. When the snapshots are complete, resume the RLINK by typing:
    # vxrlink -g hrdg resume rlink_name

    where rlink_name is the name of the paused RLINK.

  4. Start the application using the data volumes trial-hr_dv01 and trial-hr_dv02 that you have snapped off.
  5. Use the recovery function of the application to recover it, and then run the application. For example to recover a vxfs file system, use fsck.
    # fsck -F vxfs /dev/vx/rdsk/hrdg/trial-hr_dv01
    # fsck -F vxfs /dev/vx/rdsk/hrdg/trial-hr_dv02
  6. When the test is complete, shut down the application. For a file system, unmount the file system.
  7. Reattach the snapshot plexes to the original data volumes.
    # vxrvg -g hrdg -P trial snapback hr_rvg

    The -P option to the vxrvg snapback command reattaches to the original volume the plexes with the prefix specified when taking the snapshot.