Metro/Global Mirror planned intermediate site outage

With the incremental resync support installed, if there is an intermediate planned outage, to restart the recovery environment running Global Mirror from the local site to the remote site, do the following steps:

Assumptions:

  1. Quiesce of host applications is not needed because operations will continue to run from the local site.
  2. Metro Mirror pairs between the local and intermediate sites have been suspended in preparation for the planned outage.
  3. Global Mirror session at the intermediate site has been stopped in preparation for the planned outage.

Recovery steps:

  1. Perform failover commands to the C volumes at the remote site (PPRC establish C volumes to B volumes) with the cascade option.
  2. Establish paths between A volumes' LSSs and 'C' volumes' LSSs.
  3. Establish A to C pairs using the incremental resync change recording bitmaps (specify INCRES in the MODE parameter) to copy changed tracks. The configuration at this point is from volumes A to C, and then C to B, with A to C running PPRC-XD and the C to B pair being in failover state.
  4. Establish paths from control unit that is the intended master of the global Mirror session at the primary (local) site and the intended subordinate control units at the primary site.
  5. Define the Global Mirror session to all LSSs that are going to participate in the session at the primary site.
  6. Join the A volumes to the Global Mirror session.
  7. Start the Global Mirror session at the local site.
Note: To decrease the amount of time needed, steps 2, 4, 5, and 6 can be done before an outage, but the session is not started until needed.

After the intermediate site has been recovered, to restore operations to the original Metro/global Mirror environment, do the following steps:

  1. Establish paths from the remote site C LSSs to the intermediate site B LSSs, if necessary.
  2. At the intermediate site secondary devices (B volumes), issue CRECOVER or PRECOVER commands to terminate the relationships from A to B.
  3. Suspend the relationships from B to C, if the relationships have not been suspended.
  4. Perform Failback from C volumes to B volumes with the cascade option.
  5. Start Incremental Resync on the A to C volume pairs with the INCRESYNC NOCOPY option.
    Note: This step is used to enable the A to B pairs to be restored by using the Incremental Resync change recording and the FORCE option.
  6. Stop the Global Mirror session between the local and remote sites. Allow the data to continue to cascade to the C and B volumes using Global Copy.
  7. Suspend the A to C Global Copy pairs, and wait for all data to drain from the C to the B volumes.
  8. At the remote site Global Copy secondary devices (C volumes), issue CRECOVER or PRECOVER commands to terminate the relationship from A to C.
  9. Suspend the relationship from C to B.
  10. Issue established PPRC commands with failover from B to C and with the cascade option.
    Note: At this point, A is a primary to C and B is in failover state to C, but C is still a cascaded state from A to B, and is suspend, because no data is being transferred from the local site (A volumes).
  11. Perform the failback from B to C by issuing the Failback commands at site B with the cascade option
  12. Establish A to B pairs using the incremental resync change recording bitmaps (INCRES in the MODE option) to copy changed tracks from A to B with YES specified for the FORCE option to bypass relationship verification.
  13. Start the Incremental Resync function for the local PPRC volume pairs (A volumes to B volumes).
  14. After the A volumes to B volumes pairs have reached full duplex, start the Global Mirror session at the intermediate site.
Parent topic: Recovery scenarios for Metro/Global Mirror with incremental resync