Failure as a result of failback to the local storage system

When the local storage system or the local Data Center is operational again, processing can once more be moved back to the local storage system and the local Data Center. The applications must be terminated on the standby system and restarted on the local system. If the home pubset is also affected, the standby system must be shut down.

In symmetrical configurations (known as X configurations) failure as a result of failback can be prevented or deferred to a more suitable time using the SWAP-REMOTE-COPY function. Without interrupting the application, this function enables source units to be changed into target units and target units on which the applications run to be changed into source units. From this point onwards, secure operation with remote copy data replication is ensured, even without failback.

Actions in the remote Data Center

1. Terminate the applications and export the pubsets

2. Switch to the source path (i.e. set the target unit to NOT-READY or READ-ONLY).
   If the remote storage system contains only target units affected by the failure, the entire remote storage system can be specified. Switchover takes place simultaneously for all the devices affected.

   /SET-REMOTE-COPY-ACCESS
     UNIT=*BY-SYMMETRIX(SERIAL-NUMBER=<remote serialnumber>
     ,LOGICAL-VOLUMES=*ALL-TARGETS)
     ,TARGET-ACCESS=*BY-SOURCE
   

   For this function, the values DETACHED-DEVICES=*ACCEPT and NOT-DEFINED-DEVICES=*ACCEPT must be set for SYSTEM-ADMIN-SCOPE or TASK-ADMIN-SCOPE.

   The default setting causes all units generated in BS2000 to be set to NOT-READY and all other units to READ-ONLY. Target units in BS2000 must be NOT-READY as otherwise duplicate (i.e. non-unique) VSNs would result in the system. The READ-ONLY status is required for other target units as these are not otherwise be recognized by the relevant system. Target units that are not correctly included in the default setting must be processed separately.

   If the remote storage system has target units other than those affected by the failure, these must be specified separately. Device selection depends on the settings for SYSTEM-ADMIN-SCOPE or TASK-ADMIN-SCOPE and is made using:

   /SET-REMOTE-COPY-ACCESS
     UNIT=*BY-VOLUME(VOLUME=<vsn>) / *BY-PUBSET(PUBSET=<cat-id>) 
          <MN of target unit>
     ,TARGET-ACCESS=*BY-SOURCE
   

3. Use /DETACH-DEVICE to detach the target units (see the “Commands” manual [1])

Subsequently replication with data recovery can be resumed. 

Actions in the local Data Center

1. Prepare the local storage system:

   • Disable all the channels and remote links (to prevent access).

   • Reconnect all the channel or remote link cables.

   • Start the local storage system.

   • Service should check the local storage system for damage. This includes checking whether all tracks on the source units are readable and whether the local storage system is ready for synchronization.

2. Activate the remote links on the local storage system.
   The local and remote storage systems begin synchronization. Data is copied from the target units to the source units once the remote links have been enabled.

3. Activate the server channels on the local storage system.
   The server channels can be set to online once synchronization has started.

4. Start local system

5. Attach the source units on the local system and resume processing: /ATTACH-DEVICE (see manual “Commands” [1]).
   The reconstruction process can be checked by means of the following command

   /SHOW-SYMMETRIX-DEVICE-CONFIG
         UNIT=<MN of source unit(s) / MN of target unit(s)>,
         CONFIGURATION=*REMOTE-COPY-PAIRS,
         INFORMATION=*PHYSICAL
   

6. Import the pubsets and start the applications