Creating data consistency at a remote site using Global Copy and FlashCopy functions

You can combine Global Copy and FlashCopy^® functions to create a consistent copy of your target volume at a remote site.

1. At Site A, establish consistency at the volume level:
   1. Create a volume pair using Global Copy mode. Consider the following guidelines:
      • When you create a Global Copy volume pair, it is not synchronized. The volume pair is in a pending state throughout the Global Copy processing. While in this mode, the target volumes are not in a consistent state and cannot be used for disaster recovery.
      • The volume pair will remain in the duplex pending condition, not reaching the full-duplex state until instructed to do so (i.e., by issuing a mkpprcpair command with the -mmir option).
   2. Run in Global Copy mode until you require a consistent copy of data.
   3. To establish consistency, synchronize the volume pair. This process copies pending tracks and potential updates to be transmitted synchronously. This process ensures that the target volume is current with the source volume, at a point in time.
   4. Wait for the volume pair to reach full-duplex state by viewing the volume properties. You can monitor this activity by continuously displaying the Metro Mirror relationship properties page on the DS Storage Manager or by setting up automation to monitor when the volume pair is in a full duplex state.
   5. At this point, quiesce host I/O to create a consistent point-in-time copy of all suspended volume pairs.
2. Freeze updates to source and target LSSs To have a consistent copy, you must either quiesce host I/O or issue freeze and run requests. Freeze operations are done on an LSS basis. It ensures that volumes on the target LSSs are consistent with the source LSS volumes. Updates made to the source volume LSSs after the freeze are not reflected on the target LSS volumes. You can issue freeze commands from the command line interface or by using automated software. (Automation software is not provided with the storage unit; it must be supplied by the user. However, IBM^® has offerings to assist with this automation. For more information, contact your IBM storage representative.)
   When I/O updates are frozen to the volumes across the source LSS, the following conditions occur:

   • The write operations to source and target volumes on the LSSs are temporarily queued.
   • The established paths between the LSS pairs are disabled.
3. Resume operations after the freeze This step also occurs through the use of automation software. This operation allows I/O activity to resume for the volume pairs across the LSSs. The pairs remain suspended but not offline.
4. At Site B, copy the target volume using FlashCopy: To create a copy of the volumes, perform the following steps:
   1. Vary the target volumes online.
   2. Use FlashCopy to copy the target volume to another volume.
5. At Site A, restart your applications:
   1. Resynchronize the source and target volume pair. This process copies any changed tracks to target volumes.
   2. If you issued the freeze and run requests, you must also reestablish the paths.