Primary Storage

  • 1.  Can someone explain the "Change Volume" feature?

    Posted Fri April 01, 2022 04:59 AM
    Hi,

    We recently upgraded our V5030F to a V5035F that runs newer software. Now when creating remote copy relationships, we are asked if we want to add a Change Volume. But the documentation is close to non-existent about this option.

    Can anyone tell me the pros and cons of using a Change Volume? So far we just created without them, as that seemed to work fine on our old SAN.

    /Rasmus

    ------------------------------
    Rasmus Teglgaard
    ------------------------------

    #Storage
    #StorageAreaNetworks
    #PrimaryStorage


  • 2.  RE: Can someone explain the "Change Volume" feature?
    Best Answer

    Posted Fri April 01, 2022 06:23 AM
    Hi Rasmus,

    Change Volumes for Metro Mirror (MM) and non-cycled Global Mirror (GM) have the same function as for Global Mirror with Change Volume (GMCV), which is an asynchronous cycle-based replication. I think it is quite well documented when you search for GMCV.
    Change Volumes are optional for MM and GM, but mandatory for GMCV.
    Change Volumes are FlashCopy snapshots that are targets of the primary and secondary replicated volumes (so a FlashCopy on both sides of the replication relationship)
    The pros:
    - A point-in-time copy of your primary or secondary side allows you to have always a consistent copy of your volume. During a synchronization phase, if a disaster occurs, the data that is being copied may not arrive at its destination completely and you may end up with inconsistent (corrupted) data on the target side. If you have a FlashCopy defined (a change volume), you can go back to an older but consistent version of your data. Of course, you can still have your source data that is valid, but if it's not accessible (either destroyed or the link between your valid source and the target system is unusable), then you can restore a previous version of your data on the target system. This could reduce the RTO of the disaster recovery process.
    The cons:
    - As for any FlashCopy, you need to plan the potential capacity usage of the additional copies. Normally, it is quite low because you only copy the data that has changed since last completed copy. But if the link between the two systems is interrupted, depending on the interruption duration, the data that has changed since last completed copy will increase (also depends on the write data rate on your volumes).
    - Also, there is a limit for the total number of FlashCopy mappings defined on a system, so if you already have FlashCopy mappings defined for other purposes, you might hit a limit.
    - The same way, there is a limit in the number of volumes on a system. For one servers LUN, you add a replica and potentially 2 change volumes if you want to use them. So for 1 servers' LUN , you end up with 4 volumes.
    - And last, Change Volumes and there to track the changes since last copy. It means that when you perform a write operation on the source volume, because of the Copy on Write mechanism, you will first perform a write on the FlashCopy target. So, for a write operation made on a volume, you have 2 write operations. It might have a performance impact on your system.

    Using or not Change Volumes then really depends on your objectives and on your existing systems capabilities (capacity, performance and limits).

    I hope this helps. Maybe someone can add details or precisions.

    ------------------------------
    Erwann Auffret
    ------------------------------



  • 3.  RE: Can someone explain the "Change Volume" feature?

    Posted Fri April 01, 2022 06:30 AM
    Thanks for the thorough answer!

    /Rasmus

    ------------------------------
    Rasmus Teglgaard
    ------------------------------



  • 4.  RE: Can someone explain the "Change Volume" feature?

    Posted Tue April 05, 2022 09:29 AM

    Hi.. 

    Just want to add one thing..

    If you have a link failure and your replication stops between your systems.

    When this issue has been fixed and the two systems can communicate you need to manually restart replication but if you have change volumes replication will automatically restart.

    /Tony Eriksson



    ------------------------------
    Tony Eriksson
    ------------------------------