Skip to main content
Pure Technical Services

ActiveCluster FlashArray Relocation Procedure

Currently viewing public documentation. Please login to access the full scope of documentation.

When physically moving one array in an ActiveCluster environment to another rack or to another Datacenter, perform the following guidelines and necessary steps to disconnect, move, and then re-connect the FlashArray once relocated.



When physically moving one array in an ActiveCluster environment to another rack or to another Datacenter, perform the following guidelines and necessary steps to disconnect, move, and then re-connect the FlashArray once relocated.


Pro Services MUST be engaged if a customer would like Pure to perform this work. Support MUST NOT perform these steps for the customer. 

Note: This information CAN be shared with the customer every command show is admin level and none of the commands require support or PS to be engaged.  

This use case has two FlashArrays that are named ArrayToStay and ArrayToMove. These FlashArrays are currently connected and form an ActiveCluster at a single site or at two Datacenters. Begin by relocating ArrayToMove to a new rack in the same site or to a new Datacenter.

Also, you can use this process to change the replication IP addresses without a physical relocation. 

  1. List the ActiveCluster host connections for all pod volumes on ArrayToMove. This CLI output can be redirected ( i.e. > ~/connections) and used later to reconnect hosts when ArrayToMove is added back to ActiveCluster.
    user@ArrayToMove> purevol list --connect --cli                     
    purevol connect --host host1 --lun 1 Pod1::vol1
    purevol connect --host host2 --lun 1 Pod1::vol1
    purevol connect --host host2 --lun 2 Pod2::vol2
    purevol connect --host host1 --lun 2 Pod2::vol2
  2. On ArrayToMove, disconnect host connections from all pod volumes.
    user@ArrayToMove> purevol disconnect --host host2 Pod1::vol1
    user@ArrayToMove> purevol disconnect --host host2 Pod2::vol2​​​

    Note: If the plan is to disconnect ArrayToMove for less than 24 hours, you may skip step 3. If there is any chance at all of ArrayToMove being disconnected for 24 hours or more, it is strongly recommended to perform step 3 to persist the pods' re-stretch data on the ArrayToMove for longer periods of time. Keeping the data will speed up re-synchronization.

  3. On ArrayToMove, create a clone of pods.
    user@ArrayToMove> purepod clone SOURCE TARGET
  4. On ArrayToStay, unstretch pods to ArrayToStay by disconnecting ArrayToMove from each stretched pod.
    user@ArrayToStay> purepod remove --array ArrayToMove PODname
  5. On ArrayToStay, disconnect synchronous replication to ArrayToMove.
    user@ArrayToStay> purearray disconnect ArrayToMove

You can power off ArrayToMove only after confirming any ArrayToMove non-pod volumes, host applications, and hosts are prepared for ArrayToMove to be powered off. Once confirmed, ArrayToMove can be powered off and disconnected from all network and power cabling for moving to a new rack or for shipping to the new Datacenter. 

Post Relocation Steps

Once ArrayToMove is cabled and powered up in its new location, you can re-IP the array's management/replication as required. You can now reconnect ArrayToMove to ArrayToStay to reform an ActiveCluster configuration.

  1. On ArrayToStay, connect sync replication to ArrayToMove.
    user@ArrayToStay> purearray connect --management-address MANAGEMENT_ADDRESS [--replication-address REPLICATION_ADDRESS] --type sync-replication --connection-key
  2. On ArrayToStay, re-stretch PODs to ArrayToMove.
    user@ArrayToStay> purepod add --array ArrayToMove PODname
  3. Verify that pods are synced on ArrayToMove.
    user@ArrayToMove> purepod list
    Name  Source  Array            Status  Frozen At
    Pod1  -       ArrayToStay      online  -        
                  ArrayToMove      online  -        
    Pod2  -       ArrayToStay      online  -        
                  ArrayToMove      online  -   
  4. Connect hosts to pods on ArrayToMove. Here you can manually reconnect them using the CLI or use the CLI commands that were generated in Step 1 at the beginning of this procedure.
    user@ArrayToMove> purevol list --connect 
    Pod1::vol1  444G  1    -           host1
    Pod1::vol1  444G  1    -           host2
    Pod2::vol2  777G  2    -           host1
    Pod2::vol2  777G  2    -           host2
    user@ArrayToMove> purevol --connect host1 --lun # Podname::volname

Post Move Cleanup

Once ActiveCluster is running and you've completed their test and acceptance checks, remove any (optionally) created clone pods on ArrayToMove as shown below. 

Note: Steps 1 - 4 (below) are only necessary if you performed Step 3 (Create Clone of Pods) during the Pre-Move steps (at the beginning of this procedure). 

  1. On ArrayToMove, destroy volumes from cloned pods.
    user@ArrayToMove> purevol destroy ClonedVOL
  2. On ArrayToMove, destroy cloned pods.|
    user@ArrayToMove> purepod destroy ClonedPOD
  3. On ArrayToMove, eradicate volumes from cloned pods.
    user@ArrayToMove> purevol eradicate ClonedVOL
  4. On ArrayToMove, eradicate cloned pods.
    user@ArrayToMove> purepod eradicate ClonedPOD