SRM on EMC Clariions Replicated LUNs

Question

Hello,

I have customer EMC Clarrions Storage and vCenter SRM with MirrorView/ A.

After completion of the EMC Clarrions Storage setup and SRM Configuration, We have created real test environment for the failover and failback to check if all our configuration are working as expected. The real test failover and failback was successful and we did achieved what is expected to happen during failover and failback. But during the exercise we have notice/ observe something on storage status which raise few questions, below are our observation;

1. When initiating failover, the storage replication was suspended which is expected but shows "F" (Fractured), even though we have kept the LUN in the "Consistency Group". Is this normal?

2. We ignored the status and proceed to prepare for the Failback, since the status of the mirrored LUN is Fractured and need to do a reverse synchronization, we set the Fracture mirror to synchronized in reverse direction (recovery site to protected site) but supprisingly it did a "FULL Synchronization" of the mirrored LUN. Again the same question, Is this normal?
If the mirrored LUN will do a FULL Synchronization, then it will be a big problem for a larger LUN of 2TB with MirrorView / A. It will take time to for the replication to complete.

3. Is SRM or SRA doing the work to "Force Promote" the replicated LUN in the recovery site?
4. What are the steps SRM is initiating during the Failover to recovery site?
5. What is the exact role of the SRA when SRM is initiating the Failover to recovery sites?

Thanks in advance for the reply.

BobbyRG.

GearoidG · Answer

I think I can answer pt 2.

It is an known with SRM

In the release notes of the MV SRA http://powerlink.emc.com/km/live1/en_US/Offering_Technical/Technical_Documentation/300-007-019.pdf

Refer to the known problems

You will see one on P6

Symptom details

VMware SRM

While running planned failover scenarios, MirrorView mirrors may go “out-of-sync (waiting- on-admin state)” after failover.

Problem Description

While executing the planned failover (recovery plan) the I/O to the MirrorView source images is not quiescied by SRM or ESX servers resulting in “out-of-sync (waiting-on-admin state)” mirrors after failover.

I think this is what you are seeing

There is no easy way to avoid this issue unless VMware addresses the underlying issue i.e. MirrorView source images is not quiescied by SRM or ESX servers

From what I understand from other cases with VMWare this may be in future release of ESX/SRM but they have not confirmed so

That release notes document is pretty good for explaining some bit your questions 3,4 and 5

Hope this helps

Gearoid

BobbyRG · Answer

Hi GearoidG,

Thank you for the reply.

On the same document link, it did mention on p13 (If the mirrors are “out-of-sync” or “waiting-on-admin” state you need to synchronize the mirrors before attempting a failback.)

So the above statement confirm that a FULL Synchronization will take place when a fractured replicated LUNs are being synchronized again.

Please advice if understand it correctly from above statement.

BobbyRG

VMware

SRM on EMC Clariions Replicated LUNs

Was this post helpful?