Mirrorview test

It depends on the policy settings you have on your remote mirrors.

Are you using the write intent logs?

Is the recover policy set to Automatic?

if so just remove the FC cable that you use for the MirroView link.

The remote mirrors will became sys fractured and start using the write intent logs for any changes that are written  to production. Once the link comes back the remote mirrors will automatically start to transistion and sync back up  

Hi,

We are using WIL and recovery policy is set to manual.There are consistency groups created to for MV.

So after test is complete and we re-connect the cables what steps do we need to do? Also would it start replicating to same luns or do we create new luns on DR array?

regards,

Samir

consistency groups have the same recovery option Manual and  Automatic.

You mention that you have your mirrors set to Manual - therefore you will have to re-sync the mirrors once the link comes back on line just right click on the primary Lun within the remote mirrors and click synchronise. Each remote mirror will start to sync up - you do not have to re configure anything on the DR side in terms of replica Luns

are you wanting to promote and mount the replica Luns? ie make them accessible to the DR hosts, what flare code are you currently on?

From your first post I had assumed you were only testing the link failure process. From your latest post you mentin mounting the replica disks at DR, this is a completely different question if indeed this is the question your asking?

Hi,

The part of this DR Drill is to check the consistency of replicated volume, thus we will plug the fc cables used for replication between the boxes. Mount and check DR luns to DR server and check the data.

Primary CX3-20 - Rel 26 Patch 7

DR CX3-20 - Rel 26 patch 31

regards,

samir

if you just want to see if the data is consistent you can snapshot the DR Lun and mount it on a host and view the data.

if you want to do a full DR test and just pull the FC cables then most likely for Remote mirror will be in a consistent state, if so you would have to do a full resync back to production and depending on your Lun size and replication link size this could take quite some time.

if the lun was in a sychronised state when you promoted then only a partial sync back would be required.

Thanks for all the replies.

We shall do the test and post the results.

regards,

Samir

The flare code on the source side is a bit low, you should consider updating to the latest patch .031 - there were some updates to mirrorview that you should have.

On the test for mirrorview, be sure to review the Knowledgebook for Mirrorview - there is a section about testing the failover for mirrors.

White Paper - EMC MirrorView Knowledgebook - FLARE 30—A Detailed Review.pdf

glen

Thanks for all the suggestions.