We 're going to perform a DR test for the customer . The Production will be up during the test and customer would be bringing up the servers at the DR to test .
I am new to netapp so want to clarify the steps : at DR>
Break the Replication Mirrors
Create the igroup:
Once the luns are mapped, scan the ESX servers for the presented luns and proceed with VM recovery.
I am refering to an Old document that was prepared an year back and the document doesn't mention quiesceing . My question is whether its not necessary to quiesce the replication prior to breaking it in this scenario?
Personally I wouldn't do any of this as it disrupts your replication jobs. The way to do it with no disruption is clone the volumes on the DR destination. Attach ESX hosts to the clones and then you will be bringing up copies of the VMs in the DR environment. Biggest element by far is to be sure you have an isolated network for the DR test so there are no IP conflicts. We have done this many times with great success.
Thanks for the suggestion. However, we have 20 Luns that will be part of the test and the window is 24 ours ( includes storage , server, application) . Its a FC environment ,so wanted to know best approach as clonning would be time consuming in our case . Please advise
Cloning actually takes seconds per volume. We clone several hundred volumes in less than half an hour during each DR test. Why would it take a long time in your scenario? I do confess we don't use FC so perhaps that adds a wrinkle I'm unaware of.
Thanks again .I agree with your approach . But, I think customer wants to proceed with the old process as it was succefully performed last year .I will still try to bring it up in the discussion next week.
I need little clarification on the quesing part . I understand what it does but I want to know if there is any consequnces if I do not run quisece prior to breaking the snapmirror relationship.
If it's a qtree snapmirror relationship, I believe you have to quiesce first. If it's a standard snapmirror relationship, quiesce makes sure the job won't kick off again at some point in the future if this were just a temporary pausing of the replcation job.