With SRM you can test the recovery plan to make sure it will work should there be a real disaster. As SRM is a very clever product, it can allow you to test the SRM without actually failing over the VMs from Site A (Protected) to Site B (Recovery). Well, actually there are 2 types of Recovery Plan tests:
Test Recovery Plan
This is the “Green” play button under monitor tab of the Recovery Plan:
What happens when the green button is pressed
- This will start the Recovery Plan Test. But I called this as a Mock Test and not actual Disaster Recovery test. The reason I say this as this will test the VMs at the Protected site will still be working and SRM will bring up the VMs at the Recovery Site without impacting your Protected VMs or your actual production. There is a catch to this à You need to have a mechanism in place that will facilitate this functionality. I guess you would have guessed by now, yes, I’m talking about Storage Mechanism. But it depends on the type of Storage Vendor that you are using with SRM. Basically, it depends on the SRA configuration. Below is the test scenario based on Hitachi Data Systems VSP G800 storage system.
S-Vol based testing
Just a recap on HDS based replication. In HDS world, the Primary Volume is known as P-VOL, volume that stores the Protected VMs & the replicated copy this primary volume at the Recovery site is called as S-VOL or secondary volume.
In this testing, recovery site S-VOL is used for test failover. But there are the below restrictions:
- This method will disrupt the replication from the primary to the secondary volume;
- The S-VOL will not be available for an actual failover should the need arise;
- After testing, the pair is resynchronized with data that was stored in a bitmap. The updates are out of order, rendering the S-VOL unavailable for an actual failover should the need arise, until resynchronization is completed.
- The HUR or Hitachi Universal Replicator pair must be split in order to test using S-VOL.
- Below is the figure for the same:
S-VOL Copy (Shadow Image) based testing
This method will test the failover but without any disruption to the replication between primary & secondary Volumes. During test failover, the remote replication pair remains in PAIR status, and therefore protection continues uninterrupted. Below is the figure for the same:
Test Recovery Plan
As soon as the green button is pressed, test confirmation window will come up. Press Next and if everything is configured correctly then you will see the following results:
Run Recovery Plan
This is the “Red” button under monitor tab of the Recovery Plan:
What happens when the Red button is pressed
When the red button is pressed, this will give you option to either perform planned migration or test actual Disaster Recovery. This time I mentioned actual DR, this is because the protected VMs will be shut down at the protected site and will be powered on at the recovery site. The following window pops up when red button is pressed:
Here you select Disaster Recovery if you need to test the recovery plan in reality. If storage is configured correctly, then it work like a charm. Now note one thing for this you don’t need to have Shadow Image or snapshot of the volume in place, i.e. this method will use the real volumes attached to the hosts. As soon as the VMs are failed over the datastore will be detached from the hosts at the primary site and same happens at the recovery site when failing back. Proper DR test should be done in the following sequence:
Thank you & Enjoy Failing Over