Hello All.........I have few questions with respect to DR testing:
Setup:
There are two-2 locations (A and B).
Location A:
Server 1: MBX and CAS
Server 2: MBX and CAS
Server 3: MBX
Location B:
Server 1: MBX and CAS
Server 2: MBX and CAS
Server 3: MBX
DAG:
DAG1 consists of Server 1 and 2 from Location A and Server 3 of Location B. Server 1 and 2 (Location A) have active copies while passive copies are on Server 3 (Location B)
DAG2 consists of Server 1 and 2 from Location B and Server 3 of Location A. Server 1 and 2 (Location B) have active copies while passive copies are on Server 3 (Location A)
GOAL:
Goal is to mimic failure at Location B and activate the passive copies ofDAG2 at Location A. Not only that after activation, we plan to add another MBX Server (Server 4 at Location A) and create a DAG betweenServer 3 and Server 4 as we plan to keep this in place for few weeks before failing back to cover for the load.
Questions:
1. Is this testing scenario doable?
2. If so, can you share high level steps that would be required to carry out the testing?
3. How can this testing scenario be tested better if the goal is a bit vague?
4. Would adding Server 4 at Location A make any difference to the load at Server 3? If not, how we add the new server so that it is able to reduce the workload off of Server 3 as it would already be handling users from Server 1 and Server 2 from Location B?
5. I would appreciate comments on the design as well as to whether is it a good design? If not, how can it be made better?
6. When we have tested the failover, what steps would be needed to failback to Location B?
7. How long can we keep members of DAG disconnected from each other and then connect again? Is there any limit?
8. I would also appreciate, if you can share any blogs/articles for better understanding of DR, Failover and Failback as well.
Thanks in advance.