- We are using Always on Groups for DR and Cluster as HA.
- The cluster is spread across 3 global regions (geo cluster with remote witness) and 3 data centers with 4 physical nodes across each DC.
- We have 2 cluster nodes in Active active scenario with AlwaysOn to diff data center nodes and use them as DR with active Readonly.
- EG: DC-A has 2 nodes and DC-B has 2 nodes.
- DC-A has always on with read-write Groups to Active secondary replicas to DC-B
- DC-A Node 1 DR's to DC-B Node 1 and DC-A Node 2 DR's to DC-B node 2.
- We want to use DC-B for upgrading the Application while breaking the Always on.
- Technically we will be operating live on Dc-A where the Application team is upgrading DC-B to new versions and new schemas.
- After the upgrade is Done, we want a minimal way to restart the Always On.. so the flow will reverse in this case where DC-B is active and DC-B is secondary replicas.
- Is this possible? If not then what are the possibilities to this approach.
↧
Reverse AlwaysOn Failover Flow of Data.
↧