Let's says we have 2 nodes (N1, N2) clustered (say C1) on-site. And an Always On AG is configured between this cluster C1 and a new SQL server instance N3 (N3 is off-site).
Now, if there is a problem (failure) with N1 then N2 will become active (or vice versa) and continue to serve requests. This is due to clustering concept.
However, if there is a failure of the cluster, then as per SQL 2012 documentation, there cannot be an automatic failover as this is not supported between clustered node and separate server instance. DBA will have to manually do the failover.
Please can you confirm the following:
- Is this feature (automatic failover with clustered and separate server instance) supported in newer SQL versions (2016,2017)?
So as to be able to use the automatic failover feature we have only 3 options:
a) Cluster all the server instances into 1 cluster and setup a disk backup strategy (since single shared disk is used in clustering concept)
b) Don't cluster at all and only setup AG between SQL server instances.
c) Cluster all server instances into 1 cluster and setup AG between them - however bit confused in practicality of this scenario - What is the purpose of having AG within cluster?
Any guidance and reference links will be very helpful.