0

I have the following setup:

One Windows cluster consisting of four cluster nodes: A, B, C, D. Cluster nodes A and B in one data centre, cluster nodes C and D in a second data centre.

SQL failover cluster instance SQLFCI1 defined over nodes A and B, SQL failover cluster instance SQLFCI2 defined over nodes C and D.

SQL availability group SQLAG1 defined on SQLFCI1 with asynchronous replica being SQLFCI2. SQLAG1 associated with listener SQLListener1.

All client apps use SQLListener1 in their application connection strings.

End result

High availability because of FCI between A and B in data centre 1 as well as disaster recovery to FCI between C and D via async availability group. That DR site has high availability as well because the availability group sits on an FCI as well.

Minimal impact to applications because they all use the AG listener name that in effect can move across all four nodes.

Quorum partner in a third data centre.

Is this a valid setup?

3
  • 1
    Please add the output of select @@version in the question
    – Shanky
    Apr 10 '18 at 11:50
  • BTW what you are trying is fine and can be configured but I was thinking more in terms on distributed AG. Please also note you cannot have automatic failover for AG hosted on FCI
    – Shanky
    Apr 10 '18 at 12:03
  • The biggest risk you have with that is the single cluster. I have seen corruption bringing down all cluster services.
    – eckes
    Apr 11 '18 at 1:15
1

Is this a valid setup ?

Yes. This configuration is detailed here:

AlwaysOn Architecture Guide: Building a High Availability and Disaster Recovery Solution by Using Failover Cluster Instances and Availability Groups

If you want automated failover between data centers, look at using a Cloud Witness for cluster quorum.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.