SQL AlwaysOn Configuration on Clustered Servers

  • I have 2 Windows 2016 servers. Each server is in its own data-center.
    Each server is set up as a cluster, with 3 nodes (node1, node2, node3), each node (1,2,3) hosts an instance of SQL Server 2016 Enterprise.
    Can I set up Always on, using data-center1-node1,2,3 (As primary replicas) to data-center2-node1,2,3 (As secondary replicas)?
    I wasn't sure if the WSFC would recognize the cluster node as an independent server.
    Also, each node in data-center1 is set to fail over to its respective node, in data-center2, in case of failure. Would this need to be disabled?
    Thanks in advance

  • So you have just two database servers in total, across two datacentres? How are they each part of a 3-node cluster, in that case?

  • Dig into the server setup for Always On.  You will create a cluster, just not a failover cluster, with the nodes you want.  In your case, it sounds like you would have 3 separate clusters, each with 3 nodes:

    CL1ND1, CL1ND2, CL1ND3
    CL2ND1, CL2ND2, CL2ND3
    CL3ND1, CL3ND2, CL3ND3

    We'll say node 1 is going to be primary in your Availability Group with nodes 2 and 3 in a secondary role.  Depending on location of the other data centers and how your bandwidth is between them, you will determine if you need them setup with synchronous or asynchronous replication.  The node which is asynchronous could be used as a DR node while a secondary in synchronous replication is a great candidate for a report server or just a read-only node handling that traffic.  The nodes will have a listener which you will specify for your connections to the cluster.  If you add ApplicationIntent=ReadOnly; then traffic will be routed to your secondary server.

    Again, dig into it a bit more and the picture will become clearer.  I had some fear going into it and it turned out to be not as scary as I made it out to be.

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply