Going from Failover Cluster Instance to Availability Group with Azure Replica

  • jcourtjrSQL

    SSC Eights!

    Points: 842

    So if we have a FCI on premises, and want to add Availability Group replica in Azure. Would I first need to have an Availability Group on Premises (fully set up with non shared drives) and then add the Azure replica or can I just turn on the Availability Group Feature on and when setting up the first secondary replica, make that in the Azure VM? Essentially 2 replicas (On-Premesis FCI and Azure VM) or 3 replicas (On-Premesis Availbility group and Azure VM.)

     

  • Site Owners

    SSC Guru

    Points: 80378

    Thanks for posting your issue and hopefully someone will answer soon.

    This is an automated bump to increase visibility of your question.

  • jcourtjrSQL

    SSC Eights!

    Points: 842

    Bump

  • This was removed by the editor as SPAM

  • jcourtjrSQL

    SSC Eights!

    Points: 842

    This looks like garbage. Anyone else want to take a crack at this?

  • anthony.green

    SSC Guru

    Points: 112358

    All depends what you need from the solution.

    You can add an AG over the top of a FCI, but there are caveats, the biggest one for some is that data replication has to be async.  Now given that your syncing On-Prem to Azure, I am going to guess that's not a concern due to the transactional delays sync vs async would give you when traversing the cloud network and if you have an ExpressRoute etc.

     

    https://www.sqlserverlogexplorer.com/adding-alwayson-availability-group-existing-failover-cluster/

     

    After that it's configuring an AG as normal, then you will need to change connection strings to point to the listener etc, or leave them as the cluster name and change when you invoke the switch to Azure, whichever your business requires.

  • jcourtjrSQL

    SSC Eights!

    Points: 842

    So in other words I would only end up with 2 replicas in the end correct? The FCI on premesis and the Azure VM? (#Methodology 2 from the link) Async would be fine.

  • anthony.green

    SSC Guru

    Points: 112358

    Yeah essentially.

    The two nodes in the FCI will act like normal, providing local HA. (1 instance).  Then the Azure VM would provide DR (1 instance)

    Never done it myself, only ever read about it so don't have any gotcha's.

    But from what I seen there is a few resources out on the net for implementing this.

  • anthony.green

    SSC Guru

    Points: 112358

    As I think about it, quorum might be an issue, but as it's async anyway and has a potential for data loss, then a forced manual failover should the FCI fail would be the way to go, unless you gave the Azure VM 2 votes and used Azure Storage as a 5th vote so you had a majority should either the FCI go down/into maintenance or the Azure VM does.

Viewing 9 posts - 1 through 9 (of 9 total)

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