Active-Active cluster + AlwaysOn HA

  • Hi all, we currently have an active/active cluster and want to AlwaysOn HA in another data center.

    There will be 1 or more Availability groups per instance (we have a few hundred DB's) so heard that 100 or less DB's per AG is practical, even though there is no technical limit.

    So each AG will be: Primary Availability Replica of N DB's on Instance1 or Instance2, and we'd have a secondary replica, as most likely another Active/Active cluster, InstanceX and InstanceY.

    The goal is to have AG1 (on Instance1) replicate to InstanceX, AG2 (on Instance1) replicate to Instance X, AG3 (on Instance2) replicate to InstanceY, and AG4 (on instance2) replicate to InstanceY.

    What I'd like to know is:

    1. Does this layout make sense? and

    2. Would Instance1 and 2, the main working DB's, do regular failovers to each other, without causing an automatic failover to the secondary locations

    Gaby________________________________________________________________"In theory, theory and practice are the same. In practice, they are not." - Albert Einstein

  • spanning an availability group across 2 separate clusters is only supported for migration purposes.

    All servers need to be part of the same Windows Server Failover Cluster

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • gotcha...so, to use existing hardware, I'd need to break the cluster. Would the best recommendation be to create two new instances, one on each server, migrate from each active instance (via log shipping for example, or schedule an outage), bring online, and point our applications to the new server? As we are already working with two separate instances, this breaks nothing.

    Then we'd remove each node from the cluster, but keep the cluster in place for ALWAYS on later. Is that the rough sequence?

    Gaby________________________________________________________________"In theory, theory and practice are the same. In practice, they are not." - Albert Einstein

  • GabyYYZ (11/30/2015)


    gotcha...so, to use existing hardware, I'd need to break the cluster.

    No, you would need to expand it.

    GabyYYZ (11/30/2015)


    Would the best recommendation be to create two new instances, one on each server, migrate from each active instance (via log shipping for example, or schedule an outage), bring online, and point our applications to the new server? As we are already working with two separate instances, this breaks nothing.

    Then we'd remove each node from the cluster, but keep the cluster in place for ALWAYS on later. Is that the rough sequence?

    expand the cluster to include the new nodes, install the required new instances and create the availability group.

    Read my stairway to HA on this site starting at this link

    http://www.sqlservercentral.com/Authors/Articles/Perry__Whittle/354339/[/url]

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • Oops, I mean KEEP the cluster, but evict the nodes at the end after I've copied over to the new instances. Will definitely take a look at that. Thanks!

    Gaby________________________________________________________________"In theory, theory and practice are the same. In practice, they are not." - Albert Einstein

Viewing 5 posts - 1 through 4 (of 4 total)

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