Post reply

Fallback scenario installing SP1 on SQL 2014 cluster

  • January 13, 2016 at 7:42 am

    #314869

    Lectori Salutem,

    I have to install SP1 for SQL server 2014 on a clustered environment with 2 nodes. Actual it is not that difficult to install the SP (download the SP, move the active node to the other node, install the SP on the node, move back from node 2 to 1, install SP on node 2 – done)….. but what if somethings goes wrong ??

    Assume that during installing the SP on node 1 something is going wrong - which preparations do I have to make before starting installing the SP?

    We are running on VM so I can make a VM-Snapshot from node 1 (and/or node 2) just before I start but what happens if I have to restore the VM-Snapshot? Is the data from the snapshot still in sync with the data from node 2 ???

    What do I have to do with the QUORUM disk which resides on one of the nodes (while restoring a VM-Snapshot) ??

    What will happen if I have to reinstall SQL server on one of the nodes when the installed version gets corrupted – and what are the recommendation reinstalling or rebuilding it?

    In other words – what is a recommendable FALLBACK scenario installing a SP on a 2-node SQL cluster?

    Please help me out with a (detailed) plan on how to accomplish this.

    Any help is appreciated.

    Guus Kramer

    The Netherlands

  • January 13, 2016 at 7:16 pm

    #1851791

    I would first install the sp on the secondary node which is not active. Then fail over the cluster to the updated node and wait for few days before applying the sp on the previous active and current passive node. This way you can avoid any issues or failback to the previous active node if any issues on the updated one.

    If you operating system is windows 2012, it has dynamic quorum. Even if we loose two votes, in this case node2 and disk, the cluster should be up and running on last surviving node.

  • January 14, 2016 at 1:00 am

    #1851823

    LS,

    After some digging (using some different terms) on google I found this artical;

    http://blogs.msdn.com/b/john_daskalakis/archive/2014/09/05/how-to-install-a-service-pack-at-a-sql-server-2012-failover-instance-best-practices.aspx

    PART 4 is explaning what to do in case of a failure.

    The hint on waiting to upgrade the seconde node after a while is part of the ACC update to test if it will work as excpected.

    Once upgrading the PRD environment the update should be done in a maintenance time window which can not stay open for several days/weeks.

    The reason why I posted this topic was - What if something happens unexpected !!

    Guus

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

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