Cluster Upgrade -- some suggestions please

  • We currently have a three node cluster running on Windows Server 2008 SP1 (not R2) and two of the Nodes are Active with four MS SQL 2008 SP1 Instances on each and the third node is passive for all 8 instances. Below I'll call the Active Nodes A1 and A2 and the Passive node P1 just to keep it simple 🙂

    I need to install several updates plus MS SQL 2012 as a new Instance on the cluster, but given this requires many updates to accomplish I wanted to bounce our suggested plan to the group in-case others have experienced problems going this route or with any of these updates.

    Here are the updates and applications we plan on rolling out:

    - KB948465 – Windows Server 2008 Service Pack 2

    - KB956250 – Microsoft DotNet 3.5 Service Pack 1 Update

    - KB2546951 – Microsoft SQL Server 2008 SP3 on all Instances

    - KB968930 – Windows Management Framework Core Package (WinRM 2.0 and PowerShell 2.0)

    - Microsoft .NET Framework 4.0 – Required for future software updates

    - Microsoft OLEDB Provider for DB2 v4.0

    My plan is to install Windows Service Pack 2 on the Passive node (P1), then after reboot move all Instances from A1 to P1 then back to A1, then move all instances from A2 to P1 and back just to verify everything moves across okay. If no hiccups then we'll move all instance from A1 to P1, install Windows SP2 to A1, then move the instances back and do the same to A2.

    Next we'll install KB956250 then KB2546951 on the Passive node (P1) and move the four instances on A1 and A2 over to P1, one node at a time, and if everything moves without issues move the Instances on each node one at a time to P1 and update each node as before..

    After all this I'll install KB968930, DotNet 4.0, and OLEDB Provider for DB2 to passive node then once again move everything over and back one node at a time to test, and if all works install on Active nodes as before.

    Then after all this is done I should be ready to install SQL 2012.. Sheesh! It's a lot, but I'm not sure how else to get everything caught-up and tested in between updates so if something fails I can pinpoint the cause.

    Any other suggestions or ideas for this? Or has anyone ran into caveats with any of these updates on a Failover Cluster?

    Thanks,

    Sam

Viewing 0 posts

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