June 25, 2014 at 8:51 am
I am going to be installing SP4 on SQL Server 2005 installed on a Windows 2003 R2 Active/Passive cluster (production). I do not have a 'test' cluster.
I have researched this a bit, and it seems that you just install the service pack on the active node.
Some of my searches have mentioned a server restart. I'm not sure if one is required on a clustered server, or if a failover or two is necessary.
Can you recall if your SP4 install in a cluster required an outage?
Thanks!
June 25, 2014 at 11:50 am
You Will have to install the SP on all the potential nodes that could run your SQL instance as the local executables needs to be upgraded.
June 25, 2014 at 12:31 pm
My understanding is that the service pack is installed on the active node - not 'each' node.
See Ken Davis post in this thread.
My question is whether a reboot is required when installing SQL Server 2005 SP4 in a cluster, or if a failover/failback does the trick.
June 25, 2014 at 2:22 pm
Sorry, but the SP needs to be installed on both nodes. I would suggest the following:
1, Locate the "Passive" node (Node 2) and install the SP on it.
2, Failover from node 1 to node 2 and install the SP on node 1
3, (Optional) Fail back to node 1 after installing the SP if you want node 1 to be the preferred node.
Good luck!
June 26, 2014 at 7:23 am
I appreciate that you have responded.
Can you please supply a reference to your assertion?
Thanks!
June 26, 2014 at 9:48 am
Hi again,
When trying to refresh my memory I Think you are right. You don't need to run the installation from both nodes, but on both nodes. As long as you run the SP installation on the Active node, the SP installation will first upgrade the passive node and then the ative node.If the passive upgrade will fail (this you can check by checking the sql server version when failing over) there is a switch you can run with the sp installation...eg spxxxx.exe /passive. This switch will just install the SP on the passive node.
A node restart is not required, but a SQL service restart is required. But as Always, I would prefer to do the node restart as you can failover the instance during the restart.
I hope this was an answer to your question and I am really sorry for the missunderstanding from my side....
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply