My three node SQL Server 2008 R2 physical server cluster needs to be shutdown, moved along with its storage to a new location, and restarted once in the new location. My desire is to have a clean, controlled shutdown/startup sequence and, if possible, for all SQL Server instances to remain on their original node during this process. So, if SQLInstA is on NodeA and SQLInstB is on NodeB and NodeC is empty, there is no failover during server shutdown and I’ll bring up the services individually upon startup.
I’ve read various recommendations regarding setting SQL Server offline, pausing nodes, stopping cluster services, setup startup types to manual.
My first thought was to set the SQLInstA and SQLInstB services offline in Failover Cluster Manager. Offline does not trigger a failover so they’ll stay put. But, rebooting will cause failover, regardless of my setting them to have only one preferred owner. The service doesn’t restart, even though I didn’t “disable auto start”, so even if both instances ended up on the same node, the services would be down on startup, which isn’t horrible. Just manual failover work afterward.
I could stop the cluster services, but that triggers failover as well. Is it important to stop cluster services in this case?
Any thoughts on how to accomplish this?
Best Answer
Shutting down the cluster for a move is pretty straight forward.
To bring things back online.
A word of advice, make sure that EVERY cable is properly labeled at both ends before you unplug anything. I've done this kind of move a few times. A couple of hours of labeling the cables will save you hours of pain when getting things back up and running.