We have SQL Server 2012 in a two-node cluster with Windows Server 2012 (node and disk majority). Neither of the nodes are marked as preferred owner and the "Prevent Failback" option is selected. Here is the problem we are having:
- Node A current host, Node B inactive node.
- Node B is updated and the process fails the cluster over from A to B
- Node A is updated
- After node A is back on line, the process fails over the SQL Server cluster back to Node A. However, the Windows Cluster is still on Node B.
We have prevented this by setting the inactive Node to preferred owner before the updating process begins, but that entails manually setting it every time we want to update.
Does anyone know how to prevent the failback to the pre-update active node?