Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It only takes a minute to sign up.
Sign up to join this community
I have an active passive SQL Cluster. When I stop the SQL services on active node from SQL Server Configuration Manager, the SQL cluster doesn't failover instead shows as Partially online.
If SQL services are stopped from Configuration Manager, it will not failover to another node?
I have to prepare test cases for my client, so it is really urgent for me to clear this doubt.
When configured as a Cluster in Windows, you need to control the SQL Server Services from the Failover Manager, not SQL Server Configuration Manager. In Windows Failover Manager, you can fail the resources/services to a particular node and then carry on with your needed activity.
The database engine and SQL Agent are cluster aware by default. Other SQL Server services (SSRS, SSAS) can be added as cluster aware resources, but their default behavior is not cluster aware. As far as I know, SSIS is still the only service that cannot be made cluster aware (i.e. it has to be installed on all nodes and started/stopped with SSCM on each node).
We accidently stopped the SQL Agent from the Configuration Manager, and it showed Partially Online as you mentioned. For the cluster to get happy again, we had to offline the Agent from Cluster Manager, and then bring it back online from the same place. The point to be made: always control failover or instant restarts or offlines from the FCM.
SQL_Hacker's answer is exactly right. Use the Failover Cluster Manager which is designed for managing clusters.
If you stop the service from the Configuration Manager, the instance may not failover if you do not allow the cluster to perform automatic Failovers, set during the initialization of your resource/services.
If you suspect this setting is set properly but the cluster does not failover as expected, run the validation tests (deselect the disk tests) to make sure your cluster is healthy.
