As discussed in the previous part of this series, in a production environment you should never shut down a cluster as it failed the entire purpose of High Availability. However, the current series is focusing on learning purpose. If you are running the setup in your test environment, where you don’t want to run all the VMs continuously, you need to cleanly shutdown the entire setup, before shutting down your PC. Some of the following methods are still applicable for update scenarios where you may want to update the cluster node / SQL Server with future Security Patches & Service Packs.
Make sure, all the applications are running from one node. If any application is currently running from the other node, do a failover. Connect to the cluster using Failover Cluster Administrator and the browse to the Nodes. Select the Node, where no applications are currently running, right-click on it and select Pause. This will pause the selected node, and a failover to that node will not be possible.
Make sure none of the resources are present on that node, which includes the Quorum of the cluster. Now log on to the node, and manually shutdown the node. Now all the resources are present on the first node.
Now one node of the cluster is down, it’s time to shut down the other node of the cluster. To start with, take the current applications offline. In our case, I first took the SQL Server installation offline followed by the MSDTC. Never take the Quorum Disk offline.
Starting the VMs:
While starting your cluster, you need to reverse the entire order of Shutdown the cluster.
- Start the Domain Controller of your Lab environment.
- Start the Storage Server, make sure both the DC & the Storage servers are up before proceeding further.
- Start the node of your Windows Cluster where the quorum & offline applications reside.
- Once the node is up and running, start the other node of the cluster.
- Open Failover Cluster Administrator and start the MSDTC followed by the SQLServer.
- Make sure the other node is up and running (verify whether all the services with start-up type automatic are started). Once the node is up, Resume the cluster node.
Your cluster is back online and available for you.
The complete series is only meant for training & learning purpose. While installing a SQL Cluster in a production environment, you need to make sure that your Hardware matches the Hardware Compatibility List for installing SQL Server Failover Cluster. The cluster validation wizard should pass all the tests and if some of the tests are failed, make sure they are passed before proceeding further. Please refer to Books Online for additional information on How to Setup a SQL Server 2008 R2 Failover Cluster?
Here is the complete list of all the posts on this series:
- Building a SQL Server 2008 R2 Cluster: Part I
- Building a SQL Server 2008 R2 Cluster: Part II
- Building a SQL Server 2008 R2 Cluster: Part III
- Building a SQL Server 2008 R2 Cluster: Part IV
- Building a SQL Server 2008 R2 Cluster: Part V
- Building a SQL Server 2008 R2 Cluster: Part VI
- Building a SQL Server 2008 R2 Cluster: Part VII