1. Stop Storage for Windows and High Availability (SFW-HA) on all nodes in the cluster, leaving the items that the configured resources manage up and running using the below command from a command prompt:
   hastop -all -force
2. On the first node that the SQL instance will be uninstalled from, launch the Veritas Enterprise Administrator (VEA) and perform the following actions, import the disk group that contains the SQL instance, and mount all of the volumes using the same drive letters that SFW-HA MountV resources are configured to use.
3. From Windows Explorer, make a copy of the SQL data directory and rename it. One copy / backup of this directory is needed for each node in the cluster that can host the SQL instance (e.g if there are 4 nodes in the cluster that can host the SQL instance, 4 copies / backups of the directory are required). This directory backup will be used to uninstall the SQL instance on the second node (this directory is required and the SQL instance on the second node will not uninstall if this is not done).
   For example, if the SQL data directory  name is F:\ProdSQLData, copy this to F:\1_ProdSQLData. If there are more than two nodes in the cluster, make another copy of the directory using the number three, "3", at the beginning; F:\3_ProdSQLData.
4. From Control Panel > Add or Remove Programs (Windows 2003 and below), or Control Panel > Programs and Features (Windows 2008), uninstall the SQL instance.
5. After the SQL instance has successfully been uninstalled, in the VEA, deport the disk group.
6. On the next node to uninstall the SQL instance, launch the VEA, import the disk group that contains the SQL instance, and mount all of the volumes using the same drive letters that SFW-HA MountV resources are configured to use.
7. Replace the SQL data directory with the copy / backup directory from step 3 (rename the existing SQL data directory to "OLD_" and rename [a] copy / backup directory to the SQL data directory.
   Following the example above, rename the directory F:\ProdSQLData to F:\OLD_ProdSQLData, and rename F:\1_ProdSQLData to F:\ProdSQLData.If there are more than two nodes in the cluster, the next directory to be used on the next node will be F:\3_ProdSQLData.
8. From Control Panel > Add or Remove Programs (Windows 2003 and below), or Control Panel > Programs and Features (Windows 2008), uninstall the SQL instance.
9. After the SQL instance has successfully been uninstalled, follow steps 5-8 on the remaining nodes, replacing the F:\ProdSQLData directory with the appropriate backup copies created in step 3 above.
10. Once the SQL instance has been removed from all nodes in the cluster, all of the F:\OLD* directories can successfully be deleted.

At this point, the SQL instance should be removed from both nodes and we are ready to delete the service group.

1. Start VCS on all nodes using the below command from a command prompt:
   hastart -all
2. Open the Cluster Manager - Java Console and delete the SQL Server service group:
   Right-click the service group > Delete > Answer Yes to delete all of the resources

At this point, the SQL instance has successfully been uninstalled from all nodes in the cluster and the resources to manage it have been deleted.