This article was previously published under Q300183
BUG #:236421 (SHILOH_BUGS)
If the SQL Server cluster resource is in an "online pending" state and the cluster service sends a termination request to the resource, the resource may not successfully come online again on the node until the cluster service is restarted.
The cluster service usually sends a termination request to a resource to perform a failover.
When the cluster resource DLL terminates a resource that is in the "online pending" state, it does not communicate this information to the cluster service correctly. This puts the cluster service in an unknown state and prevents this resource from coming back online.
To resolve this problem, obtain the latest service pack for SQL Server 2000. For additional information, click the following article number to view the article in theMicrosoft Knowledge Base:
290211 INF: How to Obtain the Latest SQL Server 2000 Service Pack
Restart the cluster service on the node.
Microsoft has confirmed that this is a problem in SQL Server 2000. This problem was first corrected in SQL Server 2000 Service Pack 1.
Steps to Reproduce Behavior
Configure the SQL Server resource so that it cannot connect successfully; for instance, configure it to use a port that is already in use.
Start the SQL Server resource so that it remains in an "online pending" state.
Perform a failover to another node.
Fail the SQL Server group back to the original node.
The cluster will fail repeatedly until the cluster service is restarted.