Article ID: 252974
Article Last Modified on 11/1/2006
APPLIES TO
- Microsoft Windows 2000 Advanced Server
- Microsoft Windows NT Server 4.0 Enterprise Edition
This article was previously published under Q252974
SYMPTOMS
When you shut down a cluster node that currently owns cluster resources, you may receive the following error message:
CAUSE
During shutdown, the Cluster service first stops the network heartbeat to the other node. This ensures that the surviving node eventually takes the resources and restarts the services. This avoids a situation in which a service simply hangs during shutdown but both nodes continue to exchange the heartbeat. In this situation, a failover would never occur.
If a service requires a long time to complete its shutdown, the other node may graft the disk resources while the service is still writing to the disk. This causes the error message.
This behavior has been observed with the MSDTC resource and Microsoft Exchange.
This is a general design issue that conflicts with the design goal of Cluster Server for a guaranteed failover time. The fix described below extends the allowed time for all resources from 9 seconds to 15 seconds. In addition, the Generic Service resource terminates the controlled service during shutdown if it does not exit within 3 seconds.
RESOLUTION
To resolve this problem, obtain the latest service pack for Windows 2000. For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
260910 How to Obtain the Latest Windows 2000 Service Pack
Windows 2000
For additional information about a fix for Windows 2000, click the article number below to view the article in the Microsoft Knowledge Base:
272569 System Process Loses Delayed-Write Data with Windows Clustering
Windows NT 4.0
To resolve this issue, apply the hotfix described in the following Microsoft Knowledge Base article:
251007 Some Cluster Disks Are Not Available After Installing SP6
This fix also includes new Microsoft Transaction Coordinator (MSDTC) components, and includes the modifications described in the following article in the Microsoft Knowledge Base:
256326 Cluster Server Cannot Use Disk Beyond Device Number 25
To work around this issue, before you initiate a shutdown, move all groups in Cluster Administrator to another node. Or, use a batch file in which you first stop the Cluster service (with the net stop clussvc command), and then use Shutdown.exe from the Resource Kit to shut down the node.
STATUS
Microsoft has confirmed that this is a problem in Windows NT 4.0. This problem was first corrected in Windows 2000 Service Pack 2.
MORE INFORMATION
The problem described in this article is indicated by the following entry in the Cluster log:
Additional query words: Event ID: 26
Keywords: kbhotfixserver kbqfe kbbug kbfix KB252974
