Article ID: 823955
Article Last Modified on 1/31/2007
APPLIES TO
- Microsoft Windows 2000 Datacenter Server
- Microsoft Windows 2000 Advanced Server
SYMPTOMS
After you complete the following steps, all groups you move to cluster node 1 are moved back to the active cluster node 2.
Scenario
Node1 has the groups and is paused. The failback policy says Node1 is Preferred Owner. The resources will stay on Node1 and will stay online. Node2 is restarted. After Node2 starts, the Cluster service sees that there is now an Active non-paused node so it moves the groups over.
- You move some groups to cluster node 1.
- You pause cluster node 1 to maintain these groups.
- You restart active cluster node 2.
All the groups are moved to the active cluster node even though there is no failback policy. All the groups are moved to the active cluster node 2 even though there is no failback policy defined.
CAUSE
This behavior occurs because when the active cluster node starts, the Cluster service enumerates all the groups and resources. If the Cluster service finds a failback policy, the group or groups are moved depending on the failback policy settings. However, if cluster node 1 is paused and cluster node 2 is not paused, all the groups are moved to cluster node2. Pause state takes precedence over a failback policy. The node is actively participating in cluster operations but cannot take ownership of resource groups and cannot bring resources online
WORKAROUND
To work around this behavior, resume the paused cluster node before you restart the active cluster node. To do this, follow these steps:
- Click Start, point to Programs, point to Administrative Tools, and then click Cluster Administrator.
- In the left pane, click the paused cluster node.
- On the File menu, click Resume Node.
- Restart the active cluster node.
STATUS
This behavior is by design.
MORE INFORMATION
Setting node properties
For more information about setting node properties, visit the following Microsoft Web site:
For additional information about failover/failback policies on Microsoft Cluster Server, click the following article number to view the article in the Microsoft Knowledge Base:
197047 Failover/failback policies on Microsoft Cluster Server
Keywords: kbclustering kbprb KB823955
