Open Software Solutions

After you install the update in Microsoft Knowledge Base article 921181 on a computer that is running Windows Server 2003 with Service Pack 1 (SP1), you can create a Majority Node Set (MNS) cluster that has two physical nodes and a file share witness. This model is widely encountered in Microsoft Exchange Cluster Continuous Replication (CCR) configurations. In some particular situations, the cluster service cannot start.

For example, consider the following scenario:

•	You have two physical cluster nodes, Node A and Node B.
•	You have a computer that holds the MNS file share.
•	All resource groups are held by Node A.
•	You shut down Node A.
•	All resources fail over to Node B.
•	You shut down Node B.
•	You restart Node A.

In this scenario, the cluster service may not start. Additionally, the following events are logged in the System event log:

Event Source: Microsoft-Windows-FailoverClustering
Event Category: None
Event ID: 4522
Date: Date
Time: Time
Type: Error
User: SYSTEM
Computer: ComputerName
Description:
Majority Node Set resource has failed to come online. The resource has detected that it does not have the latest copy of server cluster database. If this is a two node cluster, try starting Cluster service on the other node, and let this node join the cluster. If that does not work, use the /forcequorum startup option to start the Cluster service.

Notes

•	This problem only occurs in this particular configuration. It does not appear in more than two physical nodes configurations.
•	This problem also occurs in Windows Server 2008 and in Windows Server 2003 SP2.

Back to the top

This problem occurs because Node A may not have cluster state information that was updated by Node B. Therefore, Node A may run the cluster by using incorrect state information. The file share witness feature helps prevent this problem by detecting that the cluster state has changed. The file share witness feature prevents the cluster node that contains outdated cluster state information from running the cluster. When this action occurs, Event ID 4522 is logged in the System event log.

Back to the top

To work around this problem, use one of the following methods.

Back to the top

Method 1

1.	Start the node that you shut down last. In the example situation, this is Node B.
2.	Start the node that you shut down first. In the example situation, this is Node A.

Back to the top

Method 2

If you start Node A first by mistake and you receive the Clussvc 4522 event, start Node B after. The cluster will then be restored.

Back to the top

Method 3

Implement an MNS cluster solution that has three or more physical nodes instead of two physical nodes and a file share witness.

Back to the top

Method 4

Set the startup type of the cluster service to Manual, start all nodes, and then start the service manually.

Back to the top

This behavior is by design.

Back to the top

The following event is logged in the System event log in Windows Server 2008:

Log Name: System
Source: Microsoft-Windows-FailoverClustering
Date: Date
Event ID: 1561
Task Category: None
Level: Critical
Keywords:
User: SYSTEM
Computer: ComputerName
Description:
The cluster service has determined that this node does not have the latest copy of cluster configuration data. Therefore, the cluster service has prevented itself from starting on this node. Try starting the cluster service on all nodes in the cluster. If the cluster service can be started on other nodes with the latest copy of the cluster configuration data, this node will be able to subsequently join the started cluster successfully.

If there are no nodes available with the latest copy of the cluster configuration data, please consult the documentation for ‘Force Cluster Start’ in the failover cluster management snapin, or the ‘forcequorum’ startup option. Note that this action of forcing quorum should be considered a last resort, since some cluster configuration changes may well be lost.

For more information, click the following article number to view the article in the Microsoft Knowledge Base:

Back to the top