Changes in This Update

Applies To: Operations Manager 2007

Version 6.0.6720.0 of the Windows Server Failover Cluster Management Pack includes the following changes:

  • Added support for discovery and monitoring of Windows Server 2008 R2 clusters.

  • Disabled discovery and monitoring of resources by default. If you are upgrading from a previous version of the management pack, resource monitoring will be disabled as well. Note that this monitoring is still present in the management pack, however it is not recommended to enable it as sufficient monitoring is provided by the resource group monitoring. Enabling resource monitoring may also significantly increase CPU utilization of the Operations Manager agent. For more information, see Supported Configurations.

  • Updated the management pack to discover and monitor cluster shared volume resources only on Windows Server 2008 R2 clusters.

  • Limited discovery of resource groups to a maximum of 300 resource groups per cluster. You can override this for individual clusters, but it may impact CPU usage of the Operations Manager agent. For more information, see Supported Configurations. Event-based monitoring is available for all resource groups in the cluster. Failover clusters with more than 300 hundred groups are likely to be managing Hyper-V virtual machines, which can also be monitored by System Center Virtual Machine Manager.

  • Disabled discovery of clustered objects on an Operations Manager root management server (RMS) cluster running Operations Manager 2007 R2. RMS clusters will still be discovered if you were using a previous version of the cluster management pack on Operations Manager 2007 R2. If you want to remove the clustered RMS objects, you will need to delete the old version of the management pack, and then import this version.

  • Fixed an issue where the DiscoverClustering.vbs script would fail to submit discovery data when there were WMI provider issues on the cluster node. This would also generate an alert. The script now generates empty discovery data when this condition is encountered; this will not undo discovery of existing cluster resources. An event (event id 7001 from source Health Service Script in the Operations Manager log) will be generated to the local event log signaling that this condition was encountered.

  • Updated display strings and knowledge in the management pack.

  • Updated discoveries to run every 4 hours.

  • Updated the MAC Address property for the network adapter to use hyphens instead of colons.

  • Disabled the Windows Time Service monitor for Windows Server 2008 and Windows Server 2008 R2 clusters because the cluster service has no dependency on the Time service.

  • Disabled the WMI service monitor for all operating system versions of the management pack because the cluster service has no dependency on the WMI service.

  • Added a Sync Time parameter to the discoveries and monitors for resources and resource groups to ensure that these do not run simultaneously or immediately after an agent is restarted.

  • Disabled the following rules on Windows Server 2003, Windows Server 2008 and Windows Server 2008 R2 clusters, which would cause discovery to instantly run when cluster properties changed:

    • Cluster object property changed

    • Cluster object added

    • Cluster object removed

    • Cluster object changed property value

    • Cluster object space changed due some addition

    • Cluster object space changed due some deletion

    Cluster changes are now discovered on a schedule (by default, every 4 hours). It is not recommended to enable these rules, as they and the scheduled cluster discovery need to share exactly the same configuration in order for discovery to work predictably. For example, if one rule has the override Discover Cluster Resources set to “false” and the scheduled discovery has the override set to “true”, these resources will be randomly discovered and undiscovered depending on which rule that actually ran the discovery.

The April 2009 update to this management pack included the following changes:

  • Fixed an issue with the Resource State monitor that could cause high CPU utilization on clusters with a large number of resources. Also, the default interval of this monitor is changed to run every five minutes. This interval determines how quickly Operations Manager detects that a cluster resource fails. For more information, see the Supported Configurations section.

  • Updated the default interval of the Resource Group State monitor to run every five minutes. This interval determines how quickly Operations Manager detects that a cluster resource group fails. For more information, see the Supported Configurations section.

The January 2009 update to this management pack included the following changes:

  • The resource group names now include the cluster name so that the resource group can be easily distinguished when adding them to an object group.

  • Introduced support for monitoring Windows Server 2008 failover clusters.

  • Fixed issues with the Delete Group task to address problems with handling dependencies between resources and offline resources.

  • Fixed an issue where cluster resource group discoveries would fail for groups that had leading or trailing spaces in their names.

  • Fixed an issue with how the management pack handled moving resource groups when the preferred node was offline.

The April 2008 update to the Windows Server Failover Cluster Management Pack included the following changes:

  • General changes to the Windows Server Failover Cluster Management Pack

  • A resource group with at least one offline resource was reported as offline.

  • Resources containing the backslash (“\”) not monitored properly.

  • Discovery rules that react to cluster changes will not work in low privilege environment and may not work even if the Run As profile for the cluster management pack is configured.

  • Alert/State/Event views moved from Microsoft.Cluster.Management.Library.mp to Microsoft.Cluster.Management.Monitoring.mp.

  • Enable EventLog Replication property has been removed from the Cluster Node class.

  • Resource groups with special characters such as quotation marks in their name cause discovery problems.