Article ID : 969597
Last Review : March 27, 2009
Revision : 1.0
Source: Microsoft Support

Symptom

You may experience very high CPU utilization on Cluster servers after installing the Cluster Management Pack Update 6.0.6505.0 for SCOM 2007. Regardless of the number of Processors (Cores), the load is evenly distributed, with no particular process identified as solely responsible for this high CPU usage.

Usually the Cluster Service (clussvc) and Resource Monitor (resrcmon) processes are identified as top CPU users.

This effect has been noticed especially on Cluster servers with a large (60+) number of resources.
Cause
The issue is caused by a problem in the WMI polling mechanism of the Resource State Monitor.

Resolution

Note The following processes are only appropriate when using the Cluster Management Pack referenced earlier.

To workaround this issue, you can implement one of the following options:

Option #1: Override the default frequency of the Resource State Monitor:
1. Open the Manager Console.
2. Choose Authoring -> Management Pack Objects -> Monitors.
3. Type Resource State Monitor in the Look for field, click Find now.
4. Right-click on the Resource State Monitor, choose Overrides: Override For all objects of type: Cluster Resource.
5. Put a check in the box beside WMI Poll Interval and increase the value under Override Setting.

Option #2: Disable the Resource State Monitor on the cluster nodes
1. Open the Manager Console.
2. Choose Authoring -> Management Pack Objects -> Monitors.
3. Type Resource State Monitor in the Look for field, click Find now.
4. Right-click on the Resource State Monitor, choose Overrides -> Disable the monitor.

*** The agents on the cluster nodes will now be made aware of this configuration change ***