3.5.2 Removing the cause

The following describes how to remove the cause. JP1/OA can be used to propose recovery plans to improve the performance of system resources based on information related to events and metrics that have caused a bottleneck. You can easily resolve a problem by taking action according to the proposed recovery plans.

Procedure

1. In the Analyze Bottleneck window, click Check Recovery Plan.

   

   A window for setting the conditions to generate recovery plans appears.

2. In Specification Method, make sure that the radio button for Populate from Selected Event is selected, and then click Select Event.

   A window displaying a list of events from which recovery plans are to be created appears.

3. Select the event that caused the bottleneck, and then click OK.

   In 3.3 Analyzing bottlenecks, the CPU was found to be a bottleneck. Search the list and select a warning event related to CPU usage for the entire host.

4. Specify the target value for CPU usage for the entire host in Target Value and click the Generate Recovery Plans button.

   This will display a recovery plan that brings the CPU usage for the entire host close to the target value.

   If multiple recovery plans are displayed, compare the details of each plan, and then consider which plan will be effective in resolving the problem.

   Note

   Depending on the cause of the bottleneck, appropriate recovery plans might not be displayed. In this case, take action based on the cause that you analyzed in the Analyze Bottleneck window.

5. Modify the configuration as necessary based on the information provided in the recovery plan.

6. In the Analyze Bottleneck window, make sure that all graphs (VM, volumes, and MP) are the same, and that the graphs display a straight line.

   If both the line of the graph for the VM on which the failure occurred and the line of the graph for the CPU of ESX that was determined to be the bottleneck are below the threshold, the problem no longer exists.

To Page Top