Hitachi

JP1 Version 12 Infrastructure Management: Getting Started


3.3 Analyzing bottlenecks

This section describes how to analyze a bottleneck for a problem that has been identified as an analysis target.

Context

The following figure shows the concept of bottleneck analysis.

[Figure]

The bottleneck analysis starts from the VM on which the error occurred. The bottleneck analyzed here serves as the starting point of the following analyses: the analysis of the importance and scope of the error, which will be performed later, and the analysis of the cause of the error.

Procedure

  1. In the E2E View window, to get a overview of the business system configuration.

  2. The VM in which the error occurred serves as the starting point of the analysis. Check whether there is a warning or error (indicated by a yellow or red icon) in the related components.

    [Figure]

    Analyze the warnings or errors (indicated by yellow or red icons) that exist in Disk and CPU of the server, and in Volume of the storage system.

  3. Click Volume of the storage system.

    User resources are displayed with their associations to the disk indicated by blue lines.

    Inform the storage administrator of the disk and volume for which a warning is displayed, because the disk for which the warning is displayed is associated with the volume.

    Proceed to the next operation to check whether the CPU for which an error is displayed is the bottleneck, because the CPU is not associated with storage.

  4. To check whether this CPU is the bottleneck, click the icon of the CPU. Then, from the displayed menu, select Verify Bottleneck to open the Analyze Bottleneck window.

    The CPU (bottleneck candidate) appears in a graph in the upper pane, and the VM (the VM where the problem occurred) appears in a graph in the lower pane.

  5. Compare the graph of the CPU with the graph of the VM to make sure that both graphs indicate the same trend.

    [Figure]

    If both graphs indicate the same trend, you can determine that the candidate is the bottleneck.

    In this example, because the graphs of the CPU and VM indicate the same trend, you can determine that the CPU is the bottleneck.

    Tip

    In the following cases, assume that the resource displayed in the E2E View window with a high sharing rate is the bottleneck, and determine that the resource is the bottleneck in steps 4 and 5:

    • When a warning or error (indicated by a yellow or red icon) is not displayed in steps 2 and 3
    • When you cannot determine the bottleneck in steps 4 and 5

    If you still cannot determine the bottleneck, perform a bottleneck analysis in the Performance Analysis View window.

    Note

    The method for analyzing errors that occur on switches is as follows.

    • The cause of the error might be a link-down that was detected when a configuration change results in a port no longer being used. If this is the case, in the detailed window for the switch, use the Assign Normal button to change the port status to Normal. Doing so will make the port status Normal until the status changes again, such as when the port is next used.
    • To check the resources related to a specific port, select the target port in the detailed window for the switch, and then click the Show Related Resources button. A list of related resources appears. From the list, you can select related resources and then open a new E2E View window in which those resources have been set as the base point.

Next steps

Use the Analyze Bottleneck window to analyze the impact and severity of the problem.