Capacity Advisor Release Notes

Double Counting Utilization Data

It is possible to have the same process included in two monitored workloads. When this happens, the resources used by that process are added into both workloads, resulting in double counting.

This double counting can be seen in the profile viewer in Capacity Advisor. The profile graph for a system is different when viewed outside a scenario than when viewed inside the scenario. Data shown outside a scenario is the actual system total, whereas system data shown inside a scenario is the sum of the workloads. Double counting causes the trace inside a scenario to be higher.

Impossible Data Is Not Filtered

Occasionally erroneous data readings occur. Capacity Advisor filters out utilization data that is below 0, but it does not filter out all data that is unlikely or impossibly high. These stray readings have little effect on averages or on 90th-percentile values computed from the data, but they can significantly alter peak values.

Warnings During Initial Data Collection from a Virtual Machine

Depending on the order the data is collected, the capcollect command can generate the following warning the first time data is collected from a virtual machine:

Warning: unable to determine the 
    HPVM Host for Guest "guestname"
Valid CPU utilization data is
    not available.

HP SIM Displays Data in Multiple Time Zones

Some times in HP SIM are displayed in the time zone used by the browser, whereas other times are in the time zone of the CMS. Capacity Advisor times are always in the time zone of the CMS.

Default Report Begin and End Times Are Not Always Midnight

The start and end times of the default report are not always midnight to midnight, as might be expected.

Saving Capacity Advisor Report ZIP Files Can Make Internet Explorer Nonresponsive

If you are using Internet Explorer and you attempt to save a report ZIP file, two File Download popup windows can appear, and older versions of Internet Explorer can become nonresponsive. In other situations, the report creation screen becomes nonresponsive after saving the ZIP file.

Scenarios Can Model Unsupported Configurations

Capacity Advisor cannot always verify that the scenarios defined can actually be implemented on the systems being analyzed. For example, running a 4-way virtual machine on a 2-way VM Host is not supported, but it is possible to construct such a scenario.

Possible Data Loss When Virtual Machines are Shut Down and Restarted

Obtaining accurate utilization data for a virtual machine requires information from the Utilization Provider on both the virtual machine and its VM Host. The CPU utilization values for a virtual machine gathered from the Utilization Provider running within the virtual machine does not accurately reflect the true CPU utilization, but does provide the Memory, Network, and Disk utilization values. To obtain accurate CPU utilization values for a virtual machine, the values from the corresponding FSS partition on the VM Host are used. This requires matching the virtual machine with its FSS partition on the VM Host. If a virtual machine is shut down and restarted, it might run in a different FSS partition after restarting than it did before. When capcollect runs, it notes the current assignment of virtual machines to FSS partitions and uses that assignment to associate CPU utilization data from an FSS partition with the correct virtual machine. To prevent misleading CPU utilization values being recorded for a virtual machine, if a virtual machine has been restarted during a particular 24-hour period from midnight (in the GMT time zone) to the following midnight, all CPU utilization values in that period preceding the restart are not stored.

OVPA Network Data Differs from Capacity Advisor Network Data

Sometimes the network data imported from Openview Performance Agent (OVPA, also know as the Measureware Agent) does not match the data collected by Capacity Advisor. The differences seem related to the use of auto port aggregation.

Capcollect Can Fail After Updating the CMS from 11.11 to 11.23

One of the steps of the update process sometimes leaves some .save files in the directories reserved for Java plugins. When this happens the Java virtual machine tries to load these empty files and fails. This, in turn, causes the capcollect, vselicense, and capprofile commands to fail, even when the version of Java is correct, with a message similar to the following:

 # capcollect
 /opt/vse/lbin//capcollect[33]: could: The specified number is not 
 valid for this command.
 /opt/vse/lbin//capcollect[33]: could: The specified number is not 
 valid for this command.
 Unable to find an installed Java JRE meeting the minimum version 
 requirements.
 This command requires a minimum version of 1.4.2.07
 Unable to find a java executable.

The OTHER Workload is Deleted When the Last Workload on a Cluster Member is Deleted,

When the first monitored workload is created on a system a workload named OTHER is automatically created. When the last monitored workload is removed from a system the OTHER workload is automatically deleted. This is true for all systems, but on a Serviceguard cluster the workloads can automatically move to follow the cluster's packages. When the last monitored workload moves off of a cluster member, the OTHER workload is deleted.

When the OTHER workload is automatically deleted, its utilization history is deleted as well.

If a cluster member is shut down for any reason, all packages move off of the system to other cluster members. In turn, the CMS sees this and moves all the workloads associated with packages off of the system. Unless at least one workload is not associated with a package, the OTHER workload is deleted from the system that was shut down.

When a cluster member is restored, a new OTHER workload is created with no utilization data. This makes planning using the OTHER workload difficult.

Capacity Advisor Report Wizard GUI and capreport Command Use Different Default Values to Generate Reports

Capacity Advisor Report Wizard GUI and capreport command use different default values to generate reports for specific parameters. The following table shows the default values for each case:

The capreport Command Default Time Interval for Different Workloads Varies Within a Report

When two or more targets are used with data available in an overlapped date interval, the capreport command calculates a wrong report interval.

	IMPORTANT: The report content is right, the report date interval is wrong. This problem occurs only when no date interval is provided in the command line options.

Null Message Included in capreport Output

If the options on the capreport command line cannot be parsed, the message Null can be returned.

Report Help Link Does Not Work in capreport HTML Reports

In the HTML report generated by the capreport command, clicking the “?” help-link has no effect.

capovpaextract Command Dependencies

This command uses the remote execution facilities provided by HP SIM and, therefore, requires the configuration of SSH authentication within HP SIM for the specified managed node.

Capacity Advisor supports the following versions of OVPA :