Capacity Advisor

HP Insight Dynamics - VSE Version 4.1 for ProLiant Installation and Configuration Guide

OpenPegasus Is Required for VSE Management of Linux Systems  OpenPegasus is required for VSE Management of Linux systems. RHEL distributions include OpenPegasus but do not install it by default.

Manually select OpenPegasus during RHEL installation, or install it later through the YaST installation and configuration tool from your distribution CD/DVD.

HP Capacity Advisor Version 4.1 User's Guide

Internet Explorer Times Out Before Report Generation Completes  In the case where you request a report using data from many managed systems (for example, over 100), Microsoft Internet Explorer may end the session connection before the report is complete. This ends the report creation; no report is generated for viewing or download.

One way to address this behavior is to create a registry key on your local Microsoft Windows system to extend the amount of time that Internet Explorer will wait before closing the connection, as follows:

Windows Registry Editor Version 5.00
[HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Internet Settings]
"MaxConnectionsPer1_0Server"=dword:00000005
"MaxConnectionsPerServer"=dword:00000005
"ReceiveTimeout"=dword:5265C0

HP Capacity Advisor Help

Different Paths to Profile Viewer Inconsistently Show Loads on VM Hosts  

A profile view accessed from the Visualization tab (which is showing historic resource utilization for real systems) may differ from a profile view accessed from the scenario editor (which is showing a simulation of the VM host that may have been modified for purposes of the simulation). This is expected behavior. The utilization values shown in the Visualization tab are based on the actual observed global utilization values collected on the VM Host. The utilization values for a VM Host in the Edit Scenario window Profile Viewer shows the simulated utilization computed from the utilization values for each VM and, possibly, one “OTHER” workload that is currently placed on that VM Host.

HP Capacity Advisor Version 4.1 User's Guide

In Rare Cases, VM Guest Network or Disk I/O Utilization May Appear to Exceed the I/O Utilization of the VM Host  When viewing historical network and disk I/O utilization profiles in the Capacity Advisor Profile Viewer, it may be noticed that the I/O utilization of a VM guest exceeds the utilization for its VM host at a given instance of time. Related to this, since the I/O high-water mark displayed for a VM guest (dashed blue line in the Profile Viewer graph) for a guest is based on the observed high-water mark for the host, it may be observed that a VM guest's I/O utilization will occasionally exceed the high-water mark value. In reality, it should never be possible for a VM guest's I/O utilization to exceed that of its host, since the host utilization is actually the aggregation of I/O utilization for all VM guest's running on that host.

The reason this may occur (in rare cases), is based on how data is collected separately, and in some cases from separate sources for hosts and guests, and how this data is interpolated and averaged into 5-minute sample intervals before being stored in the Capacity Advisor database. When sharp "peaks" occur in I/O utilization, it is very possible that these peaks will not align exactly in the same time intervals between host and guests, resulting in this anomaly. These anomalies become insignificant and in effect go away when averaging and analyzing this data over sample intervals greater than 5-minutes when possible.

Always collect data for both the host and guest(s) at the same time. Note that from within the Profile Viewer, a new data collection for the host also collects data for all of its guests. In most cases, minimizing the disparity in collection times between host and guest(s) will prevent large discrepancies from occurring (for example: seeing guest data exceed the high-water mark for the host). Minor discrepancies between host and guest I/O utilization may be unavoidable.