Using the Workload View

The Workload View, accessed from the Workload tab, allows you to view all workloads in the VSE. This section describes the purpose of the Workload View, workload types, how to view and modify workloads, and the main components of the Workload View.

Purpose of the Workload View

The Workload View allows you to view all of the workloads in the VSE, using a compact table format. The Workload View shows the real-time utilization metrics of all workloads, allowing you to quickly spot overloaded and under used resources. From this view, you can modify workload properties by selecting the workload name, or you can select workloads for further actions.




	NOTE: Some workloads have a system hostname as the workload name. When selecting items in the Workload View, be aware that you are selecting workloads and not systems.

For information about the different types of workloads that appear in Virtualization Manager, see “Understanding Workload Types”. For a description of the components in the Workload View, see “Components of the Workload View”.

Understanding Workload Types

The term “workload” applies to each of the following distinct but closely related concepts in the VSE Management Software:

Monitored workload A workload that can be monitored by Virtualization Manager but has no policy associated with it. Monitored workloads do not belong to an SRD and are not managed by Global Workload Manager (gWLM).
For information about creating and modifying monitored workloads, see “Creating Monitored Workloads”.
Managed workload A workload that belongs to an SRD and is managed by Global Workload Manager (gWLM). A managed workload has a policy assigned to it. You can change a monitored workload into a managed workload by assigning a policy to it. Select the check box next to the workload in either the Visualization View or the Workload View, then choose PolicyApply gWLM Policy.... To complete this action, you will be required to create a new SRD. For more information on creating an SRD on gWLM, see the HP Global Workload Manager Version 4.1 User's Guide.
NOTE: A system may contain monitored workloads or managed workloads, but may not contain both types of workloads.
Whole-OS workload The set of all processes running on a system (for example, server, nPartition, virtual partition, or virtual machine). The utilization metrics displayed for the whole-OS workload allow you to monitor the total resource utilization on the system. The whole-OS workload is not displayed as a workload in the Visualization View, but it is represented by the compartment for the system. It appears in the Workload View, where you can click the workload name to display the Workload Definition Properties screen. From that screen, you can modify the workload name and description but you cannot define any Workload Definition Criteria for the whole-OS workload.
NOTE: In the current release of the VSE Management Software (version 4.1), the whole-OS workload is the only type of workload that is supported on systems running Microsoft Windows.

OTHER workload Whenever there are user-defined workloads on a system, Virtualization Manager creates a special workload named “system-name.OTHER”. This workload accounts for all processes on the system that do not match any of the user-defined workload definitions. It is displayed like any other workload, except that you can only modify its name and description, and you can only delete it when no other workloads remain on the system. Under most circumstances, the “OTHER” workload is automatically deleted when the last user-defined workload on the system is removed.

The OTHER workload is also created on Integrity VM hosts; it is not created for Microsoft virtual servers or ESX. This workload accounts for all manageable resources that are not allocated to any virtual machine on that host. This workload cannot be deleted.




	NOTE: Global Workload Manager tracks processes in the “OTHER” workload differently than Virtualization Manager and Capacity Advisor. For more information on how gWLM tracks processes, see the HP Global Workload Manager Version 4.1 User's Guide.

Creating Monitored Workloads

You can create monitored workload definitions from the either the Visualization View or the Workload View, as follows:

Visualization View

From the Visualization tab, select a system by clicking the compartment check box. Virtualization Manager highlights the compartment.
Choose CreateWorkload Definition... from the Virtualization Manager menubar. The Create Monitored Workload Definiton page is displayed.

Workload View

From the Workload tab, choose CreateWorkload Definition... from the Virtualization Manager menubar. The Create Monitored Workload Definiton page is displayed.

On the Create Monitored Workload Definition page, you can define general workload properties and specific workload definition criteria. General workload properties include the following information:

The name of the system where the workload will reside. The drop down menu lists all available systems.
If Serviceguard clusters are available, the name of the Serviceguard cluster on which to create the workload. A message displays a list of members of that cluster (cluster nodes). After selecting a cluster, the page automatically refreshes to populate the Package selection box, which lists all the Serviceguard packages on the chosen cluster.
The workload name, which identifies the workload in the Visualization View and Workload View. Each workload must have a name that is unique across the CMS.
Optionally, a brief description of the workload.

You can specify criteria that define the processes included in the workload. In the Specify Workload Definition Criteria field, a drop down menu lets you define criteria by four selection criteria types:

by user: All processes owned by the specified user ID are included in the workload.
by executable: All processes executing the specified file are included in the workload. You can specify alternate names for the same executable file, including wildcard patterns.
by application: All processes determined to be part of the specified application are included in the workload.
by process map: All processes identified by the process map are included in the workload.

The workload definition criteria displays as a row in the table on the bottom of the page. Any process on the specified system that matches any row of the table is included in the workload.

When you create a new workload, there are no initial criteria defined. You can add new rows and delete rows that have been added. You cannot modify an existing row; instead, delete the row and add a new row in its place.

Viewing and Modifying Workload Definitions

You can view and modify existing workload definitions from the either the Visualization View or the Workload View, as follows:

Visualization View

From the Visualization tab, select the check box next to Show Workloads on the Virtualization Manager toolbar. Workloads display in the compartments.
Hover over the information icon to display an information popup.
Click the Edit or View Workload properties link on the bottom of the page. The Monitored Workload Definiton Properties page is displayed.

Workload View

From the Workload tab, click the name of the workload.
Alternately, select the check box next to the workload name and then choose ModifyWorkload Definition... from the Virtualization Manager menubar. The Monitored Workload Definition Properties page is displayed.

Updating System Workloads

You can update all monitored workloads on the system:

In the Visualization View, select a system by clicking the compartment check box. Virtualization Manager highlights the compartment. (You can select only one system at a time.)
Select the ToolsUpdate System Workloads menu action.

All monitored workload definitions on the selected managed system are replaced. This action first deletes all existing workload definitions on the selected managed systems. It then pushes out the workload definitions on the CMS that apply to the selected managed systems to those systems. Data that had been previously stored by the recovered workloads is not lost; rather, it is accessible after the workload definition is pushed back to the selected managed system.




	NOTE: This action only affects monitored workloads and has no effect on managed workloads. Therefore, this operation cannot be used to restore gWLM workload definitions.

Components of the Workload View

Figure 2-3 shows the major components of the Workload View.

Figure 2-3 Example Workload View

The Virtualization Manager tabs, with the Workload tab selected (see “Virtualization Manager Tabs”).

The VSE Management menu bar (see “Virtualization Manager Menus”).

By default, the Workload view limits 50 workloads per page. If your display contains more than 50 workloads, a pagination link displays (Next >>), allowing you to toggle to the next page of workloads. You can continue to toggle forward through the entire range of workloads, or use the << Previous link to toggle back to the previous page.

You can change the number of workloads displayed per page by modifying the user preference for workload pagination. For information, see “Modifying User Preferences”.

The selection check boxes. Use the check boxes to select one or more workloads, then choose an action to perform from the VSE Management menu bar.

The workload name. You can click the name to display the Workload Definition Properties screen for this workload. From there, you can modify the properties that define the workload.




	NOTE: Some workloads have a system hostname as the workload name. When selecting items in the Workload View, be aware that you are selecting workloads and not systems.

By default, workloads are sorted alphabetically by the Name column header. The arrow next to the header name shows the alphabetic sort order; by default, A to Z. When you click on the Name column header, the alphabetic sort order reverses. You can also change the sort order criteria by clicking on other column headers. For example: clicking on the Type header sorts by workload type and highlights that column:




	NOTE: When sorting on the Utilization columns, sorting is done by current percentage of maximum, not absolute utilization. The meter's utilizations will be in sort order but the numeric values to the right of the meters will not appear in order.

The workload type is displayed as one of the values defined in Table 2-3.

Table 2-3 Workload Type Values

Value	Definition
Cluster	A workload associated with a package on a Serviceguard cluster.
Fair Share Scheduler	A managed workload based on a Fair Share Scheduler (FSS) group.
Integrity VM	The whole-OS workload on an HP virtual machine.
Monitor	A workload monitoring a user, process, application, or a process map on a system.
nPartition	The whole-OS workload on an nPartition.
MS Virtual Server VM	The whole-OS workload on a Microsoft Virtual Server virtual machine.
Parked	A parked workload (not associated with a specific system). Parked workloads are useful while migrating a workload from one system to another, in order to preserve Capacity Advisor historical data.
Processor Set	A managed workload based on a processor set.
Server	The whole-OS workload on a standalone server. This could be a virtual machine host.
SRD Member	A workload that is part of a Shared Resource Domain (SRD).
Virtual Machine	The whole-OS workload on a generic virtual machine.
Virtual Partition	The whole-OS workload on a virtual partition.
VMware ESX VM	The whole-OS workload on a VMware ESX virtual machine.

For a description of the types of workloads, see “Understanding Workload Types”.

The system on which the workload is defined. To examine the system properties, click the system name (or IP address). This will display the SIM System Page for that system. This field remains empty for parked workloads because the system property is set to “none”.

For managed workloads, this field displays the name of the gWLM policy associated with the workload. To examine or modify the policy definition, click the policy name. This displays the gWLM Create or Edit Policies screen for that policy. For servers and monitored workloads, this field is displayed as “---”.

The real-time utilization metrics from each server and workload. This figure only shows the CPU Utilization metric. In the actual screen, the other metrics appear to the right for memory utilization, network I/O, and disk. You can view the historical data for any metric by clicking its meter, which displays the Capacity Advisor Profile Viewer screen. If no data has been collected yet, you can begin collecting data from that screen. For more information on the Capacity Advisor, see the HP Capacity Advisor Version 4.1 User's Guide.

An error indicator appears if there is a problem communicating with the managed system. The utilization meters might show one of several error indicators instead of valid data. When you hover the cursor over the error indicator, a message appears further explaining the error.