HP-UX Workload Manager User's Guide

Chapter 4 127
4 How do I use WLM?
This chapter the basic steps needed to use WLM. The remaining
chapters explain these steps in detail.
The WLM configuration allows you to:
• Treat an entire vPar or nPar as a workload
• Create workloads based on PSETs or FSS groups to share a system
or partition among several workloads
You then assign shares-based or goal-based SLOs to the workloads.
Optionally, you indicate when each SLO is active.
If your configuration has workloads with performance goals, you can also
specify a means for collecting the performance data of those workloads.
Your performance collecting programs then use a utility called wlmrcvdc
or an API to provide the data to WLM. The utility and the API are
included with WLM and are explained in the section “What methods
exist for sending data to WLM?” on page 493. For more details about
using WLM with performance metrics, see Appendix H, “Advanced WLM
usage: Using performance metrics,” on page 467.
Next, WLM compares, at every interval, the reported and desired
performance levels for each SLO. WLM then adjusts, if needed, CPU
resources for the SLOs’ associated workloads based on the resources
needed to satisfy the SLOs and their assigned priorities.
Also, WLM generates SLO status information that is made available
through EMS.
Phasing in your WLM implementation
This section suggests one way in which to implement WLM. This is only
a suggestion and may not be appropriate in all situations.
Step 1. Implement a static configuration with usage goals.
A static implementation has a set number of workloads that are active at
all times. One way to implement such a configuration is to set up usage
goals for the workloads.