HP-UX Workload Manager User's Guide

Configuring WLM
Defining SLOs
Chapter 5202
Figure 5-5 Usage goal conceptually
Here is an example of a usage goal statement:
goal = usage _CPU 60 80;
When WLM requests an allocation change, the request is proportional to
how far the current utilization is below low_util_bound or above
high_util_bound. Use cntl_kp or cntl_convergence_rate in a tune
structure with _CPU_group_name as the metric value to adjust the
proportion, causing the utilization to move into the desired range more
quickly or slowly. For example, a cntl_kp value of 10 will cause the
utilization to move in larger jumps than will a value of 0.1.
To get an idea of how such a tune structure would look, consider a
workload named buy. Its tune structure would look like the following:
tune _CPU_buy {
cntl_kp = 0.1;
}
The cntl_kp value must be between 0 and 1,000,000 inclusive. The
default value is 1.
For more information on cntl_kp or cntl_convergence_rate, see the
section “Tuning a workload’s SLO convergence: cntl_kp (optional)” on
page 224.
Goal: Keep utilization between 50% and 75%
Utilization
(CPU used) / (CPU allocation)
0
50
75
25
100
➊
➋
➌
With utilization above 75%, the workload is pretty
In this range, the workload is using a reasonable
With utilization less than 50%, the workload is not
busy and is using most of its allocation. WLM
increases its allocation to ensure it gets enough
CPU. This moves the utilization to less than 75%.
CPU very efficiently and would suffer with less CPU.
CPU to give to other workloads. With a smaller
amount of its allocation. It seems to efficiently use
the CPU it has, but it would not be able to use more
very busy. WLM takes away some of its unused
CPU allocation, the workload’s utilization moves
above 50%.
➊
➋
➌