HP-UX Workload Manager User's Guide

Configuring WLM
Tuning the metrics and the SLOs
Chapter 5 229
Tuning a workload’s SLO convergence:
cntl_convergence_rate (optional)
NOTE You can also tune convergence with the cntl_kp tunable discussed in the
section “Tuning a workload’s SLO convergence: cntl_kp (optional)” on
page 224. (If cntl_convergence_rate is not zero, it is used instead of
cntl_kp.)
Using the cntl_convergence_rate tunable changes the allocation
algorithm to take the goal value into consideration, effectively
normalizing the result by the goal’s magnitude. Given the normalization
that cntl_convergence_rate provides, as well as the greater analysis
required to fine-tune the cntl_kp tunable, you may want to use
cntl_convergence_rate first. In many cases, setting
cntl_convergence_rate = 1.0 will provide the desired convergence and
meet your needs.
For each performance and usage goal, WLM must determine the
appropriate number of CPU shares needed to achieve that goal. To do
this, WLM creates a controller for each such goal.
NOTE Although convergence tuning is optional from a syntactic point of view,
you should use cntl_convergence_rate or cntl_kp for each controller
to fine-tune it to the characteristics of its workload.