HP-UX Workload Manager User's Guide

Configuring WLM
Tuning the metrics and the SLOs
Chapter 5230
To determine the new CPU shares allocation, each controller effectively
executes an algorithm that, when plugging in cntl_convergence_rate
(as opposed to cntl_kp), is represented as follows:
New CPU allocation = (Allocation last interval) +
(cntl_convergence_rate / 0.10) * (P / goal)
where
cntl_convergence_rate
Is a tunable parameter for the controller that indicates
the number of CPU shares by which to adjust a
workload’s allocation when it differs from its goal. It is
specified in the WLM configuration file.
P
Is the deviation from the goal and is calculated as
shown in the previous section, “Tuning a workload’s
SLO convergence: cntl_kp (optional)”.
There are two special cases to consider in the previous formula:
• For goal = 0, “(P / goal)” becomes “P”
• For usage goals, goal becomes the average of the low_util_bound
and high_util_bound values: “(low_util_bound + high_util_bound) /
2”
Use the optional cntl_convergence_rate tunable to tune convergence.
It can be used in global, metric-specific, and metric/SLO-specific tune
structures. Its value is inherited from a higher level tune structure if you
do not specify it at a given level.
Its syntax is:
cntl_convergence_rate = number_of_shares;
where
number_of_shares
Is a floating-point value that represents the number of
shares (either absolute or relative) by which a
workload’s CPU allocation is adjusted when its service
level differs from its goal by 10%. When the difference
is less than 10%, the adjustment is proportionally