HP-UX Workload Manager User's Guide

Example configuration files
par_manual_allocation.wlmpar
Chapter 9 307
# current partition’s applications relative to the importance of the
# applications in all the other partitions.
#
# When managing partitions, WLM equates 1 core of CPU resources to 100
# shares. The cpushares statement causes the SLO to request 100 shares, or
# 1 core, multiplied by the metric num_cpus. So, if num_cpus = 7, the SLO
# requests 7 cores for the partition.
#
slo slo_myslo {
pri = 1; # Change this value
cpushares = 100 total per metric num_cpus;
}
#
# The following structure sets up the metric num_cpus that is used in the
# cpushares statement above. This set up allows you to change the metric’s
# value using wlmsend on the command line. See Steps 7 and 8 above for
# usage and monitoring examples.
#
tune num_cpus {
coll_argv = wlmrcvdc;
}
For more information on WLM partition management, see Chapter 7,
“Managing SLOs across partitions,” on page 255.
par_manual_allocation.wlmpar
This file is a configuration file for the WLM global arbiter. In
combination with the WLM configuration file in the previous section,
this file can migrate cores across HP-UX Virtual Partitions and/or
nPartitions based on the number of cores you request for a partition by
using the wlmsend command. The way WLM manages cores depends on
the software enabled on the complex (such as Instant Capacity, Pay per
use, and Virtual Partitions). This file also shows how several additional
par structure keywords (including utilitypri) are used (these are
commented out).
Activate the WLM global arbiter’s configuration in only one partition.
However, activate the WLM configuration file in each partition.