HP-UX Workload Manager User's Guide

Example configuration files

manual_entitlement.wlm

Chapter 9298

# Tell wlmrcvdc to watch for metrics coming in via command lines:

# % /opt/wlm/bin/wlmsend wls1_grp.desired.cpucount 1

# or

# % /opt/wlm/bin/wlmsend wls1_grp.desired.cpucount 2

tune wls1_grp.desired.cpucount {

coll_argv = wlmrcvdc ;

}

# Check for new metrics every 5 seconds.

# Also, turn on absolute CPU units, so resources on a 4-core box are

# represented as 400 shares instead of 100 shares.

tune {

wlm_interval = 5;

absolute_cpu_units = 1;

}

For information on the cpushares keyword, see “Specifying a

shares-per-metric allocation request (optional)” on page 205. For

information on using wlmrcvdc and wlmsend in this manner, see

“Sending data from the command line” on page 495.

manual_entitlement.wlm

This next example is helpful in determining what type of performance to

expect from a workload at various amounts of CPU resources. With that

information, you can create better service-level objectives for the

workload. As such, this example is a good model to follow when setting

up your WLM configuration.

# Name:

# manual_entitlement.wlm

# Version information:

# $Revision: 1.14 $