HP-UX Workload Manager User's Guide

Example configuration files

entitlement_per_process.wlm

Chapter 9292

# Purpose:

# Demonstrate the use of a shares-per-metric goal.

# A machine hosts a group’s web servers. However, the group would

# like to use the rest of the CPU cycles without impacting the web

# servers, which are often not busy. From past experience, they

# have determined an httpd base entitlement (allocation) of 10

# percent of the CPU resources gives good response for random queries,

# and that sustained web processing can be done by providing an

# additional 5 shares per active (not sleeping) process. The httpd

# processes and their children (CGI scripts, servlets, etc) that were

# recently active are included in the count. See the glance(1)

# documentation for more information on available metrics.

# Components:

# The glance toolkit is used. See glance_app(1M) for more

# information on the glance data collectors.

# Allocation change frequency:

# Because wlm_interval is not explicitly set, the 60 second default

# is used. This means wlmd will collect measurements and make

# allocation changes every 60 seconds.

# Dependencies:

# This example was designed to run with HP-UX WLM version A.01.02 or

# later. It uses the cpushares keyword introduced in A.01.02, so

# is incompatible with earlier versions of HP-UX WLM.

# prm structure

# Create workload groups. Designate which workload binaries will

# be placed in each. We will be actively managing one workload,

# apache, and leaving the rest of the CPU resources to workload

# group OTHERS.

# See wlmconf(4) for complete HP-UX WLM configuration information.

prm {

groups = OTHERS : 1,

servers : 2 ;

apps = servers : /opt/hpws/apache/bin/httpd ;