HP-UX Workload Manager User's Guide

Managing SLOs across partitions
Overview
Chapter 7256
Figure 7-1 WLM managing partitions
WLM
EMS
configuration
wlminfo /
wlmgui
WLM monitoring tools
( wlminfo or wlmgui )
Service (EMS)
Event Monitoring
SLO stats
(pass/fail)
WLM daemon (wlmd)
Arbiter
Define
Shares-per-metric
Workload Grp A
Allocation
Usage Data
Collector
Workload Grp B
Controller
Controller
(metric goal)
(usage goal)
Fixed
Allocation
Workload Grp C
(no goal)
Message log
/var/opt/wlm/msglog
Statistics log (optional)
/var/opt/wlm/wlmdstats
Global arbiter statistics
log (optional)
/var/opt/wlm/wlmpardstats
Metric Data
Collector
Workload Grp D
(no goal)
Audit data files (optional)
/var/opt/wlm/audit/
Workload groups get
new CPU allocations
better suited to their SLOs
workloads
and SLOs
WLM
file
configuration
global arbiter
Grp A
Grp B
Grp C
WLM
EMS
configuration
Log
files
WLM
Log
files
PRM
Par 0
Par 1 Par 2 Par 3
vPars get
cores added
or removed
to better
suit the SLOs
HP-UX server
WLM
file
configuration
PRM
WLM creates a new PRM configuration
wlminfo /
WLM global arbiter
daemon (wlmpard)
With nPars, a
core is
deactivated
on one nPar, then
another nPar
WLM can manage
vPars, nPars,
or a combination
is activated
a core on
wlmgui
file
file
WLM
App1 App2
App3
WLM
EMS
wlminfo/
wlmgui
App1 App2
App3
Log
files
configuration
file
WLM
Usage Data
Collector