HP-UX Workload Manager Toolkits User's Guide

HP-UX WLM Duration Management Toolkit and HP-UX WLM Toolkit for Base SAS Software
How do I use DMTK / SASTK?
Chapter 6 133
# express_lane workload group with the prmrun command:
#
# % prmrun -g express_lane <app_name> <app_arguments>
# users = jdoe : express_lane, +marketing_netgroup : express_lane;
# Line is commented out because the undefined user and netgroup
# would result in an error if this file were checked by wlmd.
}
# Set the wlm_interval value shorter than the default value of 60 (seconds)
# if the condition metric express_lane_procs_active needs to be updated
# more than once a minute. Here, the interval is set to 5 seconds.
# A shorter wlm_interval allows the SLO to react more quickly.
tune {
wlm_interval = 5;
}
# Use a tune structure with glance_prm to determine the number of active
# processes in the express_lane group. This structure generates the metric
# express_lane_procs_active, which is used to make the SLO below inactive
# when no processes are present in the express_lane workload group.
tune express_lane_procs_active {
coll_argv = wlmrcvdc glance_prm -i 2 APP_ACTIVE_PROC express_lane;
}
# Next, create the SLO slo_express_lane. This SLO has a priority of 1.
# Also, it has a fixed CPU request. A priority 1 SLO has its shares
# allocated before a lower priority SLO. Thus, WLM will satisfy
# slo_express_lane at the expense of the group OTHERS. In addition, the SLO
# has a condition metric, express_lane_procs_active. This condition metric
# specifies that at least one process must be executing in the express_lane
# workload group before the SLO becomes active.
#
# This SLO is inactive when there are no processes in its associated
# group. Otherwise, the SLO provides 45 CPU shares to the group.
slo slo_express_lane {
pri = 1;
entity = PRM group express_lane;
cpushares = 45 total;
condition = metric express_lane_procs_active > 0;
}