HP-UX Workload Manager Toolkits A.01.07.01 Release Notes for HP-UX 11.0, HP-UX 11i V1, and HP-UX 11i V2

For more information, see wlmwlsdc(1M) and the white paper Using HP-UX
Workload Manager with BEA WebLogic Server, which offers various use cases.
This paper is available at
/opt/wlm/toolkits/weblogic/doc/weblogic_wlm_howto.html.
The new features of these toolkits are described in the next section.
What’s in this version
WLMTK Version A.01.07.01 provides the toolkits ApacheTK, DMTK, ODBTK, PPUTK, SASTK,
SNMPTK, and WebLogicTK. The new functionality of the individual toolkits is described below.
• PPUTK’s utilitydc has been modified as indicated below:
◦ -e
This new option instructs utilitydc to use the EMS resource
/applications/wlm/icod_reserves_needed. This was the default behavior before this
release of WLMTK.
◦ -g workload_group
By default, when you use the prm activation/deactivation algorithm, the CPU allocation
of the OTHERS group determines when processors are activated or deactivated.
This new option allows you to use a group besides OTHERS as the basis for the activation
and deactivation.
◦ -t
This option, which allows you to specify a tuning value for the activation/deactivation
algorithms, now has a different behavior when used with the loadavg algorithm.
◦ How utilitydc works
utilitydc no longer uses the EMS resource /applications/wlm/icod_reserves_needed
by default. If you do not use this EMS resource elsewhere, you can remove the
icod_thresh_pri and icod_filter_intervals keywords from your WLM
configuration files. (These keywords cause WLM to maintain the EMS resource.)
To use this resource, specify the -e option.
◦ Updated example configurations
The example configurations have been revised.
For more information, see the utilitydc(1M) man page.
Known problems and workarounds
This section discusses problems and workarounds.
ODBTK: No control of instance on start-up
Issue Assume a database instance is managed by HP-UX WLM. If WLM starts before that instance,
it sets the CPU allocation for the instance’s workload group based on the mincpu and maxcpu
values in the WLM configuration file. The instance’s workload group remains at that initial
CPU allocation until the instance starts, allowing wlmoradc to gather metrics on it and pass
the metrics to WLM.
Workaround Try to minimize the time between WLM start-up and instance start-up.
6 HP-UX WLM Toolkits (WLMTK) Release Notes