WLMTK Overview: Using HP-UX WLM Effectively With Your Most Critical Applications

How to use WebLogicTK

For a discussion about using WebLogicTK, read the white paper, “Using HP-UX Workload Manager

with BEA WebLogic Server,” available from the information library at http://www.hp.com/go/wlm.

The paper guides you through example WLM configurations that show how to manage various types

of WebLogic Server instances. For example, the paper shows you how to:

• Manually provide a single instance an increasing amount of CPU resources using a dynamic pSet

for benchmarking

• Separate an instance from other workloads, as well as from other instances, while automatically

maintaining performance using a dynamic pSet based on:

– CPU usage of the workload

– Server instance queue metrics

Example configuration files

The following files are example WLM configuration files. Each .wlm configuration file represents a use

case documented in the white paper.

• manual_cpucount.wlm—Helps with benchmarking and capacity-planning tasks and demonstrates

how to resize the workload (WebLogic instance) pSet by manually running wlmsend. For example,

the following command would assign two cores to the wls1_grp:

% /opt/wlm/bin/wlmsend wls1_grp.desired.cpucount 2.0

• wls_1inst_3level.wlm—Demonstrates how to control a single WebLogic instance and monitor its

execution queue to determine if it is idle (normal allocation), busy (boost of one additional core), or

very busy (boost of two additional cores).

• wls_2inst_3level.wlm—Demonstrates how to control two WebLogic instances and monitor the

execution queue of each to determine if it is idle (normal allocation), busy (boost of one additional

core), or very busy (boost of two additional cores).

• wls_1inst_CPUusage.wlm—Demonstrates how to control a single WebLogic instance as opposed to

other workloads. The WebLogic instance’s workload starts with one core. As its CPU consumption

grows, it is assigned more cores, to a maximum of four. For a discussion of usage goals, see the

wlmconf(4) manpage or the HP-UX Workload Manager User’s Guide.

• wls_2inst_CPUusage.wlm—Demonstrates how to control a pair of WebLogic instances. Each

WebLogic instance’s workload starts with one core. As each instance’s CPU consumption grows, its

workload is assigned more cores to a maximum of four. For a discussion of usage goals, see the

wlmconf(4) manpage or the HP-UX Workload Manager User’s Guide.

• wls_1inst_q_goal.wlm—Demonstrates how to control a WebLogic instance. The WebLogic

instance’s workload gets from one to four cores, which wlmd allocates in an attempt to keep the

instance’s queue_busy metric below 0.

• wls_2inst_q_goal.wlm—Demonstrates how to control a pair of WebLogic instances. Each

WebLogic instance’s workload gets from one to four cores, which wlmd allocates in an attempt to

keep the instance’s queue_busy metric below 0.

• wls_2queue_2inst.wlm—Demonstrates how to control a pair of WebLogic instances, instA and

instB. InstA is higher priority, and the administrator has created a special ”hipri” queue for high

priority work. Each WebLogic instance’s workload starts with one core. If the hipri queue depth is

nonzero, the instA workload gets an extra core. If the instA default queue depth is nonzero, the

instance’s workload gets an additional core. Lastly, if the instB queue depth is nonzero, the

instances’ workload gets an additional core if any cores are left.