[z/OS]

Tuning WebSphere MQ activation specifications for workload management on WebSphere Application Server for z/OS

Tune your WebSphere® MQ activation specifications to enable effective workload management (WLM) by calculating the optimal value for the Maximum server sessions property.

When there is a backlog of messages on the WebSphere MQ queue for your message-driven bean (MDB), you want some messages to be processed before other message based on transaction class.

The workload management (WLM) feature of WebSphere Application Server for z/OS® allows messages to be processed in a certain order based on transaction class. Workload management prioritization occurs when message IDs have been placed on the WLM queue. To allow effective WLM prioritization you must, therefore, make sure that the WLM queue is sufficiently loaded with message IDs, so that the MDB has a selection of messages, with different transaction classes, to choose from. If there are not enough messages on the WLM queue, the MDB will take the first available message on the WLM queue and transaction class will not be considered.

To ensure that the WLM queue is loaded sufficiently to allow WLM prioritization, set the Maximum server sessions property on the activation specification to a value higher than the baseline recommendation of "twice the combined number of worker threads in all the servants for the server ."

  1. To tune your configuration for WLM you need to determine the following values:
    • The average number of servant worker threads processing a given MDB.
    • The average number of available servants (some number between the minimum and maximum is started at any given time).

    These values can be estimated by using Performance Monitoring Information (PMI), other monitoring tools, or perhaps by a high-level understanding of how the message-driven bean fits into the general application flow of a specific server.

  2. Set the Maximum server sessions property on the listener port to one of the following values:
    • Twice the number of worker threads that are available for the maximum number of servants in the scalable server.
    • Twice the number of worker threads that are available in all servants.

    Too low a setting for Maximum server sessions causes idle worker threads. Too high a setting can cause extra messages to build up on the WLM queue, but the extra messages should not be sufficient to overload the WLM queue and cause the server to fail.




Related concepts
The message-driven bean throttling mechanism on z/OS
MDB throttle settings for message-driven beans on z/OS
Connection factory settings for ASF message-driven beans that use WebSphere MQ as the messaging provider on z/OS
Messaging flow for ASF message-driven beans with WebSphere MQ as the messaging provider
Tuning listener ports for workload management on WebSphere Application Server for z/OS
Tuning default messaging provider activation specifications for workload management on WebSphere Application Server for z/OS
Messaging flow for JCA message-driven beans with WebSphere MQ as the messaging provider
Concept topic Concept topic    

Terms and conditions for information centers | Feedback

Last updatedLast updated: Feb 5, 2014 9:49:51 PM CST
http://www14.software.ibm.com/webapp/wsbroker/redirect?version=compass&product=was-nd-mp&topic=cprf_tunezmdb_mqacspecwlm
File name: cprf_tunezmdb_mqacspecwlm.html