This is a fairly wide ranging subject. The most common JES2 behavior I have observed, is that work tends to start on the submitting system in the JESplex.
This behavior can be influenced with a number of JES2 initialization parameters, Scheduling Environments, and SYSTEM AFFINITY or any combination thereof. One thing to understand, is that WLM is not the superior in the relationship JES is. JES handles job selection, queueing,... WLM enables that work to be processed. >From your original email, perhaps you were referring to the following paragraph from the first reference below: <quote> JES2 attempts to use approximately the same percentage of active WLM-managed initiators in each service class on each system. JES2 does not try to use the same number of initiators on each system; the number used is in proportion to the number started on each system. If WLM has started 20 initiators on system A and 10 on system B, and 15 jobs are submitted, 10 should run on system A and 5 on system B. </quote> A mixture of JES managed and WLM managed inits may give a different pattern than you were expecting. This has also been discussed before in the archives. The short answer is, if work is meeting the goals, do you really care it the batch workload is balanced across the JESplex (NOT the SYSplex)? This is definitely a non-trivial discussion, perhaps these references can help;(watch the wraps) http://www.redbooks.ibm.com/redbooks/pdfs/sg246472.pdf Chapter 6 http://www-03.ibm.com/servers/resources/servers_eserver_zseries_zos_wlm_ pdf_wlminits_pdf_wlminitsjm.pdf http://www-03.ibm.com/servers/resources/servers_eserver_zseries_zos_wlm_ pdf_cmgbatch_pdf_wlm_goal_based_initiator_management.pdf http://www-03.ibm.com/servers/eserver/zseries/zos/wlm/ HTH, <snip> In the last few days i have been discussing with a co-worker about how WLM balance batch Workload. We're z/OS 1.10 and the distribution of jobs among the JESPLEX does not seem to be smarter than in older releases. I've read that after z/OS 1.8 the algoritm would be enhanced, but the only thing that i found out is a JES2 command ($TJOBCLASS(x),XEQ(xxxx)=(MAX=nn) ) that limits the maximum number of jobs for each member of JESPLEX. There is another command ($TMASDEF,WLMPERF=YES/NO) that should do something but i could not find it documented in any IBM manual. Actually, the manual informs that the intention is not to reach an equally balanced distribution of batch jobs over the JESPLEX. WLM keeps starting inits while CPU is less than 95% utilization even with idle capacity in others systems. What enhancement is this ? </snip> ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
