I haven't dealt with this in this detail for some time. It is important for some work at hand that I do have to correct understanding. Please advise me where I'm wrong.
Neglecting system affinity and scheduling environments for the time being. These may further restrict where jobs may run. But this is not of interest right now. In pre-WLM times and still today when it comes to *JES-managed initiators*, it is the operator or automation that starts initiators on individual systems in the sysplex. Once initiators are running, each of them asks JES for work (jobs) independently. So, the initiator which is first in asking for a job from a specific jobclass will get the one on top of the jobclass queue assigned. In other words, JES is *not* in control to tell any specific initiator on any specific system "you're gonna work this job now". JES is simply picking up the top most job in a jobclass queue and hands it over to the initiator asking for work. What is different from the above (assuming it is correct) when the initiators are WLM-managed? It is WLM instead of the operator or automation package which decides when and where to start and stop initiators for the WLM-managed jobclasses. That's it. Once initators are up and running, WLM is out of control which one will get which job to work on. It is still the initiator that is going to ask JES for work. And it is still JES that will pick up the top most job from the jobclass queue. Silence is ambiguous for me. It could mean nobody knows the truth, or it could mean this is the truth :-) So, I'd very much appreciate a confirmation if my understanding is correct. And, of course, please let me know where I'm wrong. Thanks. -- Peter Hunkeler ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
