Scott Chapman wrote: [ ... Lots of very interesting things snipped for brevitiy... ]
> If, for example, you have a batch job that spends most of it's time waiting > on I/O, a CPU discussion is mostly moot. (The RMF III DELAY panel can be > useful here.) If the job is spending most of it's time waiting on CPU, ... Another analogy, if you see several AS waiting for a volser (DEV Delay) in RMF III DELAY, you should look what is on that volser and move some datasets to other volsers. One bad example I handled in a previous assignment was: RACF DB (both primary and backup) and JES2 spool and JES2 cache plus a catalog for temp datasets were all on ONE volser. Of course, the fix was easy. Yes, my example is not about CPU, but you should get the drift. If your delay is something like this example message across SEVERAL intervals: 'Job awaiting service from JES.', then you should seriously look at your system. [1] I wish to add one note to the OP: For diagnosting of delays - You should use lots of different intervals so you can see the trend. Do not act on one interval or one LPAR only. You could run a batch job using RMF and SMF to see what trends are there. Another theoritical possibility: Perhaps one of your jobs is opening/closing/opening/closing datasets repeately in a loop and RACF+SMF are constantly writing gazillion records and pushing CPU% up to 100%. Then that is a problem as indicated by others who mentioned loops. So, IBM-MAIN members cannot diagnose your problem from a distance. >That's how I would go about things. Thanks Scott for your excellent post. Now waiting for your next post. :-D Groete / Greetings Elardus Engelbrecht [1] - I just see this message about 2 hours ago, but that was for ONE measuring interval of 100 seconds only. JES2 was busy purging about 35 000 jobs via several $TA commands and the rest of the system was somewhat idle CPU at 70%. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
