1. The number of concurrent long-running tasks You have this estimate
2. Each terminal running conversational tasks This may not be a large number if your transaction design is pseudo-conversational. EXEC CICS CONVERSE would cause a transaction to be conversational. With the large amount of memory potentially available on modern systems, this may not be as big a concern. Make an educated guess, it probably won't bite you too hard. 3. An estimate of the number of concurrent tasks from terminals running nonconversational tasks How many terminals do you have available to your users? What is their daily workload? How many will be logged on at the same time? Without knowing your environment, very hard to estimate. Do you have CICS shut-down statistics reports?For our development region MXT=85 and production is the same. We only get to 10% of MXT on our processors (10 CICS regions). 4. An estimate of the number of concurrent nonterminal tasks. Our CS** tasks number 15 (CSSY etc) in an active CICS region. HTH On Mon, Oct 27, 2014 at 3:15 PM, Alex Wang <szwan...@gmail.com> wrote: > Dear All, > > First of all, please correct me I'm wrong to post this topic here. > > This is a question about how to set a proper CICS MXT value. > > CICS Performance guide shows me that: > > ----------------------------------------------------------------------------- > Initially, set MXT to the number of concurrent user tasks that you require > in your > system by totaling the following values: > 1. The number of concurrent long-running tasks > 2. Each terminal running conversational tasks > 3. An estimate of the number of concurrent tasks from terminals running > nonconversational tasks > 4. An estimate of the number of concurrent nonterminal tasks. > > ----------------------------------------------------------------------------- > I can estimate the number of concurrent long-running tasks. But for the > number > 2,3 and 4, I have no idea about how to estimate them. Could you please > show me > some lights on how to figuring them out? > > In fact, my major concern is how to evaluate a proper MXT value based on > current workload. > Is there any factors which could be used for this estimation? > > Many thanks for your help! > > Alex > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > -- Wayne V. Bickerdike ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
