Len, First, thanks for your help. You are correct, the task that was mentioned, is a web server, but, We don't use this for CICS Web frontend. The problem, sound like, the users do something, that acess the DB2 directly, and sometime, We can see one, two, three user's doing it. In this moment, the task of Web Server have more use of CPU% like show on D A command, and in this time, our command DA show CPU with 100% of utilization, and start problem to CICS transaction. I'm not expert under WLM, but will try speak with this people here. Thanks very much Sergio "Rugen, Len" <[EMAIL PROTECTED]> escreveu: There are no easy answers. Priority doesn't matter as much as resource consumption. If the high priority task isn't using resources, the it isn't the problem.
However, without checking, I think that is a web server. Is it a CICS web frontend? If so, then it is "essential" to your CICS transactions. Is it an efficient frontend? We can't answer that. Several years ago, even though we had some CICS response time problems and not enough capacity, I converted to WLM and response time goals. I set some pretty aggressive goals, again without checking, I think 95% of all transactions under .5 seconds. Even then, we were meeting those goals. We have 80% trivial transactions and 20% that are resource intensive. Everything is the same transid, so I can't determine before hand which transaztions are "loved ones" and which aren't. My actual dispatching priority of the CICS regions went DOWN under WLM, but the goals were being met. When the huge transactions run for 60-90 seconds so be it. I presnt the cost of the upgrades needed to resolve the issue and let the business case support itself. I have limited resources, WLM lets me set goals and importance. If management says "everything is important", then they are supporting unlimited upgrades. They can't have it both ways. Management always thought CPU was the solution. Even our huge, 90 second transactions used a trivial amount of CPU. I argued that even if the CPU component was 0 they wouldn't like the response time. The solution? Better DASD, (Shark ESS), more RAM for IMS DBCTL buffers and offloading TN3270 SSL. Sounds like you have one of those "opportunities" that can be painful. ________________________________ From: IBM Mainframe Discussion List on behalf of Sergio Lima Sent: Mon 9/10/2007 7:41 PM To: [email protected] Subject: CICS Priority Question Hello, Last week, We have some problems here with our environment. We have here a lot of transaction ONline, and have a STC here that have more priority than CICS. We don't understand very well, but this STC execute a BPXTCAFF, and a IMWHTTPD program, and looks like a OMVS System. This STC have a high priority under CICS, but the question are, It's really necessary ? We understand, that CICS must have a good priority, and for this reason, We need know, if someone know about this, and if no problem, give to CICS more priority. Thanks for now ! Sergio Lima Costa Caixa Economica Federal Sao Paulo - Brazil Flickr agora em português. Você clica, todo mundo vê. Saiba mais. ---------------------------------------------------------------------- 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 ---------------------------------------------------------------------- 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 Flickr agora em português. Você clica, todo mundo vê. Saiba mais. ---------------------------------------------------------------------- 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
