Performance report help
We have a mystery that needs to be solved, and I am having trouble coming up with a report that would help me do it. Here are the particulars along with a couple of questions: Our CPU usage had been trending upward -- as it generally does -- for several weeks, when we finally started hitting the wall hard during first shift. CICS transaction times were starting to elongate, batch jobs were taking much, much longer to run than usual, etc. We went a couple of weeks trying to mitigate things as much as possible, when suddenly, for no reason that we could detect, system usage dropped one day and we have been OK ever since. Naturally, upon inquiring we found that no one had done anything. Right. Anyway, I am trying to create a report detailing CPU usage by job for several of our heavy hitters during the days prior to the big drop. We have both RMF and TMON, but I am having little luck with either of them. For RMF, it appears that SMF type 79 records might contain what I need, but those appear to only be cut when RMF II is turned on, and we do not have RMF II activated on a regular basis. For TMON, it appears that JD (Job Delta, IIRC) records would have what we need, but, again, they are not generated on a regular basis. Questions: 1) Are there other SMF records that could tell me what I want to know? 2) Are there other TMON records that could tell me? 3) For those of you running RMF, do you have RMF II sessions running at all times? If so, does it consume a lot of overhead? 4) For those of you running TMON, do you cut JD records as a usual thing? If so, how much does it add to your TMON overhead? Thanks, Jon -- 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
Re: Performance report help
Take a look at the type 30 records. CBT file 529 (among others) has a program which can print a report from the type 30 data. On Fri, 9 May 2008 11:06:12 -0400, Jon Brock [EMAIL PROTECTED] wrote: We have a mystery that needs to be solved, and I am having trouble coming up with a report that would help me do it. Here are the particulars along with a couple of questions: Our CPU usage had been trending upward -- as it generally does -- for several weeks, when we finally started hitting the wall hard during first shift. CICS transaction times were starting to elongate, batch jobs were taking much, much longer to run than usual, etc. We went a couple of weeks trying to mitigate things as much as possible, when suddenly, for no reason that we could detect, system usage dropped one day and we have been OK ever since. Naturally, upon inquiring we found that no one had done anything. Right. Anyway, I am trying to create a report detailing CPU usage by job for several of our heavy hitters during the days prior to the big drop. We have both RMF and TMON, but I am having little luck with either of them. For RMF, it appears that SMF type 79 records might contain what I need, but those appear to only be cut when RMF II is turned on, and we do not have RMF II activated on a regular basis. For TMON, it appears that JD (Job Delta, IIRC) records would have what we need, but, again, they are not generated on a regular basis. Questions: 1) Are there other SMF records that could tell me what I want to know? 2) Are there other TMON records that could tell me? 3) For those of you running RMF, do you have RMF II sessions running at all times? If so, does it consume a lot of overhead? 4) For those of you running TMON, do you cut JD records as a usual thing? If so, how much does it add to your TMON overhead? Thanks, Jon -- 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
Re: Performance report help
When I've had to determine what job(s) have caused a large CPU increase over an interval I've used the SMF type 30 interval records. If you have Barry Merrill's MXG the information from them is in the SMFINTRV dataset. By using them you can limit the analysis to a short interval rather than over the full time of the job. Of course if you talking about several days of data it would be a lot of information. Tom Kelman Commerce Bank of Kansas City (816) 760-7632 -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Jon Brock Sent: Friday, May 09, 2008 10:06 AM To: IBM-MAIN@BAMA.UA.EDU Subject: Performance report help We have a mystery that needs to be solved, and I am having trouble coming up with a report that would help me do it. Here are the particulars along with a couple of questions: Our CPU usage had been trending upward -- as it generally does -- for several weeks, when we finally started hitting the wall hard during first shift. CICS transaction times were starting to elongate, batch jobs were taking much, much longer to run than usual, etc. We went a couple of weeks trying to mitigate things as much as possible, when suddenly, for no reason that we could detect, system usage dropped one day and we have been OK ever since. Naturally, upon inquiring we found that no one had done anything. Right. Anyway, I am trying to create a report detailing CPU usage by job for several of our heavy hitters during the days prior to the big drop. We have both RMF and TMON, but I am having little luck with either of them. For RMF, it appears that SMF type 79 records might contain what I need, but those appear to only be cut when RMF II is turned on, and we do not have RMF II activated on a regular basis. For TMON, it appears that JD (Job Delta, IIRC) records would have what we need, but, again, they are not generated on a regular basis. Questions: 1) Are there other SMF records that could tell me what I want to know? 2) Are there other TMON records that could tell me? 3) For those of you running RMF, do you have RMF II sessions running at all times? If so, does it consume a lot of overhead? 4) For those of you running TMON, do you cut JD records as a usual thing? If so, how much does it add to your TMON overhead? Thanks, Jon -- 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 * If you wish to communicate securely with Commerce Bank and its affiliates, you must log into your account under Online Services at http://www.commercebank.com or use the Commerce Bank Secure Email Message Center at https://securemail.commercebank.com NOTICE: This electronic mail message and any attached files are confidential. The information is exclusively for the use of the individual or entity intended as the recipient. If you are not the intended recipient, any use, copying, printing, reviewing, retention, disclosure, distribution or forwarding of the message or any attached file is not authorized and is strictly prohibited. If you have received this electronic mail message in error, please advise the sender by reply electronic mail immediately and permanently delete the original transmission, any attachments and any copies of this message from your computer system. * -- 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
Re: Performance report help
snip I am trying to create a report detailing CPU usage by job for several of our heavy hitters during the days prior to the big drop. snip If you don't have MXG or SAS, you can get some rough SMF 4,5,30 info from Mr. Yaeger's do anything to anything ICEMAN, eg http://bama.ua.edu/cgi-bin/wa?A2=ind0507L=ibm-mainP=R15258I=1X=4FA88C76F3FC5F13F7 (LIST archive) Jack Kelly 202-502-2390 (Office) -- 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
Re: Performance report help
--snip--- We have a mystery that needs to be solved, and I am having trouble coming up with a report that would help me do it. Here are the particulars along with a couple of questions: Our CPU usage had been trending upward -- as it generally does -- for several weeks, when we finally started hitting the wall hard during first shift. CICS transaction times were starting to elongate, batch jobs were taking much, much longer to run than usual, etc. We went a couple of weeks trying to mitigate things as much as possible, when suddenly, for no reason that we could detect, system usage dropped one day and we have been OK ever since. Naturally, upon inquiring we found that no one had done anything. Right. Anyway, I am trying to create a report detailing CPU usage by job for several of our heavy hitters during the days prior to the big drop. We have both RMF and TMON, but I am having little luck with either of them. For RMF, it appears that SMF type 79 records might contain what I need, but those appear to only be cut when RMF II is turned on, and we do not have RMF II activated on a regular basis. For TMON, it appears that JD (Job Delta, IIRC) records would have what we need, but, again, they are not generated on a regular basis. Questions: 1) Are there other SMF records that could tell me what I want to know? 2) Are there other TMON records that could tell me? 3) For those of you running RMF, do you have RMF II sessions running at all times? If so, does it consume a lot of overhead? 4) For those of you running TMON, do you cut JD records as a usual thing? If so, how much does it add to your TMON overhead? unsnip-- You really need to look at the SMF-30 records, at a job-step level. IIRC, Berry Merrill's package can help you in the analysis. I'll look around to see if I have anything that might be helpful and forward it to you off-list. You might need a pl/1 compiler (IEL00AA). -- 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
Re: Performance report help
- Original Message - From: Jon Brock [EMAIL PROTECTED] Newsgroups: bit.listserv.ibm-main Sent: Friday, May 09, 2008 11:08 AM Subject: Performance report help We have a mystery that needs to be solved, and I am having trouble coming up with a report that would help me do it. Here are the particulars along with a couple of questions: Our CPU usage had been trending upward -- as it generally does -- for several weeks, when we finally started hitting the wall hard during first shift. CICS transaction times were starting to elongate, batch jobs were taking much, much longer to run than usual, etc. We went a couple of weeks trying to mitigate things as much as possible, when suddenly, for no reason that we could detect, system usage dropped one day and we have been OK ever since. Naturally, upon inquiring we found that no one had done anything. Right. Anyway, I am trying to create a report detailing CPU usage by job for several of our heavy hitters during the days prior to the big drop. We have both RMF and TMON, but I am having little luck with either of them. For RMF, it appears that SMF type 79 records might contain what I need, but those appear to only be cut when RMF II is turned on, and we do not have RMF II activated on a regular basis. For TMON, it appears that JD (Job Delta, IIRC) records would have what we need, but, again, they are not generated on a regular basis. Questions: 1) Are there other SMF records that could tell me what I want to know? 2) Are there other TMON records that could tell me? 3) For those of you running RMF, do you have RMF II sessions running at all times? If so, does it consume a lot of overhead? 4) For those of you running TMON, do you cut JD records as a usual thing? If so, how much does it add to your TMON overhead? Jon, FWIW, I always run RMFMON II. The overhead is minimal compared to when I use it, which is often. YMMV, Tom Conley -- 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
Re: Performance report help
Questions: 1) Are there other SMF records that could tell me what I want to know? Yes. Job/Step Accounting -- SMF Type 30. If you have/had interval accounting turned on. - Too busy driving to stop for gas! -- 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
Re: Performance report help
Take a look at the type 30 records. Ah, yes. That would be the section in the book on my desk with a sticky-label bookmark . . . that says, Type 30 CPU. In my own handwriting. Heavy sigh Perhaps I should just give up now. CBT file 529 (among others) has a program which can print a report from the type 30 data. OK, thanks for that tip. I have downloaded and compiled it and have it running now. I'll see what it gives me. Thanks, Jon -- 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
Re: Performance report help
It has been a long time since I looked at it, but I don't think we can use MXG. We don't have SAS, and isn't that an MXG requirement? Thanks, Jon snip When I've had to determine what job(s) have caused a large CPU increase over an interval I've used the SMF type 30 interval records. If you have Barry Merrill's MXG the information from them is in the SMFINTRV dataset. By using them you can limit the analysis to a short interval rather than over the full time of the job. Of course if you talking about several days of data it would be a lot of information. /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
Re: Performance report help
Thanks for that information, Tom. I think I might start running it on a full-time basis. I'll need to check the parms first, though. Jon snip Jon, FWIW, I always run RMFMON II. The overhead is minimal compared to when I use it, which is often. YMMV, Tom Conley /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
Re: Performance report help
OK, thanks for that. I'm having vague memories (the only kind I have any more) about going through this exercise before. I can't find any of the jobs I had worked up, though; it may be that I didn't get any useful information at the time. I'll see what I get, though. Thanks, Jon snip Yes. Job/Step Accounting -- SMF Type 30. If you have/had interval accounting turned on. /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
Re: Performance report help
Yes, SAS is a pre-req for MXG. Have you looked at the RMF Spreadsheet Reporter? It's a bit high-level for what I think you need, but the process of acquiring the data, getting it into the spreadsheets and preparing the charts may give you some insight on how to take a left turn at Albuquerque and perhaps roll your own to get the info you need. Earlier, I think I read a post by another forum member to look at DFSORT to produce something. That is a very good idea. Gary Green I can use all the help I can get with my fight against cancer! Please support my efforts! Thank you. http://www.active.com/donate/tntsonj/tntsonjGGreen -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Jon Brock Sent: Friday, May 09, 2008 3:10 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Performance report help It has been a long time since I looked at it, but I don't think we can use MXG. We don't have SAS, and isn't that an MXG requirement? Thanks, Jon snip When I've had to determine what job(s) have caused a large CPU increase over an interval I've used the SMF type 30 interval records. If you have Barry Merrill's MXG the information from them is in the SMFINTRV dataset. By using them you can limit the analysis to a short interval rather than over the full time of the job. Of course if you talking about several days of data it would be a lot of information. /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 No virus found in this incoming message. Checked by AVG. Version: 8.0.100 / Virus Database: 269.23.14/1425 - Release Date: 5/9/2008 12:38 PM -- 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
