On Fri, 28 Mar 2008 10:39:52 -0400, Knutson, Sam <[EMAIL PROTECTED]> wrote:
>I don't think anyone has mentioned the white paper from Riaz Ahmad (IBM) >WSC that detailed a customer situation with problems dealing with >current volumes using MANx data sets and some WSC testing that emulated >it. > >http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP101130 > >or http://tinyurl.com/2gnojb > >z/OS System Management Facilities (SMF) Recording with MVS Logger > >WP101130 > >Abstract: The SMF (System Management Facilities) Recording to MVS Logger >describes the z/OS 1.9 facility which allows recording of the SMF data >to MVS Logger Logstream. The paper documents a case study for a customer >environment which has difficult time keeping up with the SMF data >recording to MANx data sets. This new facility provides a solution where >SMF data can be recorded to the MVS Logger logstream instead of MANx >data sets. > Thanks. The paper says this about the customer situation: "During the peak hours of this interactive workload the customer has experienced long intervals when SMF data was produced at a rate exceeding the capacity to offload. In reality, there are limits placed on the offload process by the customers SMF data collection design, which involves collecting data offloaded from the SMF MANx datasets into a single daily collection dataset. Each offload is appended to the end of the collection dataset (using DISP=MOD processing) and this effectively limits the offload processes to dump the MANx datasets one at a time, due to the enqueue on the daily collection dataset." This supports what I have been saying about being able to keep up if you use IEFU29, run dumps to FICON DASD / virtual tape and run your dump STC in a high enough service class (SYSSTC if required). Other than a problem situation (looping transaction etc.) does anyone know of any shops where data is written so fast that the SMF address space (writes to MANx) can't keep up as opposed to the offload process not being able to keep up? Kees mentioned it and I have seen it happen, but only as a result of a problem - not a "normal" thing. I know the logger can handle much higher logging rates. Mark -- Mark Zelden Sr. Software and Systems Architect - z/OS Team Lead Zurich North America / Farmers Insurance Group - ZFUS G-ITO mailto:[EMAIL PROTECTED] z/OS Systems Programming expert at http://expertanswercenter.techtarget.com/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.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
