On Thu, 17 Mar 2022 15:43:10 +0400, Peter <dbajava...@gmail.com> wrote:
>Hello, > >Could someone please direct me on which SMF record can produce the report >showing the highest CPU consumers who caused the memory shortage ? > >Today in one of our LPARs address space was recursively abending and it >monopolized the memory due to which TCPIP couldnt Dispatch the unit of work >and the Connectivity went down for a while. > >After canceling the abending Address space the LPAR started functioning >back and it was evident that the memory got relieved once the address space >was cancelled. > >Is there any SMF records which can tell me the memory consumption of the >address space which made the other Address space not to dispatch the UOW ? >This report basically will help me to show to my management people who dont >understand mainframe. > >Regards, >Peter > You can analyze/report on z/OS address space (interval) activity / consumption from the SMF type 30 subtypes 2 for long-running job-steps/task-steps (those active longer than your SMF 30 recording interval defined in PARMLIB SMFPRMxx). This condition occurred recently at a site I support (second time in less than 2 months), where the obvious symptom was in the SYSLOG with a CICS STORAGE VIOLATION - that precipitated SMF 110 subtype 1 (CICS transaction) record flooding for an extended period (up to an hour) - so seeing the spike in CPU-consumption (looping transaction activity) was the first symptom and also there are various REGION/MEMORY metrics in the same data source. Scott Barry SBBTech LLC ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN