Thanks! That's a much crisper version of what I had roughly surmised. That's an interesting case in the blog post. I wonder if bash has been / could be fixed. (Not that I run it on z/OS or write loops in it on Pi.)
Cheers, Martin Martin Packer WW z/OS Performance, Capacity and Architecture, IBM Technology Sales +44-7802-245-584 email: martin_pac...@uk.ibm.com Twitter / Facebook IDs: MartinPacker Blog: https://mainframeperformancetopics.com Mainframe, Performance, Topics Podcast Series (With Marna Walle): https://anchor.fm/marna-walle Youtube channel: https://www.youtube.com/channel/UCu_65HaYgksbF6Q8SQ4oOvA From: "Andrew Rowley" <and...@blackhillsoftware.com> To: IBM-MAIN@LISTSERV.UA.EDU Date: 17/09/2021 09:26 Subject: [EXTERNAL] Re: RMF SMF "broken" records Sent by: "IBM Mainframe Discussion List" <IBM-MAIN@LISTSERV.UA.EDU> On 17/09/2021 5:32 pm, Martin Packer wrote: > Off topic but are there usable start and stop timestamps for things running > in BPXAS address spaces? > > I've wondered about z/OS Unix substeps and thingies but not had cause to > dig too deep. (I know where I am with conventional batch jobs and steps > when it comes to SMF - but I don't see many jobs with Unix in them.) In many cases a whole step/substep SMF record is written, which has the start and stop times the same as any other type 30. However it is also possible to have many processes run in one address space, in which case you get multiple process sections without timestamps. It is usually possible to relate them by process ID. Where a process has a parent process, the parent might be in the same or another address space. The SMF record for the parent process might be produced much later than the record for the child. Process IDs can also be reused. I have written code to do it but it was not easy. You can see what the results look like here: https://www.blackhillsoftware.com/news/2019/08/27/comparing-bash-and-bin-sh-on-z-os/ That was one job that resulted in 20,000 type 30 records in 1 minute! It would be nice if unix processes used the same service/report class settings and charged their resources back to the parent job when the userid/session/process group are the same. For a unix job, a lot of work can be in the additional address spaces e.g. this script took 66 seconds of CPU time, but only 0.02s showed up in the batch job. -- Andrew Rowley Black Hill Software ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN