It might well interest y'all to know that I will be putting a slide on SRB into my revamped "zIIP Capacity and Performance" presentation. (I intend to give this presentation several times this year around the world. So votes for this for eg SHARE Boston would be appreciated.)
I had an hour-long conversation with Dave Surman (who leads the SRB effort) and John Campbell (for Db2) among others yesterday on just this subject. Two thoughts - to start with :-) : 1) SRB is on by default. You might well turn it off if you don't want/need it - eg for test LPARs where SRB isn't part of the test. 2) It DOES matter if LPAR A gobbles up zIIP - such that LPAR B gets constrained for it (FSVO "constrained"). OK, a third: 3) You might want to assess - eg with SMF 30 - how you choreograph shut down and restart of an LPAR. OK, a fourth: 4) Your automation for shutting down might want spiffying up. Possibly around the SRB messages, possibly not. And a fifth: 5) At a certain maintenance level SMF headers for RMF tell you if this LPAR has SRB on and which kind. (For all 7x.) And in SMF 70-1 the same is true for ALL LPARs on the machine. I think that makes at least one slide. :-) Outline of the above discussed at length on that call. And a parting thought: When my team sees a crop of z15 machines we'll note whether SRB is on for each LPAR and how well we think it's used round any IPLs we see. Cheers, Martin Martin Packer zChampion, Systems Investigator & Performance Troubleshooter, IBM +44-7802-245-584 email: [email protected] Twitter / Facebook IDs: MartinPacker Blog: https://www.ibm.com/developerworks/mydeveloperworks/blogs/MartinPacker Podcast Series (With Marna Walle): https://developer.ibm.com/tv/mpt/ or https://itunes.apple.com/gb/podcast/mainframe-performance-topics/id1127943573?mt=2 Youtube channel: https://www.youtube.com/channel/UCu_65HaYgksbF6Q8SQ4oOvA From: Mike Schwab <[email protected]> To: [email protected] Date: 07/02/2020 00:27 Subject: Re: z15 System Recover Boost [EXTERNAL] Sent by: IBM Mainframe Discussion List <[email protected]> Any additional IPL / Recovery BOOST processing power is over and above normal processing. It will not reduce the processing power on other LPARs. Unless you DON'T have any spare processors (or CP capacity). On Thu, Feb 6, 2020 at 4:30 PM Feller, Paul <[email protected]> wrote: > > The final word on the zIIP thing is going to be our capacity person. If that person is okay with the situation then I'm okay with it. From where I sit I don't want any more overhead for zIIP management then needed. We only have 3 zIIP on the box. 10 of the lpars get 1 zIIP and the 11th lpar gets 2 zIIP. That 11th lpar is an important lpar. Yes we have the weights set for the zIIP that the tech lpars are bottom on the list. > > If you are suggesting that we config offline the 2nd zIIP before BOOST is done, my answer would I don't know what would happen. That would be a question for IBM to answer. That is why we are looking at testing BOOST=SPEED to see how that reacts. That's why we all have the tech lpars in our environments to try things out before we roll to production. > > > Thanks.. > > Paul Feller > AGT Mainframe Technical Support > > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Tony Harminc > Sent: Thursday, February 06, 2020 3:42 PM > To: [email protected] > Subject: Re: z15 System Recover Boost [EXTERNAL] > > On Thu, 6 Feb 2020 at 16:31, Feller, Paul <[email protected]> wrote: > > > Well we have our z15 in and running. We have been testing the System > > Recovery Boost on our tech lpars. What we have found makes us think > > we will be setting the option to BOOST=SPEED for our tech lpars. I > > normally define each lpar with a minimum of 1 zIIP with 1 zIIP in > > reserve. With BOOST=SYSTEM the 2nd zIIP is getting put online and it stays online for an hour. We really don't want to do that on our tech lpars. We may also look at this related to the other lpars as we roll this out to see what happens. > > What's the down side to having that zIIP online to that lpar? Is some other work losing out because (presumably lower importance) "tech" > work is being done on the zIIP? > > > Just daydreaming here, but it would be nice if there was a time limit > > you could set for how long BOOST was active (unless I missed > > something). As I stated before BOOST was active for an hour on our tech lpars and they really don't need that BOOST for that long. > > Could you not vary it offline at the right time using existing automation? > > Tony H. > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] 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 [email protected] with the message: INFO IBM-MAIN
