That's the debate I'm having. The CF is only to support VSAM RLS and most of the time only one member of the sysplex will be up so the load on the CF is not going to be high.
I just can't persuade the outsourcer we don't need a dedicated engine for the CF. On Mon, Mar 6, 2023, 9:49 AM Seymour J Metz <[email protected]> wrote: > The issue is performance. Depending on what you are doing, the degraded > performance may be acceptable. > > > -- > Shmuel (Seymour J.) Metz > http://mason.gmu.edu/~smetz3 > > ________________________________________ > From: IBM Mainframe Discussion List [[email protected]] on behalf > of Laurence Chiu [[email protected]] > Sent: Sunday, March 5, 2023 2:05 PM > To: [email protected] > Subject: Re: Running a Coupling Facility using a CP for a test Parallel > Sysplex 0 anyh gotcha's? > > The debate I am having with the outsourcer is whether or not it's feasible > or even practical to run a test CF on a general purpose engine. They say it > requires a dedicated engine and I think that is nonsense! > > If I can get that over the line (and that is the challenge) then I can > suggest what you recommended below. And since I don't need one CF, I might > need even fewer resources. > > > On Sun, Mar 5, 2023 at 5:39 PM Mike Schwab <[email protected]> > wrote: > > > To avoid impacting other systems, I would drop your LPAR weights by a > > total of 6%, trim your LPAR memory to reuse for the ICFs, then create > > your two ICF partitions with 3% of 1 CPU and the reclaimed memory.. > > > > On Sat, Mar 4, 2023 at 2:55 PM Laurence Chiu <[email protected]> wrote: > > > > > > Thanks for the input. > > > > > > > > > On Sun, Mar 5, 2023 at 7:41 AM Mike Schwab <[email protected]> > > wrote: > > > > > > > > > > > > > > https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.researchgate.net%2Fpublication%2F342570694_Coupling_Facility_Configuration_Options_-_Updated_2020&data=05%7C01%7Csmetz3%40gmu.edu%7C436800a18ba54430093908db1daca96d%7C9e857255df574c47a0c00546460380cb%7C0%7C0%7C638136399655620282%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=HYoo%2FosMEZw5JHKVlB0T%2F2llYz5vokvdUSyxKpIB3Do%3D&reserved=0 > > > > > > > > > > I am familiar with that document and even provided a copy to our > > outsourcer > > > to read but clearly they hadn't > > > > > > This is a direct link to IBM for that document. > > > > > > > https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ibm.com%2Fdownloads%2Fcas%2FJZB2E38Q&data=05%7C01%7Csmetz3%40gmu.edu%7C436800a18ba54430093908db1daca96d%7C9e857255df574c47a0c00546460380cb%7C0%7C0%7C638136399655776514%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=tnP0qr0DTsTupn4v2ayadx5%2FmP%2BKlg8RXncytv1PUiA%3D&reserved=0 > > > > > > > > > > > > > CF is not counted on SCRT, shown on RMF reports. > > > > Won't cost you on z/OS, may on some vendors. > > > > > > > > > > I don't care so much as this CF is only System B and for development so > > > using a general purpose engine is not an issue for us. The MSU charges > > are > > > going to be low and we are prepared to pay those if it gets us our > > parallel > > > sysplex > > > > > > > > > > > > > Thin CFs go to enabled wait when work is completed, restart when > > > > interrupt says there is work. > > > > > > > > Estimate is 3% light sharing to 13% heaving sharing (of z/OS > workload). > > > > > > > > Thin CF would use internal links so no I/O overhead to another CPU. > > > > > > > > For the testing CF on the same system as the test Sysplex that is > fine. > > > But they say there are no spare links from System B to System A if I > > wanted > > > to run a test Sysplex on System B and access a CF on System A. > > > > > > This is their response I had to manage > > > > > > > > > To give an idea of what I am facing, this is their response to my > > proposals. > > > > > > Using a General Purpose CP (GCP) as a coupling facility on System > B(z13 > > at > > > WithDrawn From Marketing Licensed Internal Code) > > > • There are no spare unallocated GCP on System B i.e no “parked” GCP. > > > • All GCP’s, on System B, are allocated as shared, across all LPARS. > i.e. > > > no dedicated GCP’s. > > > • Sharing GCP’s to use for z/OS and as a coupling facility is strongly > > not > > > recommended FYI coupling facility engines run CFCC (coupling facility > > > control code) rather than z/OS. > > > • This possibly I believe is now exhausted. > > > > > > I think all these points are contestable, specially after reading the > IBM > > > document from a specialist in this area > > > > > > I just need to get some authoritative voice onto the case, ideally the > > > author of the document but that might not be easy. > > > > > > ---------------------------------------------------------------------- > > > 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 > > > > ---------------------------------------------------------------------- > 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 > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
