Overallocate PDS directories! Or change IPL/NIP processing so that everhing can be PDSE.
I had to walk to school. There were sidewalks all the way, so walking a mile or so was no problem. My reading of the "good old days" before my time is that they were characterized by: High infant mortality Mob violence Rampant discrimination ... As for my early years, my first computer was a 650, a nostalgia killer if ever there was one. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 ________________________________________ From: IBM Mainframe Discussion List <IBM-MAIN@listserv.ua.edu> on behalf of John Eells <ee...@us.ibm.com> Sent: Friday, October 5, 2018 6:45 AM To: IBM-MAIN@listserv.ua.edu Subject: Re: S106 abends after copying into LINKLIST Jesse 1 Robinson wrote: > I sympathize with IBM's predicament in reading the future maintenance tea > leaves. The 'fix rate' for a product might be subject to guesstimation from > past experience. But the effects of future enhancements like SPEs and > customer requirements are a bundle of uncertainties wrapped in unknowns. The > change from six month to two year release cycles further clouded space > predictions. > > I think customers are best off expecting data set expansions. A few like > LINKLIB and LPALIB can be deliberately oversized as likely candidates for > increase in a variety of components. But the migratable sysres volume is > limited to whatever size installation has settled on. Secondary extents, in > my view, allow for unpredictable expansion with acceptable risk. > While we have long provided some cushion in the initial numbers and in ServerPac (and before it, CBIPO and even IPO) allocations, we have really left further guessing to you. I expect that we will err on the side of more free space pretty soon to help alleviate out of space problems in this new(er) era of larger system software volumes, particularly because system software is such a small fraction of the disk space requirements for nearly any shop out there. In the long term I think everyone is better off moving all their software volumes to -54s, doubling or even tripling all the primary space allocations, and using thin provisioning to manage space at the volume level rather than the data set level. It will be approximately true that only occupied space takes up actual disk real estate when you do that. (I say "approximately" because there is an increment size, and on the average every data set will have an extra half-increment.) System software data set level space management and x37 abends during APPLY and ACCEPT processing will, I would hope, become a fading memory in a few years. Unlike some other memories, nobody will miss the "good old days." This will be more like the stories about how much more complicated life used to be when you had to walk to school. It was always uphill both ways and it was always cold and snowing. At least, that's what people used to tell their kids who rode those cushy heated (FSVO "heated," at least in Maine) buses, right? -- John Eells IBM Poughkeepsie ee...@us.ibm.com ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
