In our older environment (no real sysplex), each system has its own master catalog, but the root file system is part of sysres volume which is actually shared between systems. We don't ship new copies of the root file system, instead we IPL with a new version of sysres each month. The root file system is not in the master catalog, it is in a ucat on sysres.
In our new environment (all sysplex), the 'IBM root' file system is still on a shared sysres (in a ucat on sysres), but it has the name of the sysres in the dataset name (using &SYSR1). Again, we never replace the file system by itself, we replace sysres every month. The real (sysplex) root file system is shared across the plex. Bart -----Original Message----- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Bonno, Tuco Sent: Friday, October 26, 2012 7:13 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: maintain, distribute multiple copies same-named ZFS files operating environment here is all z/os 1.13 how am i suppsed to clone and DISTRIBUTE (with emphasis on *distribute*) zfs files? CURRENTLY, I have a situation where 6 lpars, 3 production and 3 test, all share the same master catalog. In that master catalog, the entry for OMVS.ROOT reads as follows, (i.a.,) : "volser ---- &hfsv1" ; and there are 6 3390s out there, one for each lpar, each of which has a copy of OMVS.ROOT on it. Currently, each such omvs.root is an HFS file. Whenever any lpar is ipled, the value of &hfsv1 is set bmo of an entry in an appropriate IEASYM-- in an appropriate PARMLIB, and each lpar runs with its own copy of the root, called .... OMVS.ROOT . whenever i have a NEW COPY of omvs.root that needs to be distributed, i (1) drain the 3 test lpars; (2) go to the pack in each test lpar whcih contains omvs.root and do a "DELETE/NOSCRATCH" for the omvs.root on that pack ; (3) from a 7th lpar, i copy the new version of omvs.root to the pack in each test lpar which is supposed to have a copy of omvs.root on it, using a DISP=(NEW,KEEP). Later on, those 3 "test" lpars are ipled as production lpars. in this methodology, the entry in the master catalog is never touched, and nothing in any of the 3 production lpars is ever impacted by whatever i may be doing in any test lpar. (This g.p. methodology is how i maintain ALL the o/s image dsn-s) (and, btw, i use this methodology to distribute a couple DOZEN omvs/unix-system-services dsn-s, not just omvs.root) the recent emphasis from ibm is, has been, to convert one's HFS files to ZFS. so, now, given that zfs files are really vsam files, (1) how does one maintain multiple copies of zfs files off of one master catalog, given that you can't catalog vsam files using symbolic values in the catalog entry ; (2) next, assuming that there is some method -- unknown to me at the present moment -- to maintain multiple copes of identically-named vsam files off of one catalog , how do i go about distributing new copies of zfs files, using my methodology that I described above? First of all, if I follow the above-described scenario, the first time i try to delete a zfs file (a vsam file, now) , isn't that going to also erase the catalog entry for it (and wreak some havoc for the remaining 5 lpars, which are also still using the catalog) ? Second of all, there is/are the VVDS entries, something I need not bother myself about in my current methodology (b/c in that, no vsam files are involved) -- how do i keep them sync-ed up? (3) i tried to consult several books about this , i.a., "z/os distributed file service file system implementation" , "distributed file sevice zSeries file system Administration" and " volume 9, ABCs of system programming" and "z/FS reorganization tool" , and i just can't get this to work , at least in the context of the methodology i currently use to distribute omvs/unix-system-services dsn-s. Maybe i need to totally change my methodology? EXACTLY how are other people doing this kind of thing? TIA /s/ tuco bonno; Graduate, College of Conflict Management; University of SouthEast Asia; "I partied on the Ho Chi Minh Trail - tiến lên !! " ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN