Copying IBM-MAIN this time > -----Original Message----- > From: Neil Duffee [mailto:[email protected]] > Sent: Wednesday, July 10, 2013 12:46 PM > To: Gibney, Dave > Subject: RE: PS-E not releasing with HSM Migrate/Recall > > Ah, well. It was a flash in the pan suggestion since we don't HSM here. The > only other twinge I can think of is whether secondaries exist since it's > Conditional vs. Yes. Are there more than one extent on each of the volumes? > > hmmm... just looked at what you cut/pasted and I think that answer is "there > are no secondaries". Allocated extents: 2 c/w VolSer: + indicates to me there > are probably two volumes with each having one primary extent. That would > mean they don't qualify for Conditional Space Management. I'll bet you get > equivalent results with Conditional Immediate. You might try YES to see if > you have better luck. > > If I remember correctly, Extended causes a primary extent to be allocated on > each of 'Unit count' volumes and, as you fill the file, secondaries *might* be > allocated followed by additional volumes if you have DynVolCount set. We > also don't do Extended yet; we're not a big enough shop for it to benefit.
x37 abends are almost a non-existent with stripped/extended as a default vis DATACLAS Yes, it is stripped on 2 volumes, and it did not require allocation of any secondary extents. " a nonzero secondary space allocation has been specified for the data set" But, it is still desirable (since secondary is defined) to reduce the actual disk usage to the current utilization on recall. I haven't looked, but I am pretty certain that all those empty blocks aren't taking up space on my virtual tape. " a nonzero secondary space allocation has been specified for the data set" > > Aside: is migrate/recall how Space Management is achieved by HSM? Just > curious. Yes > > --------> signature = 6 lines follows <-------- > Neil Duffee, Joe Sysprog, uOttawa, Ottawa, Ont, Canada > telephone:1 613 562 5800 x4585 fax:1 613 562 5161 > mailto:NDuffee of uOttawa.ca http:/ /aix1.uOttawa.ca/ ~nduffee > "How *do* you plan for something like that?" Guardian Bob, Reboot > "For every action, there is an equal and opposite criticism." > "Systems Programming: Guilty, until proven innocent" John Norgauer 2004 > > > -----Original Message----- > From: Gibney, Dave [mailto:[email protected]] > Sent: July 10, 2013 15:27 > To: Neil Duffee > Subject: RE: PS-E not releasing with HSM Migrate/Recall > > I run Primary (and Secondary) Space management daily. The datasets have > secondary specified: > General Data Current Allocation > Management class . . : ATMBKUPS Allocated megabytes : 3,500 > Storage class . . . : SCDSKTAP Allocated extents . : 2 > Volume serial . . . : PTAP08 + > Device type . . . . : 3390 > Data class . . . . . : TMBKUPLG > Organization . . . : PS Current Utilization > Record format . . . : VB Used megabytes . . : 78 > Record length . . . : 2100 Used extents . . . : 1 > Block size . . . . : 32760 > 1st extent megabytes: 1750 > Secondary megabytes : 2500 Dates > Data set name type : EXTENDED Creation date . . . : 2013/07/10 > SMS Compressible. . : YES Referenced date . . : 2013/07/10 > Expiration date . . : ***None*** > > The Current utilization remains the same after any flavor of migrate/recall > > > -----Original Message----- > > From: Neil Duffee [mailto:[email protected]] > > Sent: Wednesday, July 10, 2013 12:16 PM > > To: Gibney, Dave > > Cc: IBM-Main > > Subject: RE: PS-E not releasing with HSM Migrate/Recall > > > > Dave: I get the daily digest so this might already have been pointed out... > > > > We don't HSM here (we're an Innovation shop) but is the "HSM > > migration/recall" process you talk about the same as "Space Management > > cycle" from the manual? I would suspect that HSM preserves the actual > > dataset size(s) during migration/recall unless you specify otherwise > > ie. via parms. For FDRABR, it's DATA={ALL | USED}. > > > > z/OS V1R12.0 DFSMSdfp Storage Administration: > > Conditional. If a nonzero secondary space allocation has been > > specified for the data set, release unused space at Space Management > > cycle time. This option only applies to physical sequential and > > partitioned data sets. If specified for an extended format VSAM data > > set, this option is processed as if Yes were specified. > > > > --------> signature = 6 lines follows <-------- > > Neil Duffee, Joe Sysprog, uOttawa, Ottawa, Ont, Canada > > telephone:1 613 562 5800 x4585 fax:1 613 562 5161 > > mailto:NDuffee of uOttawa.ca http:/ /aix1.uOttawa.ca/ ~nduffee > > "How *do* you plan for something like that?" Guardian Bob, Reboot > > "For every action, there is an equal and opposite criticism." > > "Systems Programming: Guilty, until proven innocent" John Norgauer > > 2004 > > > > -----Original Message----- > > From: Gibney, Dave [mailto:[email protected]] > > Sent: July 8, 2013 15:29 > > Subject: PS-E not releasing with HSM Migrate/Recall > > > > Hi everyone, > > I may be all wet, but this is my best resource for help on this > > issue until I get back on support with z/OS 1.13. > > I am running z/OS 1.11 (unsupported). Back in January, I applied all > > maintenance to bring it up to RSU1301. > > For years, I have had a MGMTCLAS of ATMBKUPS with conditional partial > > release and a DATACLAS of TMBKUPLG: > > [snip] > > > > I know that the multi-volume striping will not do the partial release > > at close, but I thought that HSM migration/recall was resulting in the > > release of unused space in these datasets. > > Last week, I had several failures to allocate these datasets and > > noticed that even after recall, these datasets remain at 65616 tracks > > usually about 5% used. > > > > Have I been fooling myself that this large (never run out while > > actually writing the SMF data) and release at migration ever worked? I > > suspect IBM will not be real helpful until I can reproduce on z/OS 1.13 > > which > will be later this year. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
