home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!USCMVSA.BITNET!LDW
- Message-ID: <IBM-MAIN%92072923410645@RICEVM1.RICE.EDU>
- Newsgroups: bit.listserv.ibm-main
- Date: Wed, 29 Jul 1992 21:37:00 PDT
- Sender: IBM Mainframe Discussion list <IBM-MAIN@RICEVM1.BITNET>
- From: Leonard D Woren <LDW@USCMVSA.BITNET>
- Subject: Re: UNIT and VOLUME specification
- Lines: 73
-
- On Wed, 29 Jul 1992 12:25:37 EDT, BENSON@POKVMCR3.VNET.IBM.COM said:
- > Leonard, in answer to you comments - (a) SMS is our strategic
- > direction and our desire is that everyone use it, (b) I can't
- > speak for the future but that is certainly true today, (c) no.
-
- Then you have to fix it so that sysres can be SMS managed. This means
- replacing the SYS% conversion silliness with something that can
- actually be used. How many shops have *every* dataset on sysres named
- SYS1.something??? Until CBIPO ships the system that way, *I* never
- will. And I don't think you want to limit things that way anyway.
- A more general solution is needed. I'd suggest some prefixing method
- to specify the volser. This limits the dsnames to less than the full
- 44 characters, but this shouldn't be a problem for sysres datasets.
- For example, let me define an index level, say, altres, such that any
- reference to altres.SYS1.LINKLIB means SYS1.LINKLIB on ALTRES.
- (Actually, what would be really nice would be the ability to specify
- any volser via syntax like "ALTRES:SYS1.LINKLIB". I won't hold my
- breath...)
-
- Until sysres can be cleanly SMS managed, and until *all* other IBM
- utilities can work with SMS volumes, IBM cannot reasonably state that
- SMS is the strategic direction. IBM finally fixed the problem of
- IDCAMS BUILDIX requiring its internally-created IDC work files to be
- on non-SMS by answering an old unrelated SHARE requirement -- allow
- external sort to be used. This was actually a good solution for that
- particular case. But IBM's answer to complaints that IEHMOVE doesn't
- work on SMS volumes of "IEHMOVE is obsolete, use IEBCOPY and DFDSS" is
- inappropriate for one glaring reason -- users will have IEHMOVE-
- unloaded datasets for a long time, and they need to be able to reload
- them. IEHMOVE could be fixed rather easily -- THROW AWAY the silly
- code that does the manipulation of its work datasets via SVC 32, and
- either allocate them with SVC 99, or (best) allow them to be specified
- in the JCL. Also, change the creation of the output dataset from SVC
- 32 to SVC 99. If these reasonably simple changes were made, many
- advantages would result. IEHMOVE could run without APF authorization.
- Performance would be improved, since VIO work files could be used.
- And unloaded datasets from obsolete device types could be reloaded to
- VIO and then converted over to current dasd. Have I submitted a SHARE
- requirement? No. Why bother, when IBM has made it clear that they
- don't understand, and have no intention of fixing IEHMOVE? If you can
- assure me that the requirement will be accepted when submitted, and
- targetted for near-term delivery, then I'll submit it. Otherwise, why
- should I waste my time?
-
-
- On Wed, 29 Jul 1992 13:32:08 EDT,
- Jerry Bryan <BRYAN@WVNVM.WVNET.EDU> said:
- > >Specifically, if ALTRES (for example) is permanently resident, why
- > >do the following fail without UNIT=nnnn?
-
- > (...)
- > I think the simple answer is not quite so simple. The real reason
- > (at least historically under OS/360) is that ALTRES might be a tape.
-
- Irrelevant. Note that I specifically was referring to PERMANENTLY
- RESIDENT disks.
-
- > For example, vary ALTRES offline. Should the JCL be parsed
- > any differently while ALTRES is offline? If ALTRES were offline,
- > how on earth could you figure out what the proper device type was?
-
- Well, someone else already commented on the JCL parsing issue. I
- assume you just meant "should it mean something different if ALTRES is
- offline?" Yes! It's quite simple. VOL=SER= without a UNIT= should
- work only for permanently resident dasd. If you reference a
- non-mounted disk volume without UNIT=, the job should fail just as it
- does now if you omit the UNIT=. All I'm asking for is that UNIT=
- should be implicitly assigned by MVS if the VOL=SER= refers to a
- permanently resident dasd. Quite simple, if the designers had any
- imagination. In fact, this is so simple that a JES2 internal text
- exit could be easily written to do exactly that! Hmmm...
-
- /Leonard
-