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%92072718545419@RICEVM1.RICE.EDU>
- Newsgroups: bit.listserv.ibm-main
- Date: Mon, 27 Jul 1992 16:53:00 PDT
- Sender: IBM Mainframe Discussion list <IBM-MAIN@RICEVM1.BITNET>
- From: Leonard D Woren <LDW@USCMVSA.BITNET>
- Subject: Re: Esoteric Device Message in ESA 4.2
- Lines: 103
-
- On Mon, 27 Jul 1992 14:53:00 PDT,
- Dave Friedman 510-642-5823 <SPGDNF@UCBCMSA.BITNET> said:
- > (...)
- > If the dataset is cataloged via UNIT=DISK and DISP=(,CATLG) in jcl, the
- > catalog shows a DEVICETYPE of X'3010200F'. The same thing happens if we
-
- 3010200F is the UCB device type for 3390.
-
- > catalog the dataset using IDCAMS DEFINE ... NONVSAM DEVTYPE(3390), and
- > if we catalog the dataset with a WYLBUR command (apologies to Leonard).
-
- Whether it's JCL, IDCAMS, or Wylbur (a horse is, uh, oh never mind),
- it's the CATALOG SVC that's doing the real work, so getting the same
- value in the catalog entry is to be expected.
-
- > However, if we catalog the dataset using IDCAMS DEFINE ... NONVSAM
- > DEVTYPE(DISK), the catalog shows a DEVICETYPE of X'0003200'.
-
- However, in this particular case (the same would happen with IEHPROGM),
- the program has no way of knowing the correct device type. Duh, why
- not scan the online volumes to locate it? Because this logic hasn't
- changed since most disks were removable, and they had to make it work
- for non-mounted disks. An interesting exercise would be to try to get
- IBM to fix IDCAMS and IEHPROGM to look for an online volume in the
- specified unit group and pick up the correct device type. Good luck.
- (Why do I have to specify UNIT= *at all* if I specify VOL=SER= for an
- online disk??? That's one of my pet peeves with MVS.)
-
- > I guess the first question is why is IDCAMS using X'0003200'? However,
-
- This is documented somewhere, but I don't remember where. The X'20'
- in the third byte is the device type (dasd), the same as in 3010200F.
- The first halfword (0003) is the relative entry number in your
- UNITNAME macros of the UNITNAME DISK.
-
- > this is also true in our ESA 3.1.1 system, and it does not generate any
- > JES2 messages there.
-
- Must be a new feature, or maybe just a higher maintenance level.
-
- > IEF348I dsname WAS FOUND TO BE CATALOGED USING ESOTERIC DISK
- > INSTEAD OF A GENERIC
-
- Well, that's true.
-
- > The ESA 4.2 messages manual says that this is to warn the user in case
- > the system I/O configuration changes.
-
- Yep. If you change the *relative* position of the UNITNAME=DISK in
- your gen, all datasets cataloged using 00032000 will likely get
- catalog errors when access is attempted.
-
- > The manual seems to be advising
- > the systems programmer to recatalog the dataset using a generic.
-
- You should apar the manual (i.e., send in a reader comment form) if
- that's what it says. The person responsible for creating the dataset
- should fix it.
-
- > In our shop the systems programmer does not mess around with
- > application files.
-
- That's probably the norm.
-
- > Anyway, it would be a never-ending task, since we recommend to
- > our application staff that they use esoterics rather than generics.
-
- Esoterics are definitely preferable for creating new datasets. But
- for cataloging existing datasets, the actual device type should
- *always* be used, for exactly the above reason.
-
- > ... However, I wonder if anyone else has encountered this problem and
- > found a way to suppress the JES2 message.
-
- Maybe you could use console automation to intercept the message and
- recatalog the dataset correctly? Then you'd only get the message once
- per incorrect catalog entry, and there would be no ongoing manual
- effort involved in correcting them.
-
- > Is there some reason why JES2 should be giving people this "reminder"
- > every time they access a file cataloged with IDCAMS and an esoteric?
-
- Yep.
-
- > It seems like nagging.
-
- Yep. Using an esoteric in IDCAMS or IEHPROGM should be considered an
- error. This is IBM's way of convincing you to do it right, without
- actually making it fail.
-
- > Is this message a result of the IBM
- > orientation toward SMS? We are not yet an SMS shop.
-
- Either that or a response to constant problems caused by the UNITNAME
- positions changing. The manual says that the systems programmer can
- never move or delete UNITNAMEs once they're defined, because it'll
- cause problems. Most systems programmers don't know this. I've
- always ignored it, because I've always considered those catalog
- entries to be erroneous. When I first came here, I recataloged
- thousands of incorrectly cataloged datasets so that I could clean up
- the UNITNAMEs.
-
- /Leonard
-