On Fri, 26 Sep 2008 11:47:36 -0500, Tim Hare <[EMAIL PROTECTED]> wrote:
>Somewhere in our migration to z/OS 1.9 (pretty sure it wasn't this way on
>1.7), our IEFUJV exit started getting invoked for OMVS address spaces. Since
>this exit tries to enforce TIME= and account numbers on job cards, it fails our
>OMVS address spaces.
>
>We coded, in SMFPRM, SUBSYS(OMVS,NOEXITS) as a work-around. This
>solved the immediate problem but introduced another. Now, file predecessors
>in our job scheduling software don't work when the dataset is closed by FTP.
>
>The vendor confirms that they dynamically install IEFU83 for SYS, SYSJES2,
>and SYSTSO to monitor SMF records (14/15) to determine when a dataset has
>closed and mark the file predecessor complete.
>
>Coding SMFPRM to pick up the IEFU83 exit point for the OMVS: SUBSYS(OMVS
>(IEFU83) doesn't work because it seems to pick up IBM's IEFU83. We probably
>need to code the vendor's module name on an exit statement for
>SYSOMVS.IEFU83 and we're working on that, but I have had other questions:
Yes. Code IEFU83 as part of your SUBSYS(OMVS,EXIST(...)) and use a
PROGxx member as follows:
EXIT ADD EXITNAME(SYSOMVS.IEFU83) MODNAME(vendu83)
DSN(vendor.loadlib)
>
>1. If we let EXITS default at the SYS level, and we code EXITS at the
>SUBSYS level, SUBSYS seems to "inherit" the other exits when we do D
>SMF,O; is that correct?
No, it should only take what you code if you code
SUBSYS(OMVS,EXITS(exit1,exit2,exit3))
>
>2. If so, does that mean we need to code SYS(NOEXITS) and then SUBSYS
>(EXITS(IEFwhatever, IEFwhatever)) for each SUBSYS?
No. See above.
>
>3. Some doc seems to imply that when multiple IEFU83 exit modules are
>defined, they _all_ get invoked. Which one's return code is honored?
Depends... Check the FM:
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/IEA2E470/2.30.8?SHELF=IEA2BK71&DT=20060622045210#HDRRSPU83
>
>4. Our exits are named IEFxxxxx because we've had them for years, from
>before the EXIT statement existed. Is current "best practice" to give them
>installation-specific names and add EXIT statements in PROGxx for them?
Ther are many ways to skin a cat. Some folks do them as usermods.
Some load them via IEALPAxx into MLPA. What we do: Because
of a shared IBM sysres here but different exit requirements, we use
a SYSLIB LPALIB statement in PROGxx that has our local LPALIB ahead
of SYS1.LPALIB and put our SMF exits there. We use the standard
names so a PROGxx EXIT ADD is not needed. This is because exits
defined in SMFPRMxx get defined to the dynamic exits facility implicitly.
>
>Our ultimate goal is to make sure our IEFUJV doesn't get invoked for OMVS
>address spaces, but it's sure been confusing trying to figure out the hierarchy
>of the exits.
IBM recommends that IEFUSI doesn't get invoked for OMVS either. The
same thing has to be done... either no exits for OMVS or code SUBSYS for
OMVS in SMFPRMxx and leave out IEFUSI. This has been discussed in
the archives several times at least.
Hope this helps.
Regards,
Mark
--
Mark Zelden
Sr. Software and Systems Architect - z/OS Team Lead
Zurich North America / Farmers Insurance Group - ZFUS G-ITO
mailto:[EMAIL PROTECTED]
z/OS Systems Programming expert at http://expertanswercenter.techtarget.com/
Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
