Fw: Sort capacity exceeded didn't ABEND job
William M. Klein
My guess is exactly the OPPOSITE of Frank's. My guess is that the
programmer has ADDED a reference to the COBOL "SORT-Return" special register
- and that it is this (new) reference that causes the program to continue
execution. See:
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/igy3lr10/1.2.8.13
which says (in part)
"SORT-RETURN special register ... It contains a return code of 0
(successful) or 16 (unsuccessful) at the completion of a sort/merge
operation. If the sort/merge is unsuccessful and there is no reference to
this special register anywhere in the program, a message is displayed on the
terminal."
The logic behind this is that *IF* the program makes an explicit reference
to SORT-RETURN, then it is EXPECTED that the program "wants" to take
specific "error" processing logic. If, on the other hand, there is NO
reference to SORT-RETURN, then it is assumed that the program wants to take
SORT's "default" (i.e. ABEND) action.
"Frank Yaeger" <yae...@US.IBM.COM> wrote in message
news:<OF8D1E75F4.4EEE39E8-ON88256D...@us.ibm.com>.
..
> Greg Shirey wrote:
>
> >We had a batch job running an internal COBOL sort, and it put out this
> >message:
> >
> >ICE046A 0 SORT CAPACITY EXCEEDED - RECORD COUNT 3858551
> >
> >The job returned a condition code of 0 in this step and kept running, but
> >the results weren't very good.
> >
> >The fine manual states that the program will terminate if this message is
> >generated, which would have been preferable... Is there a parameter in
> >DFSORT that kept the job from abending?
>
> For ICE046A, DFSORT either terminates with RC=16 (if NOABEND is in effect)
> or abends (if ABEND is in effect). DFSORT passes the RC=16 back to the
> COBOL program. I would guess that the COBOL program is NOT checking the
RC
> from DFSORT, and is just passing back RC=0 instead of passing back RC=16.
>
> Frank Yaeger - DFSORT Team (IBM)
> Specialties: ICETOOL, OUTFIL, Symbols, Migration
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to list...@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
Clark F. Morris, Jr.
2 default settings, one for JCL invoked sorts and one for program
invoked sorts. When I was in charge of installation of sort, JCL was
set to abend while program invoked sorts got the return code of 16.
Since SORT-RETURN is a COBOL function the handling referred to by the
book below is a COBOL handling and is from a COBOL manual this action
should be the same for both major sort packages. Also note that the
issuing of a message doesn't imply abend after issuance. See below for
history if you haven't been following the thread.
Clark Morris, CFM Technical Programming Services Inc.
Reda, John
The installation option you are referring to are RC16=ABE or NOABE. This
controls whether a user abend 16 or a return code of 16 will be issued for
an unsuccessful JCL invoked sort. The PGMRC16 option controls this
processing for program invoked sorts. As you had said this allows a site to
have different default options for JCL and programmed invoked sorts. Many
of our customers have their options set to allow JCL invoked sort to return
with a RC16 while their programmed invoked sorts will end with a user abend
16.
John Reda
Software Services Manager
Syncsort Inc.
201-930-8260
-----Original Message-----
From: Clark F. Morris, Jr. [mailto:cfm...@istar.ca]
Sent: Friday, June 06, 2003 7:47 AM
To: IBM-...@BAMA.UA.EDU
Subject: Re: Fw: Sort capacity exceeded didn't ABEND job
On the SORT-RETURN question, SYNCSORT at least up until release 3.6 had
2 default settings, one for JCL invoked sorts and one for program
invoked sorts. When I was in charge of installation of sort, JCL was
set to abend while program invoked sorts got the return code of 16.
Since SORT-RETURN is a COBOL function the handling referred to by the
book below is a COBOL handling and is from a COBOL manual this action
should be the same for both major sort packages. Also note that the
issuing of a message doesn't imply abend after issuance. See below for
history if you haven't been following the thread.
Clark Morris, CFM Technical Programming Services Inc.
----------------------------------------------------------------------
Frank Yaeger
>On the SORT-RETURN question, SYNCSORT at least up until release 3.6 had
>2 default settings, one for JCL invoked sorts and one for program
>invoked sorts.
For the record, you can set 8 separate sets of installation defaults with
DFSORT as follows:
1. JCL - batch direct invocation defaults
2. INV - batch program invocation defaults
3. TSO - TSO direct invocation defaults
4. TSOINV - TSO program invocation defaults
5-8. TD1-TD4 - time-of-day defaults
You can set ERET=RC16 or ERET=ABEND (or any other ICEMAC option) as
appropriate for each of these.
For complete details, see Chapter 5 of "DFSORT Installation and
Customization" at:
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/ICECI107/2.2?DT=20020718103509
Frank Yaeger - DFSORT Team (IBM)
Specialties: ICETOOL, OUTFIL, Symbols, Migration
=> DFSORT/MVS is on the Web at http://www.ibm.com/storage/dfsort