On Mon, 25 Mar 2013 14:00:49 -0500, Paul Gilmartin wrote:

>On Mon, 25 Mar 2013 14:38:41 -0400, Steve Thompson wrote:
>>
>>Now, go whack the person that wrote the JCL that has the following
>>similarly coded DD statement:
>>
>>//SYSUT3  DD  UNIT=3390,SPACE=(CYL,10),DISP=(NEW,CATLG)

Has this ever been allowed?  ISTR that only PASS or DELETE are 
valid for a data set with a temporary dsname.  The event that most 
often causes these data sets to be left behind is a system crash or 
initiator abend.  When the program abends, the system would 
normally delete them at step end or job end.  In any case, CATLG 
would fail on most systems because there would be no alias and 
most shops restrict the use of the master catalog.


>>This is what produces a data set name like this:
>>SYS12048.T104505.RA000.IBMUSER.R0100008
>>
>>The DISP is NOT needed for a "step only" work file / data set. So during
>>DEALLOCATION at step end, this type of data set just goes away.
>>
>Indeed.  And for good measure, I'd whack the z/OS (...OS/360?) designer
>who decided that the proper behavior in this case is to KEEP but not
>CATLG the data set.  

The design was that data sets would be cataloged upon successful 
(or unsuccessful) termination, not at initiation.  It's kind of hard for 
the system to do that upon a system crash.

>I'd have made CATLG without DSNAME a JCL syntax
>error and never initiated the job.

ISTR that if KEEP or CATLG is specified, a message is issued that says
INVALID DISP.  PASS SUBSTITUTED.

-- 
Tom Marchant

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to