On Wed, 27 Feb 2019 23:36:59 +0000, Savor, Thomas (Alpharetta) wrote:
>
>... To fix this, it required us to split many jobs in half, some even
>more...the run total got to be somewhere around 6500+ jobs in total. But we
>had to build dataset create jobs that were parm cards to a utility to create
>the files, because if JES3 scanned the JCL and couldn't find it...boom JCL
>ERROR.
>
At some point, I arrived at a paradigm such as:
//HANDLE DD DISP=(MOD,CATLG),UNIT=SYSALLDA,SPACE=(...),DSN=...
//SYSUT2 DD DISP=SHR,DSN=*.HANDLE,VOL=REF=*.HANDLE
... to either use an existing data set or create a new one. No need for
utility and parm.
JES2. Never tried it on JES3. Could be done in 2 separate steps, but why
bother?
>So, I was scarred because "I" saw this restriction as pretty stupid...and
>never saw the point.
>
I suspect that many problems such as yours arise from JES3 setup's adamantly
acting as if all steps are executed regardless of COND or IF or DYNALLOC. There
ought to be an alternative.
I'm not much impressed by any argument that a design that doesn't know at
each point whether or not a given data set exists presumes a defective process.
>To me JES2 is and was exactly what I believe it was designed for a tool to
>support and run applications very efficiently, whereas I felt that JES3 was
>getting in the way of normal application jobs. Seemed like we were constantly
>trying to please JES3 in some way...so I hope and pray this "standard" wont
>make it to JES2.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
