Hello Gil,
In the production environments that I am involved in, most TSO users are
not permitted to SUBMIT jobs.
The people schedulers have TSO access to the job scheduler product, and
maybe also to SDSF.
Batch jobs have been created(developed) under a rigorous development regime
that is managed by a source management product. The jobs have (supposedly)
been tested and peer approved before they are promoted to the libraries in
production.
The job scheduler (CA-7 or the successor to OPC) is a product which manages
job streams. Job submission is dependent on triggers... like time of day,
successful completion of one or more jobs/jobsteps, or some other event.
Submitting via the job scheduler also ensures the job is run under the
correct userid, ensuring correct access to resources.
The job scheduler monitors the progress of the job, highlighting if it
takes longer than "normal", and captures the aggregate return code / abend
code upon completion.
All nicely known and managed.
There is no need to have a job submit (INTRDR) another job, as that
introduces a wild card. One consequence is unhappy auditors. :-)
In regard to John's comments regarding additional documentation required,
just as the JCL for the STEP exists, then the relevent documentation
describing that step should also exist. Extra work then is:-
1. extract step and add jobcard.
2. remove step from old job.
3. extract step documentation to new document.
4. remove step documentation from old job document.
5. add new job to schedule, dependent on original job.
6. change existing dependencies from old job to new.
7. convince application teams that doing this once per ftp step will stop
many 0dark30 phone calls.
On Mon, 6 Mar 2006 08:41:56 -0700, Paul Gilmartin <[EMAIL PROTECTED]>
wrote:
<snip>
>Hmmm. Clearly I work in a development environment, not a
>production environment, so I'm curious about protocols.
>
>What's a "job scheduler" Is it made of silicon or carbon?
>I thought that nowadays almost all jobs (barring those actually
>submitted on physical, necrodendritic cards) go through an
>INTRDR; it's simply a matter of how they get there. Are
>programmers in a production environment likewise discouraged
>from using the TSO SUBMIT command (which, AFAIK, also uses
>an INTRDR)? How do jobs get submitted? Must a human
>bureaucrat ("job scheduler") sign off on each one?
>
>If the process is in fact automated, can't one job submit
>another through the automated sanctioned channel, as opposed
>to via INTRDR?
>
>-- gil
>--
>StorageTek
>INFORMATION made POWERFUL
>
Regards
Bruce Hewson
----------------------------------------------------------------------
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
