On 6/10/2013 2:46 PM, Farley, Peter x23353 wrote:
To your point about tailoring and dynamically submitting JCL, it really is an issue. In a typical large z/OS shop today, dynamically tailoring and submitting JCL is only permitted for test environments and users. "Production" JCL is frozen and controlled and submitted only by the scheduler software, and there is no political possibility to dynamically adjust the parameters even if it is technically feasible.
The scheduler can be set up to submit the tailoring job just as easily as the job to be tailored. And a few critical production abends should take care of the political aspect.
There *are* non-theoretical solutions to "runaway" file output. The *ix system model of using "disk quotas" per "user" makes it entirely possible to imagine z/OS "application" users with "reasonable" disk quotas specific to the application (i.e., not by job but by suite of jobs). Not the best solution? Maybe not, but ISTM to be better than having to predict what each and every process (i.e., job and file) output volume will be.
I've worked for service bureaus that established just such quotas. My objections stand, as both the single job and a suite of jobs can still fail; the difference is the number of jobs/users impacted, not the principle.
And there may well be other process models out there different from anything I know or imagine. I don't claim to have an exclusive lock on ideas to replace what we have to deal with
Ditto. Gerhard Postpischil Bradford, Vermont ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
