Mainframe, MVS and zOS Discussion

Hi,

A Job got abended because a dataset was not created in time. As the dataset created late, Job has taken yesterday's dataset and abended. Restarted the Job after the new dataset created but it failed due to GDG verification in CA-11 step. Now I resubmit the Job instead of restarting and this time Job went fine.

That makes me ask - what is the difference b/w restart and resubmit and why it failed during restart?

How did you do the "restart"? There is a job card parameter to restart a job at a particular step, and job schedulers (such as CA-7) also allow restart of jobs. If you used the job card method, for example, then you skipped the CA-11 analysis and set up, which could cause GDG issues.

Resubmit means you are running the job from the first step right on through the last step, so CA-11 analysis and set up would be performed.

Is not "restart" and "resubmit" totally different to each other? I'm not sure but can there be circumstances where in they can behave "just same"?

If the job failed for an allocation problem, then it never got started executing. In that case, restart and resubmit would be the same.

As I stated in my previous post, "restart" is an ambiguous term since it can refer to multiple things. Without knowing the type of "restart" being done, not much more can be said.

Robert Sample wrote:How did you do the "restart"? There is a job card parameter to restart a job at a particular step, and job schedulers (such as CA-7) also allow restart of jobs. If you used the job card method, for example, then you skipped the CA-11 analysis and set up, which could cause GDG issues.

I restarted the job from top - can that called as "resubmit"?

I would consider that a resubmit. Restart in the way I'm using the term would involve changing the JOB statement to include(assuming your job is using PROCs). Restart has implications for GDG processing, whereas resubmit will (especially if CA-11 or similar product is used) will handle the GDGs normally.

Thank you Robert.