Automatic restart

If the programmer submitting the job has provided for an automatic restart and the job ends abnormally, you receive the following system message:

* id IEF225D SHOULD jobname.stepname.procedure checkid RESTART

This message allows you to prevent repeated restarts at the same checkpoint or job step.

When this message appears, use the REPLY command to respond YES, HOLD, or NO, as follows:

Reply YES if the restart is to be performed at a specific checkpoint or job step for the first time. (If it is a job step restart and the step to be restarted used a card input data set that was not part of the SYSIN stream, you must return to the appropriate hoppers all cards read by the job step before it ended abnormally. If it is a checkpoint restart, follow the programmer's instructions for replacing the input cards.)
Reply HOLD if you want to defer the restart: for example, to permit another job to run first. You must issue the appropriate subsystem command when you are ready to restart the job. Also, if you want, you can cancel the job. However, cancelling the job can cause unrecoverable paging space or the failure of certain data sets to be deleted if the job was using virtual I/O.
Reply NO if a restart at a specific checkpoint or job step has been requested repeatedly. When your reply is NO, and the programmer wants a restart to be performed, he must resubmit the job for a deferred restart.

If the programmer specifies VIRTUAL=REAL (V=R), the job is processed entirely in central storage; it is not paged out. For a V=R job, the restart might be delayed while the system waits for the allocation of storage. If another job is using the required storage, you get no message, only a delay. Enter the DISPLAY A,L command to see if a system task or another job is using the storage required by the job with a V=R region. You can then stop or cancel the conflicting task or job.

Note: Any operator commands in the input stream of the job step being restarted are not executed.