Interrupts and checkpoint/restart
When programs run for an extended period of time, interruptions might halt processing before the end of a job. The checkpoint/restart functions of z/OS® let an interrupted program be restarted at the beginning of a job step or at a checkpoint that you have set.
Because the checkpoint/restart functions cause a lot of extra processing, use them only when you anticipate interruptions caused by machine malfunctions, input or output errors, or intentional operator intervention.
The checkpoint routine starts from the COBOL program
object that contains your program. While your program is running,
the checkpoint routine creates records at points that you have designated
using the COBOL RERUN
clause. A checkpoint record
contains a snapshot of the information in the registers and main storage
when the program reached the checkpoint.
The restart routine restarts an interrupted program. You can perform a restart at any time after the program was interrupted: either immediately (automatic restart), or later (deferred restart).
Setting checkpoints
Restarting programs
Resubmitting jobs for restart
z/OS DFSMS: Checkpoint/Restart