<disposition>
──┬─ LOCK ──┬──────────────────────────────────────────────────────────┤ └─ PURGE ─┘
The disposition option PURGE causes all checkpoint files to be removed at the successful termination of the job and protects the job against system failures. The LOCK option causes all checkpoint files to be saved indefinitely and can be used to restart a job even if it has terminated normally.
When a checkpoint is invoked, the following files are created:
-
The checkpoint file, CP/<JN>/<CPN>, where <JN> is a four- or five-digit job number and <CPN> is a three-digit checkpoint number
If the PURGE option has been specified, the checkpoint number is always zero, and each succeeding checkpoint with PURGE removes the previous file. If the LOCK option is used, the checkpoint number starts with a value of 1 for the first checkpoint and is incremented by 1 for each succeeding checkpoint with LOCK. If the two types are mixed within a job, the LOCK checkpoints use the ascending numbers and the PURGE checkpoints use 0 (zero), leaving files 0 through N at the completion of the job.
-
Temporary files, CP/<JN>/T<FN>, where <FN> is a three-digit file number beginning with 1 and incremented by 1 for each temporary disk file
-
The job file, CP/<JN>/JOBFILE
This file is created under the LOCK option only.
The LOCK and PURGE options are also effective when the task terminates. If the task terminates abnormally and the last checkpoint has used the PURGE option, then the checkpoint file (numbered zero) is changed to have the next sequential checkpoint number, and the job file is created (if necessary). If the job terminates normally and only PURGE checkpoints have been taken, the CP/<JN> directory is removed.
