Handling application program situations

This section describes how to handle non-standard application program situations. These situations include restart of a job step defined to AR/CTL as non-unique, cold start of an application program defined as needing restart by AR/CTL, and restart at other than the checkpoint AR/CTL has selected. AR/CTL automatically handles application program restart in standard situations, which include any application program error that does not require database recovery and a cold start of the application program.

While you correct the problem that caused the abend, determine whether the job step can be restarted or must be cold-started:

  • You can restart a job step if all processing before the latest checkpoint is correct.
  • You may need to recover the databases and/or sequential data sets and then perform a cold start of a job step if the database updates are incorrect or if the application program updated the database but never issued a checkpoint. Cold-starting a job step without database recovery is valid if the application program made no database updates or the database updates can be repeated without database corruption.

Note

If the application program abends without issuing a restart call and/or a checkpoint call, AR/CTL cannot restart the job step. If the application program issues database update calls before issuing the restart and/or checkpoint call, and the job step abends before the call, AR/CTL cannot restart the job step. The job step must be handled as a cold start, or the restart must be handled outside of AR/CTL. You can use the Check XRST, Check CHKP, and Check IMS Updates options to prevent job steps from executing without issuing restart and/or checkpoint calls correctly. For more information, see Reference.

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*