Resolving restart abends caused by indoubt checkpoints

A restart of an abended IMS batch job ends abnormally if the last checkpoint ID that was supplied by IMS Program Restart Facility is indoubt. If such an abend occurs, determine the correct checkpoint ID and restart option for the job.

About this task

Indoubt checkpoints most often occur for BMPs that are active at the time of an IMS control region abend or for canceled DL/I batch jobs. The condition is caused when the next checkpoint ID that is requested by the application is not confirmed to IMS Program Restart Facility, presumably because the logging request did not complete.

If a restart abend from an indoubt checkpoint occurs, determine if you must restart the original abended job by using the last indoubt checkpoint or the last verified checkpoint. You can also restart the job from the beginning.

Indoubt checkpoints are flagged by messages IRT010W and IRT017I.

Procedure

If a restart of an abended job ends abnormally with a PRF U3625 or an IMS U0102 completion code, review the IMS logs and message DFS682I. With this information, select one of these actions:

  • If the restart checkpoint ID that is displayed in message DFS682I matches the checkpoint ID that is displayed in message IRT010W, restart the job by using the last indoubt checkpoint ID:
    1. Access the IMS Program Restart Facility ISPF main menu.
    2. To access a list of abended jobs, Select option 1, Administer abended jobs.
    3. In the ACTION column, enter FORCE next to the name of the job that you want to restart.
    4. Restart the job.
      The FORCE option affects only this one restart.
      Tip: You can also specify the FORCE option with a temporary JCL override by adding a IRT$CNTL DD * statement followed by the AUTOXRST=FORCE control statement.
  • If the restart checkpoint ID that is displayed in message DFS682I matches the checkpoint ID that is displayed in message IRT017I, restart the job by using the last verified checkpoint ID:
    1. Access the IMS Program Restart Facility ISPF main menu.
    2. To access a list of abended jobs, Select option 1, Administer abended jobs.
    3. In the ACTION column, enter LAST next to the name of the job that you want to restart.
    4. Restart the job.
      The LAST option affects only this one restart.
      Tip: You can also specify the LAST option with a temporary JCL override by adding a IRT$CNTL DD * statement followed by the AUTOXRST=LAST control statement.
  • If the job must be restarted from the beginning, restart the job by doing these steps:
    1. To start the IMS Program Restart Facility ISPF interface, run REXX EXEC IRTXISPF.
    2. To access a list of abended jobs, Select option 2, Administer abended jobs.
    3. In the ACTION column, enter NOXRST next to the name of the job that you want to restart.
    4. Restart the job.
      The NOXRST option affects only this one restart.
      Tip: You can also specify the NOXRST option with a temporary JCL override by adding a IRT$CNTL DD * statement followed by the AUTOXRST=NO control statement.
  • Alternatively, you can specify USEJCLID=YES and a checkpoint ID that is optimal for the job restart by using the CKPTID parameter in the JCL EXEC statement.
    To restart a job by using the USEJCLID option, you must make available the log data set that contains the corresponding checkpoint log by specifying an IMSLOGR DD statement.