When a CA 11 controlled job abends and is restarted in a step subsequent to the abended step, it may be necessary to modify the last return code (LRTCD) of the step that abended to prevent subsequent steps from flushing. This occurs because the successor step never process through the JES control blocks and CA 11 is not able to make a determination about the status of the successor step. The LRTCD parm only need to be used if the restart step is a successor step that is dependent on the return code of the failed step.
The example below illustrate JOBA abended in step5 and want to restart in step6 instead. Normally, CA 11 would automatically select STEP5 for restart.
To restart the job in STEP6, you must override the START step with STEP6 and code STEP5 as the abended step. By default, =0 is used to null the return code of the fail step. However, a non-zero value can be used, causing CA 11 to honor its return code for the successor step.
-------- CA-7 QUEUE MAINTENANCE - CPU JOB IN RESTART STATUS ------
JOB: jobname CODE: JCLERR LAST-STEP: MCNT: 001
CA-7# 0001 JES#: 17349 NODE-NAME: LOCAL
REASON:
-- -- RESUBMIT FOR PRODUCTION
-- -- FORCE COMPLETE
-- x -- CA-11 RESTART/RERUN PSEUDO:
START: STEP6 END:
CC: BYPGDG: USAGE: LRTCD: STEP5 = 0 CMT STATUS: CMT SHOWS JOB IS SET FOR RESTART
-- -- SET PARM DATA FOR RMS AND RESUBMIT
PARM:
-- -- DO NOT INSERT RMS PROC BUT RESUBMIT
PROGRAM: QM50 MSG-INDX: 00 -- QM.4-X -- yy.ddd / 15:12:05
MESSAGE: 'X' THE DESIRED FUNCTION OR ENTER A COMMAND ON THE TOP LINE