Scheduler Non-CPU jobs failed to Autostart

Products

• Scheduler Job Management

Issue/Introduction

Non-CPU jobs (jobs with Station number less than 40) that are defined with ‘Automatic Start = YES’ did not have ‘START’ automatically posted to the job when its early start time is reached and all the predecessor conditions have been satisfied.  This happened when Scheduler abended shortly after AUTOSCAN completed.

Cause

The Scheduler Started Task was operator-canceled (S222) shortly after the abend, and therefore Scheduler did not have a chance to process its records recovery routine.

Sample scenario: 

06.00.57  CACM051I AUTOSCAN COMPLETED

06.01.00 MUFPROD:DB00311I - LOG AREA IS 100% FULL, SPILLING TRACK NONE, …

               (During this time, Scheduler cannot communicate with Datacom MUF due Log Area full)

06.38.45  CACM999I CA SCHEDULER ENF SUBTASK ABENDED    

06.38.45  CACM956E CA SCHEDULER READER SUBTASK ABENDED

06.38.45 IEF450I CASCHD CASCHD - ABEND=S222 U0000 REASON=00000000

Resolution

Do not cancel the Scheduler Started Task.  Scheduler must be allowed to go down normally so the events are saved in the recovery records.  When Scheduler restarts, the events that have not been processed when Scheduler shuts down are recovered.  

Non-CPU jobs with Autostart Yes do not have ‘START’ posted during AUTOSCAN.  This type of jobs is placed in the Scheduler Recovery Queue (SRQ) to be processed after AUTOSCAN completed.  In this case, the SRQ could not be created because the Scheduler started task was cancelled.