S522 — the job step waited too long

What happened

A step ended with ABEND=S522. It used almost no CPU: the job log shows it started, then nothing until the abend, often about the installation's wait limit later.

03.08.07 JOB00101  +BRPOST: POST THE BRANCH FILE NOW? REPLY Y OR N
03.38.07 JOB00101  -BRPOSTRN RUN      BRPOST   *S522

What it means

All the tasks in the job step were in a wait for longer than the job wait time limit set by the JWT parameter in the SMFPRMxx parmlib member, so the system ended the step.

Typical causes

• The program waits for an operator reply that nobody gives, for example after ACCEPT ... FROM CONSOLE in a batch program, once the console is no longer watched.
• The step waits for something else that never comes, such as a resource held elsewhere or a device or volume that is not made available.

Where to look

• The job log: the last message before the gap, and the times.
• Outstanding replies and messages on the console for that job.
• The program: does it wait for an operator or another job?

How to fix it

1. Remove the reason for the wait: in unattended batch, take answers from the job's own input (SYSIN or a control file) instead of the console, or let the scheduler hold the job for an approval.
2. Rerun once the cause is fixed.
3. TIME=NOLIMIT or TIME=1440 lets a step wait for an unlimited time. Use it only for a step that is expected to wait; on the JOB statement TIME=1440 also overrides every step's TIME.

Interview question

A nightly job started ending S522 the week the night shift was automated. What is the likely cause and the fix?