Home DevOps AutoSys Job Status Codes Explained — Every Status and What It Means

AutoSys Job Status Codes Explained — Every Status and What It Means

Where developers are forged. · Structured learning · Free forever.
📍 Part of: AutoSys → Topic 18 of 30
Complete guide to AutoSys job status codes: RUNNING, SUCCESS, FAILURE, ON_HOLD, ON_ICE, PEND_MACH, ACTIVATED, RESTART, and more.
🧑‍💻 Beginner-friendly — no prior DevOps experience needed
In this tutorial, you'll learn:
  • AutoSys has 12+ job status codes — know the abbreviations used in autorep output (SU, FA, RU, OH, OI, PE, etc.)
  • PEND_MACH means the agent on the target machine is unavailable — check disk space first
  • FAILURE requires investigation before restarting — always check the error log first
✦ Plain-English analogy ✦ Real code with output ✦ Interview questions
⚡ Quick Answer
AutoSys job statuses are like package tracking codes — each status tells you exactly where your job is in its lifecycle. RUNNING means it's on the truck. SUCCESS means it was delivered. FAILURE means something went wrong. PEND_MACH means the truck broke down.

Every AutoSys job has a status at any given moment. Understanding what each status means is fundamental to monitoring and operating an AutoSys environment. Some statuses are normal progression; others indicate something needs your attention. This article covers every status code you'll encounter.

Normal lifecycle statuses

These are the statuses a job moves through during a healthy execution.

check_status.sh · BASH
12345678910111213
# Check status of a single job
autostatus -J daily_report

# Check status with full details
autorep -J daily_report -d

# List all jobs and their current status
autorep -J %

# Filter by status (SU = SUCCESS, FA = FAILURE, RU = RUNNING)
autorep -J % -s SU
autorep -J % -s FA
autorep -J % -s RU
▶ Output
Job Name ST Exit Start Time End Time
------------------------------------------------------------
daily_report SU 0 03/19/2026 02:00:01 03/19/2026 02:08:43
weekly_reconcile FA 1 03/19/2026 03:00:00 03/19/2026 03:01:22
monthly_summary OH -- -- --

All status codes with abbreviations

The autorep command uses two-letter abbreviations for status codes. Here they all are:

status_abbreviations.txt · BASH
123456789101112
AC  ACTIVATED    -- Box is RUNNING but this job hasn't started yet
FA  FAILURE      -- Job completed with non-zero exit code
IN  INACTIVE     -- Job hasn't been triggered; waiting for schedule
OH  ON_HOLD      -- Manually held; won't run until released
OI  ON_ICE       -- Suspended; won't run even when conditions reappear
PE  PEND_MACH    -- Waiting for target machine (agent offline)
QU  QUE_WAIT     -- Waiting for available machine load/resources
RE  RESTART      -- Job is being restarted after failure
RU  RUNNING      -- Job is currently executing
ST  STARTING     -- Job sent to agent; waiting for execution to begin
SU  SUCCESS      -- Job completed with exit code 0
TE  TERMINATED   -- Job was killed (term_run_time, KILLJOB event, etc.)

The problem statuses and what to do

These statuses almost always require your attention:

problem_statuses.sh · BASH
123456789101112131415161718192021
# ── FAILURE: job failed ──────────────────────────────────────────
# Check the error log first
cat /logs/autosys/daily_report.err
# Then check the event log
autorep -J daily_report -d
# Retry (force start) after fixing the underlying issue
sendevent -E FORCE_STARTJOB -J daily_report

# ── PEND_MACH: agent machine offline ─────────────────────────────
# Check the machine
autoping -m prod-server-01
autorep -M prod-server-01
# Check disk space on the agent machine (most common cause)
# ssh prod-server-01 'df -h'

# ── TERMINATED: job was killed ────────────────────────────────────
# Check if killed by term_run_time or manually
autorep -J slow_job -d | grep term
# If due to term_run_time, increase it:
update_job: slow_job
term_run_time: 180
▶ Output
FAILURE — Error in /logs/autosys/daily_report.err:
ERROR: Database connection timeout after 30s

PEND_MACH — prod-server-01: MISSING (agent service down)
Disk usage on prod-server-01: /apps 100% — agent stopped
⚠️
PEND_MACH in bulk usually means a disk is fullWhen you see 20+ jobs all showing PEND_MACH on the same machine at the same time, the most likely cause is that the agent machine's filesystem is 100% full. SSH into the machine, check df -h, and clear space before restarting the agent service.
StatusAbbrev.Normal or problem?Action needed?
INACTIVEINNormalNo — just not scheduled yet
ACTIVATEDACNormalNo — box is running, job will start when conditions met
STARTINGSTNormalNo — brief transition state
RUNNINGRUNormalNo (unless running longer than expected)
SUCCESSSUNormalNo
FAILUREFAProblemYes — investigate and fix
ON_HOLDOHNormal (if intentional)Check if intentional
ON_ICEOINormal (if intentional)Check if intentional
PEND_MACHPEProblemYes — check agent machine
QUE_WAITQUPossible problemCheck machine load
TERMINATEDTEProblem (usually)Investigate cause
RESTARTRENormal (auto-retry)Monitor for continued failure

🎯 Key Takeaways

  • AutoSys has 12+ job status codes — know the abbreviations used in autorep output (SU, FA, RU, OH, OI, PE, etc.)
  • PEND_MACH means the agent on the target machine is unavailable — check disk space first
  • FAILURE requires investigation before restarting — always check the error log first
  • ACTIVATED means a job is inside a running BOX but hasn't started yet — this is normal, not a problem

⚠ Common Mistakes to Avoid

  • Confusing ON_HOLD and ON_ICE — they sound similar but behave differently when released
  • Not checking PEND_MACH machine disk space first — it's the most common cause and quickest fix
  • Restarting jobs in FAILURE without first reading the error log — the underlying cause might make the restart fail again immediately
  • Not distinguishing INACTIVE from ON_ICE — INACTIVE just means not scheduled yet; it's not a problem

Interview Questions on This Topic

  • QWhat does PEND_MACH status mean in AutoSys?
  • QWhat is the difference between ON_HOLD and ON_ICE in AutoSys?
  • QWhat does ACTIVATED status mean for a job inside a BOX?
  • QHow do you force-start a job that is in FAILURE status?
  • QWhat is the most common cause of jobs going to PEND_MACH in bulk?

Frequently Asked Questions

What does PEND_MACH mean in AutoSys?

PEND_MACH (PE) means the job is waiting for its target machine to become available. The Remote Agent on the target machine is not responding — usually because the agent service has stopped (commonly due to a full disk), the machine is offline, or there's a network issue.

What does ACTIVATED mean in AutoSys?

ACTIVATED (AC) means the job's parent BOX is in RUNNING state, but the job itself hasn't started yet because its own conditions (start_times, condition attribute) haven't been met. It's a normal state — the job is queued inside a running box.

What is the difference between ON_HOLD and ON_ICE?

ON_HOLD: When you release a job from ON_HOLD (off_hold), it will run if its starting conditions are currently satisfied. ON_ICE: When you take a job off ICE (off_ice), it will NOT run even if conditions are met — it waits until its conditions reoccur in the next scheduling cycle. ON_ICE is a stronger suspension.

How do I restart a failed AutoSys job?

First check the error log to understand why it failed. Fix the underlying issue. Then use sendevent -E FORCE_STARTJOB -J jobname to restart it. Don't restart without investigating — you may be running a broken job again.

What does TERMINATED status mean in AutoSys?

TERMINATED (TE) means the job was killed. This can happen because term_run_time was exceeded (AutoSys killed it after max runtime), someone manually sent a KILLJOB event, or the agent machine went down while the job was running. Check autorep -d to see the details.

🔥
Naren Founder & Author

Developer and founder of TheCodeForge. I built this site because I was tired of tutorials that explain what to type without explaining why it works. Every article here is written to make concepts actually click.

← PreviousAutoSys date_conditions and run_windowNext →AutoSys sendevent Command
Forged with 🔥 at TheCodeForge.io — Where Developers Are Forged