Homeâ€ș Blogâ€ș Ansible Automation Platform Jobs Stuck in Pending:

Life Tech

Ansible Automation Platform Jobs Stuck in Pending: Root Cause and Fix


Today I ran into one of those issues that looks complex, but turns out to be beautifully simple.

Every job I launched in Ansible Automation Platform (Tower / Controller 4.5) just sat there.
No output. No errors. No movement. Just PENDING.

And sometimes, silence is the loudest signal.

What I Saw

Everything looked healthy on the surface:

  • Job templates launched successfully
  • Jobs stayed in PENDING forever
  • No logs, no failures, no hints
  • Execution Environments looked perfectly fine

At first, I suspected Ansible itself,  playbooks, environments, something deep.

But this didn’t feel like a playbook problem.
This felt like something wasn’t even starting.

The Turning Point

When jobs don’t start at all, it’s usually not Ansible
 it’s scheduling.

So I went one level lower services.

That’s when I checked Receptor, the quiet engine behind job execution in AAP 4.x.

And there it was:

The Real Error

Running Receptor manually revealed the truth:

That one line explained everything.

What Actually Broke

Receptor relies on a Unix socket located here:

Here’s the subtle part:

  • /var/run (linked to /run) is temporary
  • It gets cleared on reboot or system cleanup
  • The directory /var/run/receptor was simply
 gone

And Receptor?
It doesn’t recreate it.

So it failed silently.
And when Receptor is down:

  • No execution capacity
  • No scheduling
  • Jobs stay in PENDING forever

No errors in UI, because nothing even reached that layer.

The Fix

Sometimes the fix feels almost too simple.

I recreated the directory:

Then restarted services:

And just like that—

PENDING → RUNNING

The system came back to life.

Making It Permanent

Because /var/run is temporary, this would happen again after reboot.

So I made it persistent using systemd:

Create:

Add:

Then apply:

Now the directory is recreated automatically every time.

What This Taught Me

When everything looks fine, but nothing moves:

  • Don’t start with playbooks
  • Don’t chase UI clues
  • Go deeper

Sometimes the failure isn’t in automation

It’s in the foundation that enables it.

A missing directory.
A silent service.
A system waiting for something that no longer exists.

Final Thought

Not every problem announces itself.

Some just sit there quietly

like a job stuck in PENDING,
waiting for you to look where no one else does.

Jay

Comments

How was your experience?
Your feedback helps improve this site.
PoorExcellent