n8n now needs two settings, not one.

1. Keep EXECUTIONS_MODE=regular (you already did that).

1. ALSO add EXECUTIONS_PROCESS=main.

Without step 2 each run forks into a child process that never reports back, so the node “spins” forever.

Add that env-var, redeploy, and your Code nodes should finish normally.

Didn't work for me. Added that variable first to the primary, after it didn't work added it to the workers as well and it still does not work, it keeps spinning

Hmmm, try adding this env-var too: EXECUTIONS_DATA_SAVE_ON_SUCCESS=all

With these three, I would think it should be good:

EXECUTIONS_MODE=regular

EXECUTIONS_PROCESS=main

EXECUTIONS_DATA_SAVE_ON_SUCCESS=all

Hopefully, this finally fixes the spinner for you!

Just to confirm once more, since I have your attention (and thank you for that). Do I add these on Primary, Workers or both?

No worries at all, and you're welcome!

If you’re running a simple single-container install (no queue/worker pair), just put the vars on the Primary service.

If you’re still running a separate worker container:

• EXECUTIONS_MODE and EXECUTIONS_PROCESS Belong on the Primary (the worker ignores them). So set only on the Primary.

• EXECUTIONS_DATA_SAVE_ON_SUCCESS (and the ONERROR one) should be set on both Primary and Worker, so whichever side finishes the job knows how to store the data.

Here are screenshots. I added the EXECUTIONS_DATA_SAVE_ON_SUCCESS but it still didn't work. Idk if I should delete it in the workers container as well as the EXECUTIONS_PROCESS

Your screenshots show Primary is in regular mode while the extra Worker container is still running in queue mode.

That mix makes the Code node spin forever.

Choose ONE setup:

1. Simple (no queue) – easiest

• Stop / delete the Worker (and any ‘webhook-processor’) service.

• Keep only the Primary container.
  Vars on Primary:

EXECUTIONS_MODE=regular

EXECUTIONS_PROCESS=main

EXECUTIONS_DATA_SAVE_ON_SUCCESS=all

Redeploy Primary.

Code nodes should finish and show their output.

2. Queue mode (needs Worker)

• Leave both Primary and Worker running.

• Set EXECUTIONS_MODE=queue on both.

• Add EXECUTIONS_DATA_SAVE_ON_SUCCESS=all on both.

• You can remove EXECUTIONS_PROCESS from the Worker – it’s ignored there.

Redeploy Primary first, then Worker.

Pick whichever style you want; just don’t mix them. After that, the Code nodes should stop hanging.

IT WORKED. THANK YOU SO MUCH

Wooooo, finally!
You're welcome dude, I am happy I could help you out and I hope you're able to rest peacefully ahaha.

Happy deving !