Executions were success, after restart the status changes to crashed

Questions
1

Describe the problem/error/question

I am deploying a new n8n queue-mode setup (main + workers, runners, Redis, Postgres).

Everything works correctly under normal operation. However, after restarting the stack to apply configuration changes, the workers become stuck while recovering crashed executions.

I discovered around 750 crashed executions in the UI, and shortly after that, all executions that were previously marked as successful suddenly switched to an error state.

What is the error message (if any)?

Please share your workflow

Was testing with simple code nodes and wait to test concurrency.

Share the output returned by the last node

Information on your n8n setup

  • n8n version:2.6.3
  • Database (default: SQLite):Postgres
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):docker
  • Operating system:ubuntu
3
  1. Using Docker Compose
  2. V2.6.3
  3. Workers recover after sometime. I think what’s happening is that it looks at previous executions, doesn’t see finish time or status because it’s looking the logs not the DB. then marks it as crashed.
    The longer i keep it running, the more crashed executions i see.
  4. No errors, just the worker being in recovery.

I made a custom image of v2.6.3 disabling the recovery process, and using it for the workers. And it’s stable so far.

5

Will recovery mode being disabled cause any issues later?
The only resource i found was this opened issue. Luckily it helped me get it fixed for now.

6

Goodmorning,

I have seen weird behaviour in older versions, but those issues were patched.
Seems like it could be a setup issue or bug. Not sure.
You would need to give us a lot more information on the instance setup to determine what it is and possibly report a bug.