Problems with stopping processes

Hi

We have a problem that we have a number of processes running and we can’t stop them and they don’t get executed.

Is there any solution for this ?

I guess you are running all workflows in the main process by setting EXECUTIONS_PROCESS=main because else that should not be able to happen as it simply kills the process and it can stop it no matter what.

The error #524 is a apparently a timeout. So it looks like it tries to stop the workflow but is somehow unable to do that. Sadly have not seen that before. What does the workflow do? Does it actually stop or is it supposed not to stop? Do you know where it does get stuck and why?

Hi

Yes I think we changed the EXECUTIONS_PROCESS=main to prevent memory runaway, based on our discussions.

We needed to restart the N8N so I can not trace where it did get stuck.

The workflow has just 3 nodes, a receiving webhook, a switch and HTTP post.

We have to monitor and see why it happens.

Kind regards
Mattias

How long do you expect these processes to run? Would you be able to share your workflow with us?

hi!

It is expected / (and normally does) last about a second. I try to can share it but it needs to be without credentials.

Here is at least a screenshot of workflow

It did happen again on another workflow which looks very similar.

We have added continue on fail on the last node.

Is there any way to make the process kill itself if it doesn´t work.

The interesting thing is that it seems that workflows have succeeded and failed (see the second picture)

Kind regards
Mattias

If you reload n8n website and then then go to “Executions” does it then still display them? Because I think that the “message” that informs the UI that the workflow did finish maybe simply gets lost because of connection problems and that is why the frontend still thinks that it is running even though it is not.

Hi

Yes it appears even when reload and it appears on both mine and other computers when logging into the same instance. So my feeling is that it not frontend related.

It started to happen after we change the setting to main as you descried it to prevent the memory spike.

I can reproduce it. Will check how it can be fixed.

2 Likes

Found the issue. Will try to release a new version with the fix later.

1 Like

Fix got released with [email protected]

2 Likes