Hello community,
We are using self hosted n8n and yesterday we made a version update to n8n and our community node in that workflow got removed. Because of that the workflow was throwing an error.
Is it intentional or this is a bug I’m not sure. I just wanted to confirm that the version updates would remove the community nodes. Can you please confim?
Here are some screenshots. It worked again when I downloaded the community node so I don’t have another screenshots.
This could be either because the typeVersion in the node, or more likely the type has changed.
I’d recommend selecting that node, copying, and then pasting into a text editor. Then add a new node of that type to the canvas (this one should work). Do the same and copy into another editor and compare the values.
I had this recently where I changed how I installed community nodes. E.g. my "type": "n8n-nodes-ldap.LDAP" changed to "type": "CUSTOM.LDAP". To fix that, I updated the previous node with the new value, then copy pasted it back to the canvas.
Good luck, would be interesting to know what changed.
What node was it, What version were you on before the upgrade and what version did you upgrade to? How do you have n8n installed as well is it through Portainer?
I wonder if maybe there was a permission change at some point or if the node was accidentally deleted. I will keep an eye out for more reports of this but it is good to know you are still running.
Hey @Jon,
From your reply I assume that’s not expected with version updates right? Hard to say about the permission change because we have multiple users but I should be able to see all the workflows and credentials and I could see that the community node was removed.
Besides the version update what might cause the community node to be deleted if I may ask?
In theory yeah things shouldn’t change like that, There is nothing that should cause a community node to be deleted really. It could be that the user running the container was different so it is using a different path but that is something you would have noticed sooner if you are using SQLite.