ERROR: UNKNOWN ERROR - check the detailed error for more information socket hang up

Get error when I use httprequest.
But I use request-promise-native,it runs well.

Sorry, sadly no idea. But does not sound like something n8n specific, more Node.js related. For such non-n8n related questions, you will have more luck on stackoverflow.

Btw. the request library (and so also request-promise-native) will soon be removed from n8n as the library got deprecated a while ago and we now also replaced it in n8n with axios. So I suggest to use axios moving forward.

For me, stability is the most important thing, not the number of functions. It was good before version 138, and there will be various problems afterwards. I can only understand that you give me advice as not to let me upgrade.

Sorry do not really understand what you mean.

Anyway we did not replace the request library for fun (because we knew it will not be, and it was not), it was necessary as it was deprecated and we had to upgrade at some point. You are more than welcome to submit a PR to help and improve the n8n anytime or stay on 0.138.0 for as long as you want.

Generally you are right. If something is working and you do not need the new functionality: Do not upgrade! You have nothing to win, you can only lose.

I just want to say that the same is based on Node.js. I use A normally, but B is not normal. This phenomenon only appeared after version 138. You should find a way to understand why this is the case? I use Docker and set up automatic updates in Nas. I really don’t know which features have been updated in each version. Faced with our feedback, whether you care or not, let me use the old version. I It’s very uncomfortable. Why do I only use the old version? What is the significance of your upgrade?

What node are you actually using or is it a custom node or function?

Would be nice to know what service you are connecting to as well that way someone can try and work out why you getting the error.

The more detail you can provide the better.

My Nas needs to be accessed from the external network. I set up DDNS and visit a link on regularly to maintain access availability. My node uses HTTP Request, which is such a simple requirement.

So are you just updating the record on with an http post to set the IP to your external one on a schedule?

Can you share your workflow (removing anything important like credentials and ips) and the full error message?

@aston_he very sorry sadly still do not understand everything you wrote but try to answer as good as possible.

If a problem appears after 0.138.0 we know exactly why it did appear. It is because of the library switch. And all the problems we are made aware of we fix. You can look into GitHub there are a lot of commits that fix those specific problems.

Regarding new versions. You can see in our changelog what changes have been made. I suggest deactivating automatic updates and only update if there is really a need. That will generally also improve the uptime of n8n a lot. After all, will with each upgrade n8n go down for a little bit.

1 Like


In my case the error was by firewall. I had blocked all connections over port 5678 to avoid connections outside the reverse proxy’s SSL (nginx port 443).

But it seems that when node tries to query DNS it uses the port on which n8n is running natively.

If you are running n8n on port 5678 (or similar) and accessing through a reverse proxy (nginx, traefik, etc) remember to keep the n8n port open for these types of communications.

1 Like