Unusable application

Lanhild

I'm guessing 9.1.3 rendered the application unusable, as the issue appeared later than the latest update of this app.

Pushing/pulling an image from a remote works correctly. However, attempting to pull an image from the same server doesn't work at all.

For instance:

I have a registry @ registry.cloudron.local
I have an app @ app.cloudron.local, using an image from the registry, on the same cloudron.

Whenever I attempt to update this application with a new image, the pull never works.

Mar 19 18:51:04 box:docker Attempt 1 failed. Will retry: Unable to pull image registry.cloudron.local/app.image:2.0.13. registry error: {"reason":"server error","statusCode":500,"json":null}

In the registry configuration in Cloudron settings, the error Invalid serverAddress: socket hang up appears whenever I attempt to save the private registry settings.

In the registry application logs, there is the following line, appearing intermittently:

Mar 19 20:32:18 time="2026-03-20T00:32:18.895638941Z" level=error msg="traces export: Post \"https://localhost:4318/v1/traces\": dial tcp [::1]:4318: connect: connection refused" go.version=go1.23.7 instance.id=redacted service=registry version=3.0.0

I have multiple Cloudrons, all with a registry application installed and facing the same issue.

As mentioned in the beginning of my topic, pulling from a remote works.

Connecting the private registry registry.cloudron.local in my.other-cloudron.local works perfectly. It seems like there might be a DNS issue.

Feel free to ask for more information.

Lanhild

@girish pinging for visibility.

joseph

@lanhild I just tried to reproduce this and it seems to work fine. Have you added the local registry app as a Private docker registry i.e under System -> Docker -> Registries? The local registry app is not automatically available for app install (same cloudron or different cloudron).

robi

It may be the https issue which local to local can't connect

joseph

I assumed .local was like .example. Did you literally mean .local ?

Lanhild

@joseph I used .local for the sake of my example

Lanhild

@joseph said:

@lanhild I just tried to reproduce this and it seems to work fine. Have you added the local registry app as a Private docker registry i.e under System -> Docker -> Registries?

That's what I'm attempting to do, but it results in the error I mentioned.

joseph

In the registry configuration in Cloudron settings, the error Invalid serverAddress: socket hang up appears whenever I attempt to save the private registry settings.

OK, I missed this one. Can you try docker login on the server itself and see if that works ?

Lanhild

@joseph it doesn't, gives me a timeout error. I'll add relevant logs in a few minutes.

Lanhild

@joseph

Error response from daemon: Get "https://registry.cloudron.local/v2/": context deadline exceeded

joseph

@lanhild We have to debug this step by step.

Does host registry.cloudron.local resolve to public IP?
Check if nginx can reach the container.. docker ps | grep com.docker.registry to get container ID. docker inspect bea487bec441 | grep IPAddress to get local IP. curl http://172.18.20.34:5000/asdf -v should get some response for the registry. It will have Docker-Distribution-Api-Version header.
Then curl https://registry.cloudron.local ? This should give some UI page or login page of the registry

Depending on where it fails, we have to proceed. FWIW, docker login to same cloudron registry works fine on my servers.

Lanhild

@joseph My answers, step-by-step as you formatted it in your message:

Yes.
Reachable, Docker-Distribution-Api-Version: registry/2.0

<html>
<head><title>302 Found</title></head>
<body>
<center><h1>302 Found</h1></center>
<hr><center>nginx</center>
</body>
</html>

docker login from the Cloudron server results in the following

Error response from daemon: Get "https:/registry.cloudron.local/v2/": net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)

docker login from any other remote works perfectly fine.

Lanhild

I can't repro the issue on the demo server.

joseph

Out of ideas, since I can't reproduce this anywhere. You can contact me on support@ if you want me to debug on the server.

klawitterb

@joseph I've now also noticed problems with the docker registry. In my case I'm also running a builderbot on the same cloudron instance. Seems like it can't connect to the docker registry, my logs on the builderbot are simply saying:

Mar 28 08:42:05 build-1774683724218 Build stream finished
Mar 28 08:42:05 build-1774683724218 Pushing...
Mar 28 08:42:20 build-1774683724218 Push stream finished
Mar 28 08:42:20 build-1774683724218 Get "https://docker.apehost.de/v2/": net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
Mar 28 08:42:20 build-1774683724218 Push failed

Probably not related I've also noticed that the logs on the docker registry container get spammed with this error:

Mar 28 08:46:32 time="2026-03-28T07:46:32.467979729Z" level=error msg="traces export: Post \"https://localhost:4318/v1/traces\": dial tcp [::1]:4318: connect: connection refused" go.version=go1.23.7 instance.id=401bc88b-1bfe-4713-ae09-1cffdf628884 service=registry version=3.0.0
Mar 28 08:46:35 2026/03/28 07:46:35 [info] 71#71: *109 client 172.18.0.1 closed keepalive connection

Lanhild

@klawitterb said:
Probably not related I've also noticed that the logs on the docker registry container get spammed with this error:
Mar 28 08:46:32 time="2026-03-28T07:46:32.467979729Z" level=error msg="traces export: Post \"https://localhost:4318/v1/traces\": dial tcp [::1]:4318: connect: connection refused" go.version=go1.23.7 instance.id=401bc88b-1bfe-4713-ae09-1cffdf628884 service=registry version=3.0.0

I'm also encountering these.

robi

Assuming it is listening on port 4318, you can't do https with localhost.

Dropping the 's' should work if the API route is available.

girish

@Lanhild https://git.cloudron.io/packages/docker-registry-app/-/merge_requests/15 . I think that probably fixes it per upstream reports - https://github.com/distribution/distribution/issues/4270 , https://github.com/distribution/distribution-library-image/issues/187 . I don't think this related to your original DNS issue though. This traces export is just some log spam .

girish

OK, with the update that log message is gone (in our registry)

klawitterb

Log spam seems to be gone now, still having trouble with the actual connection to the registry on my cloudron:

Docker registry resides on the same cloudron

Cloudron makes it easy to run web apps like WordPress, Nextcloud, GitLab on your server. Find out more or install now.

Cloudron Forum

Unusable application