Adding MLFLOW Container locally

There are quite some Deep Learning/Machine Learning orientated containers in the AppStore. What was missing for me was MLFLOW to track different training runs.

When I wanted to add the Container locally, I ran into the error:

Error response from daemon: pull access denied for docker.1panel.live/mlflow/mlflow, repository does not exist or may require 'docker login': denied: only support mainland China
I created this issue.

Why do I get this error at all, How to avoid it?

I’ve been testing this as well, and I’m running into the same behaviour you described.

Right now ZimaOS always tries to rewrite the image name to:

docker.1panel.live/mlflow/mlflow

And that mirror only works inside mainland China, so the pull fails even if you point to a normal Docker Hub image. I’ve tried several installation methods and overrides, but they all get mapped back to the same China-only registry, which is why the container never pulls correctly.

I’m still working through it and trying alternative approaches (including GHCR images and custom overrides). Once I find the clean, repeatable solution, I’ll post the full working procedure here so you can install MLflow normally on ZimaOS.

Just wanted to confirm you’re not doing anything wrong, the behaviour is reproducible. I’ll update this thread as soon as I have the fix.

1 Like