Skip to content

Conversation

kludee
Copy link

@kludee kludee commented Feb 1, 2025

Im using the Minikube. When i tried to access the host’s container runtime, with Containerd, ctr and docker i got an empty table or different errors which ended to be rabit holes.

The installed different Docker-Binaries to get an Workaround and saw that docker usese the Domain-Unix-Socket /run/docker.sock when starting dockerd.
So I changed the Unix-domain-socket in the deployment.yaml from /run/containerd/containerd.sock to /run/docker.sock and it worked. Otherwise i did not get access to the host's container runtime environment.

I suspect that Kubernetes is using the Unix-socket containerd.sock and this is causing problems. I'll investigate the behavior more closely as soon as I have more time.

Works for the Minikube
I am not aware of the impact it would have on other deployments.

The path to the Unix domain socket has been changed. Otherwise you don't get access to the host's container runtime environment. Important so that Scenario 2 (DIND) can be carried out. 

*Works for the Minikube
@kludee kludee mentioned this pull request Feb 1, 2025
@madhuakula
Copy link
Owner

I understand the issue @kludee I think we somehow need to handle both cases of containerd.sock and docker.sock. Let me check

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants