Problem running minikube on current WSL-2

[mwoodpatrick]

Is there an existing issue for this?

• I have searched the existing issues

What did you expect to happen?

I'm having a problem getting minikube to run on WSL-2 it reports having problems accessing https://registry.k8s.io/ but on the same shell I can ping that server. I would appreciate any suggestions on how to debug this. It has worked fine previously.

minikube start
😄 minikube v1.29.0 on Ubuntu 23.04 (amd64)
✨ Using the docker driver based on existing profile
👍 Starting control plane node minikube in cluster minikube
🚜 Pulling base image ...
🏃 Updating the running docker "minikube" container ...
❗ This container is having trouble accessing https://registry.k8s.io
💡 To pull new external images, you may need to configure a proxy: https://minikube.sigs.k8s.io/docs/reference/networking/proxy/
🐳 Preparing Kubernetes v1.26.1 on Docker 20.10.23 ...

Debugging Information

curl.log
dig.log
oras.log
oras_pull.log

Anything else?

minikube version
minikube version: v1.29.0
commit: ddac20b4b34a9c8c857fc602203b6ba2679794d3

wsl --version
WSL version: 1.1.7.0
Kernel version: 6.1.21.1-1
WSLg version: 1.0.51
MSRDC version: 1.2.3770
Direct3D version: 1.608.2-61064218
DXCore version: 10.0.25131.1002-220531-1700.rs-onecore-base2-hyp
Windows version: 10.0.22624.1537

Code of Conduct

• I agree to follow this project's Code of Conduct

[BenTheElder]

So we can see pulling is working fine from your host with the included debugging logs you uploaded, but we can't see what is wrong from the minikube side from the minikube output.

Can you run crane pull --verbose registry.k8s.io/pause:3.9 /dev/null from within minikube? there's a sample command in the debugging guide that runs this as a pod.

[TerryHowe]

Debugging guide https://github.com/kubernetes/registry.k8s.io/blob/main/docs/debugging.md

Looks like everything ran fine on your machine, just not working on minikube.

[BenTheElder]

cc @medyagh @spowelljr @afbjorklund 👀

[BenTheElder]

I would guess this is proxy related as hinted by minikube, but can't be certain without more information.

[medyagh]

@mwoodpatrick on the first look it appears to me the problem to be with your network access inside WSL.

however we just released 1.30 https://github.com/kubernetes/minikube/releases/tag/v1.30.0 that includes this PR kubernetes/minikube#16200

do you mind retrying with v1.30

also do you mind sharing the full curl command log with -v ?

[mwoodpatrick]

The curl command I ran (whose log is attached) was

curl -vhttps://registry.k8s.io/v2`

I updated to minikube version: v1.30.1 and have kubectl v1.26.3 when I run

kubectl run --rm -it crane --restart=Never --image=gcr.io/go-containerregistry/crane --overrides='{"spec": {"hostNetwork":true}}' -- pull --verbose registry.k8s.io/pause:3.9 /dev/null

I get

pod "crane" deleted
error: timed out waiting for the condition

[BenTheElder]

Since this is using the docker driver, can you do something like docker exec -it $minikube_container bash and run debug from there?

Otherwise will defer to Minikube maintainers, all of the logs from the host look fine.

[mwoodpatrick]

Note this issue happens even if I'm not using VPN. Please let me know of the best way to debug this,

[BenTheElder]

I think you will need to try some of these debugging steps again but inside the minikube node, from the output above it appears to be a docker container. #214 (comment)

With docker ps you can check the running containers for the minikube container and then with docker exec -it run commands interactively inside it.

[medyagh]

@mwoodpatrick for sake of a centralized conversation we could continue the conversation in slack channel
https://kubernetes.slack.com/archives/C1F5CT6Q1/p1680612729918929
thread and update this issue with the summary and finding once resolved for other users.

[mwoodpatrick]

Looks like issue was new Microsoft Linux kernel updating to https://github.com/microsoft/WSL/releases/tag/1.2.0 fixed the issue