restart faileds: webhook "sinkbindings.webhook.sources.knative.dev": Client.Timeout exceeded while awaiting headers

[kameshsampath]

minikube version
minikube version: v1.6.2 commit: 54f28ac5d3a815d1196cd5d57d707439ee4bb392

The exact command to reproduce the issue:

minikube start

The full output of the command that failed:

• first start a minikube instance using the command `minikube -p dmeo start --kubernetes-version=v1.14.7
• once the instance is started wait for some time
• stop the instance
• start the instance from same profile

😄  [knative] minikube v1.6.2 on Darwin 10.15.2
✨  Selecting 'hyperkit' driver from user configuration (alternates: [virtualbox])
👍  Kubernetes 1.17.0 is now available. If you would like to upgrade, specify: --kubernetes-version=1.17.0
🔄  Starting existing hyperkit VM for "knative" ...
⌛  Waiting for the host to be provisioned ...
🐳  Preparing Kubernetes v1.14.7 on Docker '19.03.5' ...
🚀  Launching Kubernetes ...

💣  Error starting cluster: addon phase cmd:"/bin/bash -c \"sudo env PATH=/var/lib/minikube/binaries/v1.14.7:$PATH kubeadm init phase addon all --config /var/tmp/minikube/kubeadm.yaml\"": /bin/bash -c "sudo env PATH=/var/lib/minikube/binaries/v1.14.7:$PATH kubeadm init phase addon all --config /var/tmp/minikube/kubeadm.yaml": Process exited with status 1
stdout:

stderr:
error execution phase addon/coredns: unable to create deployment: Internal error occurred: failed calling webhook "sinkbindings.webhook.sources.knative.dev": Post https://eventing-webhook.knative-eventing.svc:443/sinkbindings?timeout=30s: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)


😿  minikube is exiting due to an error. If the above message is not useful, open an issue:
👉  https://github.com/kubernetes/minikube/issues/new/choose

The output of the minikube logs command: https://gist.github.com/8e68c35f4d80e8376534eda149a56852

The operating system version: macOS Catalina v10.15.1

[medyagh]

@kameshsampath the problem seems to be the addon/coredns addon, do you mind disable and enable the addon and see how it will affect the issue?

[kameshsampath]

@medyagh - none of the addons are enabled but still same issue

[awgneo]

I have this problem as well. This addon isn't even in my addons list.

[tstromberg]

The coredns addon is built-in and required by kubeadm, so there is no possibility of tuning that behavior.

As best as I can tell, something configured a knative webhook to be contacted in certain cases, and that webhook is not available at the time that coredns comes up.

You can workaround this by using minikube delete, but it would be worth hearing how you got to this point so that we could replicate it ourselves?

I think the proper behavior here may be to get kubeadm to ignore webhooks, but I'm not sure that is possible.

[tstromberg]

I'm working on a fix for this, which resets clusters that are in a bad state.