Kubeadm Containerd Centos Errors

Kubeadm Containerd Centos Errors

So if you find yourself running kubeadm init on centos and using Containerd as the runtime, you might hit the following preflight error:

error execution phase preflight: [preflight] Some fatal errors occurred:
	[ERROR CRI]: container runtime is not running: output: time="2020-02-17T14:45:43Z" level=fatal msg="getting status of runtime failed: rpc error: code = Unimplemented desc = unknown service runtime.v1alpha2.RuntimeService"
, error: exit status 1

You check systemctl status containerd and it’s running.

You check your kubeadm init config and see it has this block:

---
apiVersion: kubeadm.k8s.io/v1beta2
kind: InitConfiguration
<<<< other stuff >>>>
nodeRegistration:
  criSocket: /var/run/containerd/containerd.sock

Well turns out if you cat roles/containerd/files/config.toml you’ll see this line, which I think is the default:

disabled_plugins = ["cri"]

Change that to

disabled_plugins = [""]

and kubeadm init should stop complaining about the container runtime.