Kubeadm init with failed node error

General Discussions
1

image
image1173×420 18.2 KB

i am using ubuntu server.
when running kubeadm init, i had the errors showing in the screenshot (have also provided the kubeadm version in the screenshot). The error looks like
May 17 06:44:44 node1 kubelet[1965]: E0517 06:44:44.563519 1965 kubelet.go:2419] “Error getting node” err=“node “node1” not found”
May 17 06:44:44 node1 kubelet[1965]: E0517 06:44:44.664897 1965 kubelet.go:2419] “Error getting node” err=“node “node1” not found”
May 17 06:44:44 node1 kubelet[1965]: E0517 06:44:44.765447 1965 kubelet.go:2419] “Error getting node” err=“node “node1” not found”

but i am not sure if this is the root cause for the problem. Can any one help on this?

2

image
image940×64 3.65 KB

3

image
image1265×454 15.5 KB

4

This is happening when your control-plane is not accessible which is sounds weird if you are already setting up the control-plane :slight_smile:

  • If you are using proxy to access outside word set NO_PROXY in your systemctl env
  • allow traffic in firewall
1 Like
5

Solved for vanilla kubernetes with CRI-O as container runtime.

Causes:
I guess it’s because of lack some module during install CRI-O. maybe runc module.
It might be bug of CRI-O install package.

Workaround:

  • Install CRIO and start it.
  • Install docker to install runc perfectly.
# yum install -y yum-utils device-mapper-persistent-data lvm2
# yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo
# yum install docker-ce -y
  • When you init k8s cluster using kubeadm, add ‘–cri-socket’ option as followed.
    ex)
    kubeadm init
    –kubernetes-version=v1.23.1
    –pod-network-cidr=192.168.0.0/16
    –apiserver-advertise-address=10.193.229.141
    –cri-socket /var/run/crio/crio.sock