bsrCh3rif07/Installing-Kubernetes

Collection of scripts that enable you to quickly set up a Kubernetes cluster with all the necessary tools.

Installing-Kubernetes

This repository serves as a cheat sheet for installing Kubernetes and the differents tools needed to the Cluster management.

1- Installing tools on all nodes:

The k8s-allnodes-containerd-setup.sh script installs all the necessary tools to set up Kubernetes with Kubeadm. After running it on the control plane and worker nodes, simply run on the control plane:

sudo kubeadm config images pull
sudo kubeadm init --apiserver-advertise-address=10.0.2.5 \
--cri-socket=unix:///var/run/containerd/containerd.sock \
--pod-network-cidr=192.168.0.0/16

2- Kubelet Container runtime network not ready issue

If you encounter the error "Container runtime network not ready" after successfully initializing the cluster, it may be due to a problem with the CNI failing to load the config. Here is a solution to fix it:

a- Download the setup CNI file

curl https://raw.githubusercontent.com/containerd/containerd/main/script/setup/install-cni | sudo tee $HOME/install-cni

b- Install Go

wget https://storage.googleapis.com/golang/getgo/installer_linux
chmod +x ./installer_linux
./installer_linux
source ~/.bash_profile

c- Run the installation file:

cd ~
sudo chmod +x install-cni
sudo sed -i '/CNI_COMMIT/d' install-cni
./install-cni

d- Finally, restart Containerd and Kubelet:

sudo systemctl restart containerd
sudo systemctl restart kubelet

3- Setup Kubectl & Install a network plugin

To manage the Kubernetes cluster, you need to set up kubectl. Run the following commands to create a directory for kubectl configuration, copy the admin.conf file to the directory, and change the ownership of the configuration file

mkdir -p $HOME/.kube
  sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
  sudo chown $(id -u):$(id -g) $HOME/.kube/config

After setting up kubectl, you can install a network plugin. For example, you can use Calico for pod networking by running the following command on the control plane:

curl https://raw.githubusercontent.com/projectcalico/calico/v3.25.0/manifests/calico.yaml -O
kubectl apply -f calico.yaml

4- Join the worker nodes

If the control plane setup completed successfully, you can join the worker nodes to the cluster by running the command output by kubeadm init in step 1. The command should look like this:

kubeadm join 10.0.2.15:6443 --token 0ks9ue.z3azsbowa7lkwxm7 \
--discovery-token-ca-cert-hash sha256:865200c503b7e2da05bd51a9f7fbce84b3f467a08b9aa614f36ce7fc40250a24