Kubeadm

Kubeadm is a tool built to provide best-practice “fast paths” for creating Kubernetes clusters. It uses kubeadm init and kubeadm join type simple commands for creating a better end user experience for cluster creation & maintenence.
It enables Kubernetes administrators to quickly and easily bootstrap minimum viable clusters that are fully compliant with Certified Kubernetes guidelines.

As a defacto cluster creation tool it has been used by projects like minikube,kops etc.

Kubeadm is focused on bootstrapping Kubernetes clusters on existing infrastructure and performing an essential set of maintenance tasks. Its scope of work includes :

• Creating a control plane node (master node) by kubeadm init & joining other master and worker nodes by kubedm join command.

• It also includes various utilities for performing management tasks like control plane upgrades on already bootstrapped clusters , token and certificate renewal.

What kubeadm is not supposed to do

• Infrastructure provisioning and direct infrastructure manipulation(as done by kops for cloud/onprem or minikube for local)
• Third party networking integration
• Monitoring /logging etc.
• Specific cloud provider intergration (This actually is handled well by Cloud Controller Manager which will be discussed in details in another blog post)

Creating a Single Control plane cluster using kubeadm

Now lets get hands dirty with installation & configuration of kubeadm to deploy a cluster with single master and 2 worker nodes.

Prerequisites

• 3 Virtual machines with Ubuntu 18.04 installed and sudo privileges. Name one node as master & other 2 as worker1 & worker2 repectively.
• 2 GB or more of RAM per machine.
• 2 CPUs or more on all nodes.

End Goal

• Setup a Kubernetes cluster with single master node & two worker nodes
• Complete networking setup for the cluster so that Pods can talk to each other

Video

Steps to be followed on all nodes

• Install container runtime on all three nodes (Docker in the current tutorial)

Ref: https://kubernetes.io/docs/setup/production-environment/container-runtimes/#docker

# Install Docker CE
## Set up the repository:
### Install packages to allow apt to use a repository over HTTPS
apt-get update && apt-get install apt-transport-https ca-certificates curl software-properties-common

### Add Docker’s official GPG key
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | apt-key add -

### Add Docker apt repository.
add-apt-repository \
  "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
  $(lsb_release -cs) \
  stable"

## Install Docker CE.
apt-get update && apt-get install docker-ce=18.06.2~ce~3-0~ubuntu

# Setup daemon.
cat > /etc/docker/daemon.json <<EOF
{
  "exec-opts": ["native.cgroupdriver=systemd"],
  "log-driver": "json-file",
  "log-opts": {
    "max-size": "100m"
  },
  "storage-driver": "overlay2"
}
EOF

mkdir -p /etc/systemd/system/docker.service.d

# Restart docker.
systemctl daemon-reload
systemctl restart docker

• Installing kubeadm, kubelet and kubectl

  We will install following packages on all the three nodes:

  • kubeadm: the tool to bootstrap the cluster.
  • kubelet: the component that runs on all of the machines in your cluster as systemd process and does things like starting pods and containers.
  • kubectl: the command line utility to talk to your cluster.

apt-get update && apt-get install -y apt-transport-https curl
curl -s https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key add -
cat <<EOF >/etc/apt/sources.list.d/kubernetes.list
deb https://apt.kubernetes.io/ kubernetes-xenial main
EOF
apt-get update
apt-get install -y kubelet kubeadm kubectl
apt-mark hold kubelet kubeadm kubectl

Setup Control Plane on node with hostname set to master

kubeadm init --pod-network-cidr=192.168.0.0/16

After running kubeadm init successfully we get output that master node is initialised & we get some command to setup kubectl & also command for joining worker nodes to this master node.

Note: We are using Calico as network plugin, so we are using "--pod-network-cidr=192.168.0.0/16" option

• Setup kubectl

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

• Setup Kubernetes network by installing Calico network plugin

kubectl apply -f https://docs.projectcalico.org/v3.8/manifests/calico.yaml

• Finally look for master Node status

kubectl get nodes

NAME      STATUS     ROLES    AGE   VERSION
master    Ready      master   9m   v1.16.0

Setup worker nodes

• Copy the kudeadm join command from the output of kubeadm init and run on the worker nodes to join them to the cluster.

• To regenerate kubeadm join command output to join worker nodes run below command on master node & use the output to join the worker nodes.

kubeadm token create --print-join-command

Final status check of cluster

• Run kubeadm get nodes on master node again to check the status of worker nodes

kubectl get nodes

NAME      STATUS     ROLES    AGE   VERSION
master    Ready      master  21m   v1.16.0
worker1   Ready      <none>   9m   v1.16.0
worker2   Ready      <none>   9m   v1.16.0

• Find cluster information

kubectl cluster-info

Kubernetes master is running at https://157.245.118.253:6443
KubeDNS is running at https://157.245.118.253:6443/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy

To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.

As seen above in kubectl cluster-info output our single control plane setup is complete & cluster is up and running with control plane master running on Node master with ip https://157.245.118.253:6443 & KubeDNS is running at endpoint as mentioned above. Next we can check the health of control plane components.

• Check health status of various components of control plane.
  

kubectl get componentstatus

NAME                 STATUS    MESSAGE             ERROR
controller-manager   Healthy   ok
scheduler            Healthy   ok
etcd-0               Healthy   {"health":"true"}   

All the control plane components are up & in healthy state.

Conclusion

As seen above kubeadm has made single control plane setup a real easy & no sweat job.
In next article of the kubeadm series we will discuss steps for setting up a HA multimaster control plane.