17 Aug 2021

Create a Multi-Cloud Setup of Kubernetes cluster

architecture

Description:

CREATE A MULTI-CLOUD SETUP of K8S cluster :

Launch node in AWS
Launch node in Azure
Launch node in GCP
One node on the cloud should be Master Node
Then set up multi-node Kubernetes cluster.

So to do this task, I have created three nodes :

Master node on AWS, Slave nodes on AWS, Azure, and GCP.

First: Setting up Kubernetes master on AWS :

k8s-1

Step-1: For installing kubelet, kubeadm, kubectl first, we need to set up a repo for this :

vim /etc/yum.repos.d/k8s.repo
# content inside repo k8s.repo
[kubernetes]
name=Kubernetes
baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-\$basearch
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg

Step-2: Installing required software :

yum install docker kubelet kubeadm kubectl iproute-tc -y

Step-3: Starting and enabling services :

systemctl enable --now docker
systemctl enable --now kubelet

Step-4: We also need to pull docker images using kubeadm. It pulls images of the config files.

kubeadm config  images pull

Step-5: Now, we need to change the docker cgroup driver into systemd

vim /etc/docker/daemon.json
{
"exec-opts": ["native.cgroupdriver=systemd"]
}

Step-6: Since we have made changes in docker, we need to restart the docker service :

systemctl restart docker

Step-7: Setting up a network bridge to 1 :

echo "1" > /proc/sys/net/bridge/bridge-nf-call-iptable

Step-8: The important step is while initializing Master, while running preflight the main thing is we need to associate the token to the public IP of instance, so that any of the other nodes can easily connect, so for this use :

--control-plane-endpoint=<PUBLIC_IP>:6443
kubeadm init --pod-network-cidr=10.244.0.0/16 --control-plane-endpoint=<public_ip>:6443 --ignore-preflight-errors=NumCPU          --ignore-preflight-errors=Mem

Step-9: Now, make a directory for Kube config files and give permission to them :

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

Step-10: Apply flannel :

kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml

Step-11: Final step: Generate token so that slave nodes could connect to master node :

kubeadm token create --print-join-command

Second: Setting up Kubernetes nodes on AWS, Azure, GCP :

(Note: Follow same steps in all the three platforms)

AWS
Azure
GCP

Step-1: For installing kubelet, kubeadm, kubectl first, we need to set up a repo for this :

vim /etc/yum.repos.d/k8s.repo
# content inside repo k8s.repo
[kubernetes]
name=Kubernetes
baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-\$basearch
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg

Step-2: Installing required software :

yum install docker kubelet kubeadm kubectl iproute-tc -y

Step-3: Starting and enabling services :

systemctl enable --now docker
systemctl enable --now kubelet

Step-4: We also need to pull docker images using kubeadm. It pulls images of the config files :

kubeadm config  images pull

Step-5: Now, we need to change the docker cgroupdriver into systemd :

vim /etc/docker/daemon.json
{
"exec-opts": ["native.cgroupdriver=systemd"]
}

Step-6: Since we have made changes in docker, we need to restart the docker service :

systemctl restart docker

Step-7: Setting up a network bridge to 1 :

echo "1" > /proc/sys/net/bridge/bridge-nf-call-iptable

Step-8: Copy-paste the token generated in the master node….

Finally, in the master node :

kubectl get nodes

You will see that all the nodes are connected and are ready : res