Step-by-Step Guide to Building a Kubernetes Cluster on CentOS 7

Environment requirements

A compatible Linux host (Debian or Red Hat based).

2 GB+ RAM per machine.

2 CPUs or more.

Full network connectivity between all machines.

Unique hostname, MAC address, and product_uuid for each node.

Required ports open (see linked documentation).

Swap must be disabled for kubelet to work.

System environment (2 CPU, 4 GB RAM): CentOS 7. Master node IP: 10.229.1.168, node1 IP: 10.229.3.251. All operations are performed as root.

Change hostnames

# master node set hostname
hostnamectl set-hostname k8s-master
# node1 set hostname
hostnamectl set-hostname k8s-node1
# add hosts entries
cat >>/etc/hosts <
Disable firewall, SELinux, and swap on all nodes
# disable firewall
systemctl status firewalld && systemctl stop firewalld && systemctl disable firewalld
# disable SELinux permanently
sed -i 's/enforcing/disabled/' /etc/selinux/config
setenforce 0
# disable swap temporarily and permanently
swapoff -a
vim /etc/fstab   # remove or comment out the swap line
Pass bridge IPv4 traffic to iptables
# create sysctl config
cat >/etc/sysctl.d/k8s.conf <
Use Alibaba Cloud yum repository for Kubernetes
# create repo file
cat > /etc/yum.repos.d/kubernetes.repo <
Install Docker and Kubernetes binaries
# install Docker
yum -y install docker
systemctl enable docker && systemctl start docker
docker -v
# configure Docker registry mirror
cat > /etc/docker/daemon.json <
Master node installation
# initialize the control plane (may need to pull images manually)
kubeadm init --pod-network-cidr 10.244.0.0/16 --kubernetes-version latest
# list required images
kubeadm config images list
# if images fail to pull, pull them manually and retag
docker pull registry.aliyuncs.com/google_containers/kube-apiserver:v1.22.1 && \
... (additional docker pull commands) ...
# retag images to official names
docker tag registry.aliyuncs.com/google_containers/kube-apiserver:v1.22.1 k8s.gcr.io/kube-apiserver:v1.22.1 && \
... (additional docker tag commands) ...
# remove the temporary images
docker rmi registry.aliyuncs.com/google_containers/kube-apiserver:v1.22.1 ...
# verify images
docker images
# re‑run init if needed
kubeadm init --pod-network-cidr 10.244.0.0/16 --kubernetes-version latest
After successful init, run as a regular user:
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
Or as root:
export KUBECONFIG=/etc/kubernetes/admin.conf
Check node status (will show NotReady until a network plugin is installed).
kubectl get nodes
Worker node (node1) installation
# pull required images on node1
docker pull registry.aliyuncs.com/google_containers/kube-proxy:v1.22.1 && \
... (additional pulls) ...
# join the cluster
kubeadm join 172.16.1.197:6443 --token ebi9py.oz4hmt72yk1wlvoe \
  --discovery-token-ca-cert-hash sha256:9990f921f6c66423fc097f81f2c4d5f2b851dc906cbce966db99de73dbce793b
If the join fails due to missing kubeconfig, copy it from the master:
scp root@k8s-master:/etc/kubernetes/admin.conf /etc/kubernetes/admin.conf
export KUBECONFIG=/etc/kubernetes/admin.conf
Install flannel network plugin
# download flannel manifest
wget https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml
# apply the manifest
kubectl apply -f kube-flannel.yml
# verify pods
kubectl get pods -n kube-system -o wide
If images cannot be pulled, manually download and load the flannel image:
# download flannel docker image
wget https://github.com/flannel-io/flannel/releases/download/v0.14.0/flanneld-v0.14.0-amd64.docker
# load into Docker and retag
docker load < flanneld-v0.14.0-amd64.docker && \
 docker tag quay.io/coreos/flannel:v0.14.0-amd64 quay.io/coreos/flannel:v0.14.0 && \
 docker rmi quay.io/coreos/flannel:v0.14.0-amd64
# reinstall
kubectl delete -f kube-flannel.yml
kubectl apply -f kube-flannel.yml
Cluster health checks
# component status
kubectl get cs
# configmaps
kubectl get configmap -n kube-system
# certificate expiration
kubeadm certs check-expiration
# generate join token
kubeadm token create --print-join-command
Cluster testing – nginx deployment
# create deployment and expose
kubectl create deployment nginx-deploy --image=nginx
kubectl expose deployment nginx-deploy --port=80 --type=NodePort
# get service details
kubectl get pod,svc
# test with curl
curl 172.16.0.188:32353
Install Kubernetes Dashboard
# apply dashboard manifest
kubectl apply -f https://raw.githubusercontent.com/kubernetes/dashboard/v2.3.1/aio/deploy/recommended.yaml
# expose via NodePort 30001
kubectl patch svc kubernetes-dashboard -n kubernetes-dashboard -p '{"spec":{"type":"NodePort","ports":[{"port":443,"targetPort":8443,"nodePort":30001}]}}'
# create admin user
cat > dashboard-adminuser.yaml <
Access the dashboard at
https://10.229.1.168:30001
(ignore the browser warning by typing
thisisunsafe
), log in with the retrieved token, and explore the UI.
Reference links:
https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/install-kubeadm/
https://blog.csdn.net/flying_monkey_1/article/details/118701275
https://blog.csdn.net/weixin_40039683/article/details/112886735
https://github.com/kubernetes/dashboard/releases