High availability cluster setup in Kubernetes - Step-by-Step Execution

Process Flow - High availability cluster setup

Start: Plan HA Cluster

↓

Provision Multiple Master Nodes

↓

Configure etcd Cluster

↓

Set Up Load Balancer for Masters

↓

Join Worker Nodes to Cluster

↓

Verify Cluster Health & Failover

↓

End: HA Cluster Ready

This flow shows the main steps to set up a Kubernetes high availability cluster, from planning to verifying failover.

Execution Sample

Kubernetes

kubeadm init --control-plane-endpoint "LOAD_BALANCER_DNS:6443" --upload-certs
kubeadm join LOAD_BALANCER_DNS:6443 --token <token> --discovery-token-ca-cert-hash sha256:<hash> --control-plane
kubectl get nodes
kubectl get pods -n kube-system

Commands to initialize HA control plane, join nodes, and check cluster status.

Process Table

Step	Action	Command/Config	Result/Output
1	Initialize first master node with control plane endpoint	kubeadm init --control-plane-endpoint "LOAD_BALANCER_DNS:6443" --upload-certs	Master node initialized, certificates uploaded, control plane endpoint set
2	Join additional master nodes to control plane	kubeadm join LOAD_BALANCER_DNS:6443 --token <token> --discovery-token-ca-cert-hash sha256:<hash> --control-plane	Additional master nodes joined, etcd cluster formed
3	Set up external load balancer pointing to master nodes	Configure load balancer with masters' IPs on port 6443	Load balancer routes API requests to healthy masters
4	Join worker nodes to cluster	kubeadm join LOAD_BALANCER_DNS:6443 --token <token> --discovery-token-ca-cert-hash sha256:<hash>	Worker nodes joined and ready
5	Verify nodes and pods status	kubectl get nodes kubectl get pods -n kube-system	All nodes Ready, system pods Running
6	Simulate master node failure	Shutdown one master node	Load balancer routes traffic to remaining masters, cluster remains functional
7	Exit	N/A	HA cluster operational with failover

💡 HA cluster setup completes when multiple masters are joined, load balancer routes traffic, and failover works

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 4	Final
Master Nodes	0	1 (initialized)	3 (joined control plane)	3	3 (all active)
Worker Nodes	0	0	0	2 (joined)	2 (all Ready)
Load Balancer	Not configured	Not configured	Configured with 3 masters	Configured	Configured and routing
Cluster Health	Not available	Healthy (single master)	Healthy (HA masters)	Healthy (masters + workers)	Healthy with failover

Key Moments - 3 Insights

Why do we need a load balancer in front of master nodes?

The load balancer distributes API requests to all healthy master nodes, ensuring availability if one master fails. See execution_table step 3 and 6 where load balancer setup and failover are shown.

What happens if a master node goes down?

The load balancer routes traffic to remaining masters, so the cluster stays functional without downtime. This is shown in execution_table step 6.

Why do worker nodes join using the load balancer DNS instead of a single master IP?

Using the load balancer DNS ensures workers can always reach an active master, even if one master fails. Refer to execution_table steps 2 and 4.

Visual Quiz - 3 Questions

Look at the execution table, at which step are additional master nodes joined to the control plane?

A Step 2

B Step 1

C Step 4

D Step 5

💡 Hint

Check the 'Action' column for 'Join additional master nodes' in execution_table.

According to variable_tracker, how many worker nodes are active after step 4?

A 0

B 1

C 2

D 3

💡 Hint

Look at the 'Worker Nodes' row under 'After Step 4' in variable_tracker.

If the load balancer was not configured, what would likely happen during master node failure?

A Cluster continues working without issues

B API requests fail because no failover routing exists

C Worker nodes automatically become masters

D New load balancer is created automatically

💡 Hint

Refer to key_moments about load balancer role and execution_table step 3 and 6.

Concept Snapshot

High Availability Cluster Setup in Kubernetes:
- Use multiple master nodes for control plane redundancy
- Configure an external load balancer to route API requests
- Join worker nodes via load balancer DNS
- Verify cluster health and failover by simulating master failure
- Ensures no single point of failure for control plane

Full Transcript

This visual execution shows how to set up a Kubernetes high availability cluster. First, you initialize the first master node with a control plane endpoint pointing to a load balancer. Then, you join additional master nodes to form an etcd cluster and control plane. Next, you configure an external load balancer to route API requests to all healthy masters. Worker nodes join the cluster using the load balancer DNS to ensure connectivity. Finally, you verify the cluster health by checking nodes and pods status. Simulating a master node failure shows that the load balancer routes traffic to remaining masters, keeping the cluster operational without downtime.