This guide outlines a structured approach to upgrading an Amazon Elastic Kubernetes Service (EKS) cluster, covering both the control plane and worker nodes.

It provides detailed steps for preparation, execution, and post-upgrade verification, ensuring a smooth transition to newer Kubernetes versions. Key actions include:

• verifying compatibility,
• backing up data,
• updating tools, cordoning nodes to prevent new pod scheduling, and systematically upgrading both managed and self-managed node groups.

The guide also emphasizes the importance of updating essential add-ons and testing workloads post-upgrade to maintain the stability and performance of your Kubernetes environment. This ensures your cluster leverages the latest features, security patches, and performance improvements, providing a robust foundation for cloud-native applications.

Step-by-Step Upgrade Guide

1. Preparation

1. Check Compatibility:

• Verify that your workloads and add-ons are compatible with the target Kubernetes version.
• Review the EKS Kubernetes version release notes.

2. Backup Data:

• Backup cluster data and configurations to ensure you can restore them if needed.

3. Upgrade kubectl:

curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
chmod +x kubectl
sudo mv kubectl /usr/local/bin/

4. Update AWS CLI:

curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
unzip awscliv2.zip
sudo ./aws/install

2. Upgrade EKS Control Plane

1. Check Current Version:

aws eks describe-cluster --name <cluster_name> --query cluster.version --output text

2. Initiate Control Plane Upgrade:

aws eks update-cluster-version --name <cluster_name> --kubernetes-version <target_version>

3. Monitor the Upgrade Process: The upgrade process can take some time. Monitor the status using:

aws eks describe-update --name <cluster_name> --update-id <update_id>

3. Upgrade Worker Nodes

Managed Node Groups

1. Cordon Nodes: Prevent new pods from being scheduled on nodes:

kubectl cordon <node_name>

2. Update Managed Node Groups: For each managed node group:

aws eks update-nodegroup-version --cluster-name <cluster_name> --nodegroup-name <node_group_name> --kubernetes-version <target_version>

3. Monitor the Upgrade Process:

Check the status of the node group update:

aws eks describe-update --cluster-name <cluster_name> --nodegroup-name <node_group_name> --update-id <update_id>

Self-Managed Nodes

1. Cordon and Drain Old Nodes:

• Prevent new pods from being scheduled and safely migrate workloads:

kubectl cordon <node_name>
kubectl drain <node_name> --ignore-daemonsets --delete-local-data

2. Create a New Node Group: Create a new node group with the desired Kubernetes version:

eksctl create nodegroup --cluster <cluster_name> --name <new_node_group> --kubernetes-version <target_version>

3. Delete Old Node Group:

• Remove the old node group:

eksctl delete nodegroup --cluster <cluster_name> --name <old_node_group>

4. Post-Upgrade Steps

1. Verify Upgrade:

• Check the nodes and ensure they are running the desired version:

kubectl get nodes

2. Update Add-ons:

• Update essential add-ons to ensure compatibility with the new Kubernetes version:

eksctl update addon --name vpc-cni --cluster <cluster_name>
eksctl update addon --name kube-proxy --cluster <cluster_name>
eksctl update addon --name coredns --cluster <cluster_name>

3. Test Workloads:

• Ensure all applications are functioning correctly post-upgrade:

kubectl get pods -A

4. Clean Up:

• Remove old node groups if not done already:

Summary

With this structured approach, including cordoning nodes before the upgrade, you can efficiently upgrade your EKS cluster. This process helps ensure that your Kubernetes environment is up-to-date with the latest features and security enhancements, maintaining the robustness and performance of your applications and providing a solid foundation for your cloud-native workloads.