Deploy self-managed Ubuntu nodes¶

Amazon Web Services provides documentation for launching self-managed node groups. Canonical publishes Ubuntu LTS and Ubuntu Pro LTS EKS images based on Ubuntu Minimal, containing the necessary components to launch Ubuntu EKS worker nodes.

This guide details the process of deploying self-managed node groups using the eksctl command-line tool.

Requirements:

AWSCLI installed and configured
eksctl (make sure to keep it updated to support latest EKS and Ubuntu versions)

Launching node groups from the command line¶

Eksctl supports Ubuntu nodes natively. To specify Ubuntu nodes, in the eksctl create nodegroup command, supply the --node-ami-family argument with one of the following supported Node AMI families:

`amiFamily`	Ubuntu release	Supported EKS versions
`Ubuntu2204`	22.04 Jammy Jellyfish	1.29 – 1.32
`UbuntuPro2204`	22.04 Jammy Jellyfish (Pro)	1.29 – 1.34
`Ubuntu2404`	24.04 Noble Numbat	> 1.31
`UbuntuPro2404`	24.04 Noble Numbat (Pro)	> 1.31
`Ubuntu2604`	26.04 Resolute Raccoon	> 1.36 (eksctl > v0.226.0)
`UbuntuPro2604`	26.04 Resolute Raccoon (Pro)	> 1.36 (eksctl > v0.226.0)

Note

If you prefer to use managed node groups, refer to this guide. In that scenario, the AMI is defined via a launch template, and EKS automatically handles the node recycling strategy.

Therefore, launching a new node group with the Ubuntu 24.04 LTS AMI would be:

eksctl create nodegroup \
  --cluster my-ubuntu-cluster \
  --name 2404-nodes \
  --node-type t3.medium \
  --nodes 3 \
  --nodes-min 1 \
  --nodes-max 4 \
  --ssh-access \
  --ssh-public-key my-key \
  --node-ami-family ubuntu2404

To use Ubuntu Pro, which enables Kernel Livepatch on the nodes and allows the cluster to run Ubuntu Pro containers, use the following command:

eksctl create nodegroup \
  --cluster my-ubuntu-cluster \
  --name 2404-pro-nodes \
  --node-type t3.medium \
  --nodes 3 \
  --nodes-min 1 \
  --nodes-max 4 \
  --ssh-access \
  --ssh-public-key my-key \
  --node-ami-family ubuntupro2404

Launching node groups from a YAML configuration file¶

If you need to change default configurations or track configuration changes as code, you can start the new node group from a cluster configuration YAML file:

---
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig

metadata:
  name: my-ubuntu-cluster
  region: us-east-1

nodeGroups:
  - name: 2404-pro-nodes
    amiFamily: UbuntuPro2404
    instanceType: t3.medium
    desiredCapacity: 3
    minSize: 1
    maxSize: 4
    ssh:
      allow: true
      publicKeyName: my-key

Save this file, say nodegroup.yaml and launch it with:

eksctl create nodegroup -f nodegroup.yaml

Validation¶

If you have kubectl installed and configured, run kubectl get nodes -o wide to get extended information about your clusters:

kubectl get nodes -o wide
NAME                             STATUS   ROLES    AGE    VERSION   INTERNAL-IP      EXTERNAL-IP      OS-IMAGE           KERNEL-VERSION           CONTAINER-RUNTIME
ip-192-168-23-188.ec2.internal   Ready    <none>   3h3m   v1.36.2   192.168.23.188   100.62.88.240    Ubuntu 26.04 LTS   7.0.0-1010-aws (amd64)   containerd://2.2.2
ip-192-168-58-220.ec2.internal   Ready    <none>   3h2m   v1.36.2   192.168.58.220   100.55.160.174   Ubuntu 26.04 LTS   7.0.0-1010-aws (amd64)   containerd://2.2.2

If your kubectl configuration is missing or incorrect for managing your cluster, you can regenerate it using the AWS CLI:

aws eks update-kubeconfig --region us-east-1 --name my-ubuntu-cluster