Guides

Fix "kubectl: command not found" on Linux, Mac and Windows

By Ravindra Bagale, AWS & DevOps trainer · LinkedIn · Published: 16 October 2026 · Last updated: 17 October 2026

"kubectl: command not found" means your shell cannot find a program called kubectl in any folder on your PATH. Install the official binary for your system (Linux, macOS or Windows), put it in a PATH folder such as /usr/local/bin, open a new terminal, and check with kubectl version --client. Then connect it to your cluster with aws eks update-kubeconfig, gcloud container clusters get-credentials or az aks get-credentials.

curl -LO "https://dl.k8s.io/release/$(curl -Ls https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
sudo install -o root -g root -m 0755 kubectl /usr/local/bin/kubectl
kubectl version --client
aws eks update-kubeconfig --region ap-south-1 --name lab-cluster
kubectl get nodes

A live example first

A student logs in to a fresh Amazon Linux 2023 EC2 instance to manage the class EKS cluster:

kubectl get nodes

-bash: kubectl: command not found

Check where the shell is looking:

echo $PATH
which kubectl

/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin
/usr/bin/which: no kubectl in (/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin)

So kubectl is simply not installed. Install it, prove it, connect it:

curl -LO "https://dl.k8s.io/release/v1.30.4/bin/linux/amd64/kubectl"
sudo install -o root -g root -m 0755 kubectl /usr/local/bin/kubectl
kubectl version --client

Client Version: v1.30.4
Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3

aws eks update-kubeconfig --region ap-south-1 --name lab-cluster
kubectl get nodes

NAME                                         STATUS   ROLES    AGE   VERSION
ip-10-0-1-25.ap-south-1.compute.internal     Ready    <none>   3d    v1.30.2-eks-1552ad0

Done. Friends, notice we picked v1.30 to match the EKS cluster version.

Why does the shell say command not found?

When you type a command, the shell checks each folder in PATH, from left to right, for a file with that name. If no folder has it, you get "command not found". So there are only two possible problems: kubectl is not installed, or it is installed in a folder that is not on PATH (for example, it was downloaded to ~/Downloads).

Install kubectl on your system

Linux (Ubuntu, Amazon Linux, RHEL)

1. Check your CPU type with uname -m (x86_64 = amd64, aarch64 = arm64).
2. Download the matching binary from dl.k8s.io.
3. Run sudo install -o root -g root -m 0755 kubectl /usr/local/bin/kubectl.
4. Run kubectl version --client.

On a Graviton (arm64) EC2 instance, change amd64 to arm64 in the URL, or you will see cannot execute binary file: Exec format error.

macOS

brew install kubectl
kubectl version --client

If brew itself is "not found" on an Apple Silicon Mac, add eval "$(/opt/homebrew/bin/brew shellenv)" to ~/.zprofile.

Windows 10 / 11

winget install -e --id Kubernetes.kubectl

Close PowerShell, open a new one, then:

where.exe kubectl
kubectl version --client

If you downloaded kubectl.exe by hand, put it in a folder like C:\tools\k8s and add that folder in System Properties → Environment Variables → Path.

minikube and Docker Desktop

• minikube ships its own kubectl: minikube kubectl -- get pods -A.
• Docker Desktop installs kubectl when you enable Kubernetes in Settings.

Connect kubectl to your cloud cluster

Amazon EKS

aws eks update-kubeconfig --region ap-south-1 --name lab-cluster

If you then see You must be logged in to the server (Unauthorized), your IAM user or role has no access entry in the cluster. Ask the cluster admin to add it.

Google GKE

gcloud components install gke-gcloud-auth-plugin
gcloud container clusters get-credentials lab-cluster --region asia-south1

Without the auth plugin, kubectl prints an error that gke-gcloud-auth-plugin was not found.

Azure AKS

az aks install-cli
az aks get-credentials --resource-group lab-rg --name lab-cluster

Errors you may see next

Real example: why version skew matters

The Kubernetes project officially supports kubectl within one minor version of the cluster. Amazon EKS, GKE and AKS all upgrade clusters on a regular schedule, and old clients can stop working when an API is removed. For example, Kubernetes 1.24 removed the old client.authentication.k8s.io/v1alpha1 auth API, and many teams found their old kubeconfig files broken after upgrading. Keep kubectl updated along with your cluster.

Mistakes students make (and how to avoid them)

1. Running sudo kubectl after installing in a user folder. sudo uses a different PATH and a different home folder, so it cannot find your binary or your ~/.kube/config. Install to /usr/local/bin and run kubectl without sudo.
2. Keeping the old terminal open. PATH changes only load in a new terminal. Close it and open a new one before you test.
3. Downloading the HTML page instead of the binary. If file kubectl says "HTML document", the URL was wrong. Download again with curl -LO and the exact URL.
4. Installing two versions. which -a kubectl lists every copy on PATH. Remove the old one so you always run the version you expect.
5. Sharing one admin kubeconfig. Every learner should connect with their own IAM identity, so access can be removed per person.

Summary table