Minikube — Notes on Running Kubernetes Locally
What it is
Minikube is the quickest way to get Kubernetes running on a laptop. It’s basically a single-node cluster in a VM or container. Good for testing, demos, or just learning the basics. Nobody uses it for production, but for training and local dev it shows up everywhere.
How it works
When you run minikube start, it spins up a small VM or container with the control plane and a worker inside. The driver decides where it runs: could be Docker, VirtualBox, Hyper-V, KVM, or others. After it boots, you talk to it with kubectl just like you would in a real cluster. Add-ons can be turned on — ingress, metrics, dashboard — so it feels closer to production, even though it’s all local.
Technical profile
Area | Details |
Main purpose | Local single-node Kubernetes |
Host platforms | Windows, macOS, Linux |
Drivers | Docker, VirtualBox, Hyper-V, KVM, containerd |
Components | Full control plane + one worker node |
Add-ons | Dashboard, ingress, metrics-server, storage-provisioner |
Control | CLI commands (minikube start, minikube addons enable) |
License | Apache 2.0 (open source) |
Resources | 2+ GB RAM, a few GB disk, CPU depending on workloads |
Setup notes
– Install a hypervisor or container runtime first (Docker is easiest).
– Download the minikube binary for your OS.
– Start it up: minikube start –driver=docker.
– Check with kubectl get nodes — should show one node “ready.”
– Turn on extras: minikube addons enable ingress.
– Stop or wipe with minikube stop / minikube delete.
Where it’s useful
– Quick Kubernetes playground on a laptop.
– Developers testing configs before sending to a team cluster.
– Workshops where every student gets their own cluster in minutes.
– CI jobs that need Kubernetes but not a full remote cluster.
Things to keep in mind
– It’s single-node, so no real HA.
– Performance tied to the laptop/desktop it runs on.
– Some add-ons behave a bit differently than in real clusters.
– Not for production — it’s a training and testing tool.
Comparison snapshot
Tool | What stands out | Best fit |
Minikube | Very easy, single-node, quick reset | Local dev, training, demos |
Kind | Runs clusters in Docker containers | CI/CD pipelines, fast ephemeral use |
MicroK8s | Snap-based, multi-node support | Ubuntu labs, edge testing |
K3s | Small binary, ARM-ready | Edge, IoT, light servers |