Skip to content

Latest commit

 

History

History
125 lines (89 loc) · 5.26 KB

kubectl-helm.md

File metadata and controls

125 lines (89 loc) · 5.26 KB
Raw

Kubectl, Helm, and Minikube Install Script

Installs latest version of kubectl, Helm, and optionally minikube. Auto-detects latest versions and installs needed dependencies.

Script status: Stable

OS support: Debian 9+, Ubuntu 18.04+, and downstream distros.

Maintainer: The VS Code and GitHub Codespaces teams

Syntax

./kubectl-helm-debian.sh [kubectl version] [Helm version] [Minikube version] [kubectl SHA256] [Helm SHA256] [minikube SHA256] [Non-root user]

Or as a feature:

"features": {
    "kubectl-helm-minikube": {
        "version": "latest",
        "helm": "latest",
        "minikube": "latest"
    }
}
Argument Feature option Default Description
kubectl version version latest Version of kubectl to install. e.g. v1.19 Use latest to install the latest stable version.
Helm version helm latest Version of Helm to install. e.g. v3.5 Use latest to install the latest stable version.
minikube version minikube none Version of Minikube to install. e.g. v1.19 Specifying none skips installation while latest installs the latest stable version.
kubectl SHA256 automatic SHA256 checksum to use to verify the kubectl download. automatic will download the checksum.
Helm SHA256 automatic SHA256 checksum to use to verify the Helm download. automatic will both acquire the checksum and do a signature verification of it.
minikube SHA256 automatic SHA256 checksum to use to verify the minikube download. automatic will download the checksum.
Non-root user automatic Specifies a user in the container other than root that will be using the desktop. A value of automatic will cause the script to check for a user called vscode, then node, codespace, and finally a user with a UID of 1000 before falling back to root.

A note on Ingress and port forwarding

When configuring Ingress for your Kubernetes cluster, note that by default Kubernetes will bind to a specific interface's IP rather than localhost or all interfaces. This is why you need to use the Kubernetes Node's IP when connecting - even if there's only one Node as in the case of Minikube. Port forwarding in Remote - Containers will allow you to specify <ip>:<port> in either the forwardPorts property or through the port forwarding UI in VS Code.

However, GitHub Codespaces does not yet support this capability, so you'll need to use kubectl to forward the port to localhost. This adds minimal overhead since everything is on the same machine. E.g.:

minikube start
minikube addons enable ingress
# Run this to forward to localhost in the background
nohup kubectl port-forward --pod-running-timeout=24h -n ingress-nginx service/ingress-nginx-controller :80 &

Usage

Feature use

You can use this script for your primary dev container by adding it to the features property in devcontainer.json.

"features": {
    "kubectl-helm-minikube": {
        "version": "latest",
        "helm": "latest",
        "minikube": "latest"
    }
}

[Optional] You may also want to enable the tini init process to handle signals and clean up Zombie processes if you do not have an alternative set up. To enable it, add the following to devcontainer.json if you are referencing an image or Dockerfile:

"runArgs": ["--init"]

Or when using Docker Compose:

services:
  your-service-here:
    # ...
    init: true
    # ...

If you have already built your development container, run the Rebuild Container command from the command palette (Ctrl/Cmd + Shift + P or F1) to pick up the change.

Script use

Usage:

  1. Add kubectl-helm-debian.sh to .devcontainer/library-scripts

  2. Add the following to your .devcontainer/Dockerfile:

    COPY library-scripts/kubectl-helm-debian.sh /tmp/library-scripts/
RUN apt-get update && bash /tmp/library-scripts/kubectl-helm-debian.sh

    ...or if you want to install minikube as well...

    COPY library-scripts/kubectl-helm-debian.sh /tmp/library-scripts/
RUN apt-get update && bash /tmp/library-scripts/kubectl-helm-debian.sh latest latest latest

  3. [Optional] You may also want to enable the tini init process to properly handle signals and ensure Zombie Processes are cleaned up if you do not have an alternative set up. To do so, add the following to .devcontainer/devcontainer.json if you are referencing an image or Dockerfile:

    "runArgs": ["--init"]

    Or if you are referencing a Docker Compose file, add this to your docker-compose.yml file instead:

    your-service-name-here:
  # ...
  init: true
  # ...

    While technically optional, `

  4. You may also want to install Docker using the docker-in-docker or docker-from-docker scripts.

That's it!