k3s

A k3s template lets you start sandboxes with k3s (lightweight Kubernetes) already installed and bootstrapped. k3s is a minimal, production-grade Kubernetes distribution ideal for resource-constrained environments, development, and testing.

Dockerfile

Create a Dockerfile to set up k3s in your sandbox:

FROM ubuntu:22.04

ENV DEBIAN_FRONTEND=noninteractive
ENV container=docker

RUN apt-get update && apt-get install -y --no-install-recommends \
    ca-certificates \
    curl \
    systemd systemd-sysv \
    && rm -rf /var/lib/apt/lists/*

# Download and install k3s binary
RUN mkdir -p /usr/local/bin && \
    curl -sfL https://github.com/k3s-io/k3s/releases/download/v1.35.5%2Bk3s1/k3s -o /usr/local/bin/k3s && \
    chmod +x /usr/local/bin/k3s && \
    ln -sf /usr/local/bin/k3s /usr/local/bin/kubectl && \
    mkdir -p /var/lib/rancher/k3s

# Copy k3s service file
COPY files/k3s.service /etc/systemd/system/k3s.service

# Copy bootstrap script
COPY files/init-k3s.sh /usr/local/share/k3s/init-k3s.sh
RUN chmod +x /usr/local/share/k3s/init-k3s.sh

# Copy bootstrap service
COPY files/init-k3s.service /etc/systemd/system/init-k3s.service

# Enable services by creating symlinks
RUN mkdir -p /etc/systemd/system/multi-user.target.wants && \
    ln -sf /etc/systemd/system/k3s.service /etc/systemd/system/multi-user.target.wants/k3s.service && \
    ln -sf /etc/systemd/system/init-k3s.service /etc/systemd/system/multi-user.target.wants/init-k3s.service

STOPSIGNAL SIGRTMIN+3

CMD ["/sbin/init"]

Bootstrap script

Create files/init-k3s.sh to ensure k3s is properly initialized:

#!/bin/bash
set -euo pipefail

fail() {
  printf '%s\n' "$1"
  shift
  "$@" || true
  exit 1
}

wait_for() {
  local attempts="$1"
  local delay="$2"
  shift 2

  for _ in $(seq 1 "$attempts"); do
    if "$@" >/dev/null 2>&1; then
      return 0
    fi
    sleep "$delay"
  done
  return 1
}

# Wait for k3s to be ready
if ! wait_for 60 2 test -f /etc/rancher/k3s/k3s.yaml; then
  fail "k3s kubeconfig did not appear" journalctl -u k3s --no-pager
fi

# Copy kubeconfig to root home for easy access
mkdir -p /root/.kube
cp /etc/rancher/k3s/k3s.yaml /root/.kube/config
chmod 600 /root/.kube/config

# Wait for nodes to be ready
if ! wait_for 60 2 kubectl --kubeconfig /root/.kube/config get nodes -o jsonpath='{range .items[*]}{.metadata.name}{"\n"}{end}' | grep -q .; then
  fail "k3s node did not become ready" journalctl -u k3s --no-pager
fi

k3s service file

Create files/k3s.service for k3s systemd integration:

[Unit]
Description=Lightweight Kubernetes
Documentation=https://k3s.io
Wants=network-online.target
After=network-online.target

[Service]
Type=notify
EnvironmentFile=-/etc/systemd/system/k3s.service.env
EnvironmentFile=-/etc/sysconfig/k3s
ExecStartPre=/bin/sh -c 'mkdir -p /var/lib/rancher/k3s/agent/containerd'
ExecStart=/usr/local/bin/k3s server
KillMode=process
Delegate=yes
LimitNOFILE=1048576
LimitNPROC=infinity
LimitCORE=infinity
TasksMax=infinity
TimeoutStartSec=0
Restart=always
RestartSec=5s

[Install]
WantedBy=multi-user.target

Bootstrap service

Create files/init-k3s.service to run the bootstrap script:

[Unit]
Description=Bootstrap k3s
After=k3s.service
Wants=k3s.service

[Service]
Type=oneshot
ExecStart=/usr/local/share/k3s/init-k3s.sh
RemainAfterExit=yes
StandardOutput=journal
StandardError=journal

[Install]
WantedBy=multi-user.target

Build and publish the image

Build the k3s template and push it to a registry that Leap0 can pull from:

docker build -t registry.example.com/my-org/k3s-template:v1 ./k3s
docker push registry.example.com/my-org/k3s-template:v1

Create the Leap0 template

Create a Leap0 template that points to the k3s image:

import { Leap0Client } from "leap0";

const client = new Leap0Client();

const template = await client.templates.create({
  name: "k3s",
  uri: "registry.example.com/my-org/k3s-template:v1",
});

console.log(template.id);
await client.close();

Start a sandbox with k3s

Create a sandbox from the k3s template:

const sandbox = await client.sandboxes.create({
  template_name: "k3s",
});

After the sandbox is ready, k3s will be running with kubectl available:

# Get nodes
kubectl get nodes

# Get pods in all namespaces
kubectl get pods -A

# Deploy a test pod
kubectl run test-pod --image=alpine --restart=Never -- sleep 1000

Was this page helpful?