Running Kubernetes in RAM

This guide covers explanation and implementation of the cloud-init pieces that let K3s run comfortably on low-disk or tmpfs-backed roots. The focus is on what user-data sets up to keep writes light and disposable.

Explanation

The netbooted node lives mostly in RAM, so we trim anything that would otherwise fill a tiny root. Cloud-init rehomes logs to tmpfs, keeps bootstrap state in /tmp and /run, and installs a lean K3s server that only writes what it must. Manifests are staged once and copied over after K3s is ready, keeping churn low.

Implementation

Volatile journald

Start by making journald write to tmpfs right after boot so log files never hit disk.

#cloud-config (snippet)
runcmd:
  - [ mkdir, -p, /run/log/journal ]
  - [ systemctl, restart, systemd-journald ]

Bootstrap state stays in RAM

Tokens and extra args live in /tmp/k3s and /run/k3s-bootstrap, so you aren’t spraying secrets or temp files onto persistent storage.

# /usr/local/bin/k3s-20-install.sh (excerpt)
K3S_TOKEN="${K3S_TOKEN:-my-cluster-token}"
K3S_COMMON_ARGS="${K3S_COMMON_ARGS:-}"
K3S_SERVER_ARGS="${K3S_SERVER_ARGS:- --cluster-init --node-label=kidd.network/role=control-plane --token ${K3S_TOKEN}}"
ARGS_DIR="/run/k3s-bootstrap"
if [ -f "$ARGS_DIR/K3S_COMMON_ARGS" ]; then
  K3S_COMMON_ARGS="$K3S_COMMON_ARGS $(tr '\n' ' ' < "$ARGS_DIR/K3S_COMMON_ARGS")"
fi
if [ -f "$ARGS_DIR/K3S_SERVER_ARGS" ]; then
  K3S_SERVER_ARGS="$K3S_SERVER_ARGS $(tr '\n' ' ' < "$ARGS_DIR/K3S_SERVER_ARGS")"
fi

Minimal server install

The installer uses only the flags it needs; anything extra can be layered in through the tmpfs arg files if you want to experiment without baking new images.

# /usr/local/bin/k3s-20-install.sh (excerpt)
INSTALL_K3S_VERSION="v1.33.6+k3s1"
K3S_ROLE="server"
curl -sfL https://get.k3s.io | \
  K3S_TOKEN="$K3S_TOKEN" \
  INSTALL_K3S_EXEC=" $K3S_COMMON_ARGS $K3S_SERVER_ARGS" \
  INSTALL_K3S_VERSION="$INSTALL_K3S_VERSION" \
  sh -s -

One-shot manifest staging

Copy staged YAML once and be done with it. It’s enough to get essentials onto the node without creating ongoing writes.

src="/etc/k3s-manifests"
dst="/var/lib/rancher/k3s/server/manifests"
for f in "$src"/*.yaml "$src"/*.yml; do
  b=$(basename "$f")
  install -D -m0644 "$f" "$dst/$b"
done

# /etc/systemd/system/k3s-ready.path (excerpt)
[Path]
PathExists=/etc/rancher/k3s/k3s.yaml
PathExists=/var/lib/rancher/k3s/server/manifests
Unit=k3s-ready-dispatch.service

Next Steps

Keep /etc/k3s-manifests lean to minimize writes during boot.
If you need persistent storage, mount it explicitly (e.g., NVMe for Longhorn) and leave the root ephemeral.
Move to deploying-k3s-on-netboot-nodes.md to tie this into the full bootstrap flow.