k3s fails to start — port 6443, 80, or 443 already in use
Symptom
[INFO] systemd: Starting k3s
[ERROR] listen tcp :6443: bind: address already in use
Or the k3s service starts but kubectl get nodes hangs, and ingress-nginx pods
crash-loop with bind: address already in use.
Root Cause
k3s wants:
- :6443 for the Kubernetes API server
- :80 / :443 for the bundled Traefik ingress (Traefik is disabled by our bootstrap in favour of ingress-nginx, but ingress-nginx still needs 80/443 free)
- :10250 for the kubelet
Anything already bound to those ports stops k3s from installing correctly. Common culprits:
| Port | Usually |
|---|---|
| 80/443 | apache2, nginx, caddy, docker-proxy from a previous compose stack |
| 6443 | Another k3s / kubeadm cluster left over from an earlier attempt |
| 10250 | An abandoned kubelet from a previous install |
Fix
1. Find what's using the port
sudo ss -tlnp '( sport = :80 or sport = :443 or sport = :6443 or sport = :10250 )'
You'll get a line per offending service — the last column is the process.
2. Stop or remove the conflicting service
# Common cases:
sudo systemctl disable --now apache2
sudo systemctl disable --now nginx
sudo systemctl disable --now caddy
# A stale k3s from a previous attempt:
sudo /usr/local/bin/k3s-uninstall.sh || true
# Old Docker compose stack:
docker compose down
3. Free port 6443 if it's a leftover k3s
# If :6443 is bound by a k3s process, wipe it fully
sudo /usr/local/bin/k3s-uninstall.sh 2>/dev/null || true
sudo rm -rf /etc/rancher /var/lib/rancher/k3s /var/lib/kubelet
4. Re-run the bootstrap
sudo sed -i '/^k3s=/d;/^ingress=/d' /etc/platform/.bootstrap_state
sudo ./platform-bootstrap/bootstrap.sh
The pre-flight step will fail fast now if the port is still occupied — that's the intended behaviour.
Verification
sudo ss -tlnp '( sport = :80 or sport = :443 or sport = :6443 )'
# expect: only k3s / ingress-nginx-controller
kubectl get nodes
# expect: node in Ready state