Configure Unraid as a Docker Swarm worker using Docker-in-Docker.
This container runs a dedicated Docker Engine inside Unraid, allowing you to join a Swarm without modifying the host Docker daemon.

Prerequisites (Unraid + macvlan):

1. In Unraid, open Settings - Docker.
2. Stop Docker, then set "Docker custom network type" to macvlan.
3. Enable a custom network on eth0 (e.g., "br0" or "eth0" depending on your setup).
4. Assign a fixed IP in the Unraid Docker template (matches the MyIP field). This IP must be
   free on your LAN and outside your DHCP pool.
5. Start Docker again.

Template configuration notes:

• Privileged: recommended for best compatibility with nested containers and features like
  overlay networking. Not strictly required, but some inner containers may fail without it.
• Hostname and /etc/hosts: update ExtraParams to set a meaningful hostname and host entry:
  --hostname='Tower' --add-host='Tower:127.0.0.1'
  Replace "Tower" with your Unraid server name or desired node name.
• ExtraParams: you can remove --gpus=1 if you do not want GPU access, or set it to a
  specific GPU UUID.

After container start (joining the Swarm):

1. On your Swarm manager node, generate a worker join token:
   docker swarm join-token worker
2. Copy the join command that it outputs. It looks like:
   docker swarm join --token TOKEN MANAGER_IP:2377
3. Enter the container console (or use docker exec) and run the join command inside this
   container.
4. Verify on the manager:
   docker node ls
   You should see this node listed as a worker.

Optional:

• If you want the inner Docker daemon to persist data, ensure the /var/lib/docker path maps
  to a stable storage path as configured in the template.
• For TLS client certs or custom daemon config, use the mapped /certs and /etc/docker paths.
• You can set the default runtime by writing /etc/docker/daemon.json in the bound config path
  (Path: Docker config). Example:
  {
  "runtimes": {
  "nvidia": {
  "args": [],
  "path": "nvidia-container-runtime"
  }
  
  }
  }