Prerequisites
- A running UCP cluster with at least one manager node
- A new host with a supported Linux distribution (e.g., Ubuntu, CentOS)
- SSH access and root or sudo privileges on the new host
- Network connectivity on ports 2376 (Docker daemon) and 2377 (Swarm)
Ensure the Docker Engine version on your new worker matches the UCP manager’s version to prevent compatibility issues.
1. Provisioning and Installing Docker Engine
On the new worker node, install Docker Engine using Docker’s official convenience script:2. Joining the Swarm Cluster
First, retrieve the worker join token on your UCP manager:Keep the join token confidential; anyone with access can join your Swarm cluster as a worker.
docker swarm join command and execute it on the new worker:
3. Automatic UCP Agent Deployment
UCP configures its agent as a global service in Docker Swarm. After the worker joins:- Docker Swarm schedules the UCP agent on the new node.
- The agent pulls and installs UCP components (e.g., UCP proxy).
- The node is automatically registered with the UCP control plane.
4. Verification in UCP Console
- Open the UCP web UI on your manager node.
- Navigate to Nodes in the sidebar.
- Confirm the new worker appears with the Worker role and a Ready status.
| Task | Host | Command / UI Action |
|---|---|---|
| Install Docker | Worker node | curl -fsSL https://get.docker.com | shsudo systemctl enable dockersudo systemctl start docker |
| Retrieve join token | Manager node | docker swarm join-token worker |
| Join the swarm | Worker node | docker swarm join --token <token> <manager-ip>:2377 |
| Verify in UCP | Manager Console | Go to Nodes and check for the new worker |