Installing Docker UCP

We need to set a few environment variables, as follows:

$ export UCP_IP=<IP address>
$ export UCP_FQDN=<FQDN>
$ export UCP_VERSION=3.0.0-beta2

Here <IP address> and <FQDN> are the public IP address and the public DNS name of the AWS EC2 instance we're installing in UCP.

After that, we can use the following command to download all the images that UCP needs:

$ docker run --rm docker/ucp:${UCP_VERSION} images --list \
    | xargs -L 1 docker pull

Finally, we can install UCP:

Installing UCP 3.0.0-beta2 on a VM in AWS

Now we can open a browser window and navigate to https://<IP address>. Log in with your username, admin, and password, adminadmin. When asked for the license, upload your license key or follow the link to procure a trial license.

Once logged in, on the left-hand side under the Shared Resources section, select Nodes and then click on the Add Node button:

Adding a new node to UCP

In the subsequent Add Node dialog box, make sure that the node type is Linux and the node role, Worker is selected. Then copy the docker swarm join command at the bottom of the dialog box. SSH into the other two VMs you created and run this command to have the respective node join the Docker Swarm as a worker node:

Joining a node as a worker to the UCP cluster

Back in the web UI of UCP, you should see that we now have three nodes ready, as shown here:

List of nodes in the UCP cluster

By default, worker nodes are configured so that they can only run the Docker Swarm work load. This can be changed in the node details though. In this, three settings are possible—Swarm only, Kubernetes only, or mixed workload. Let's start with Docker Swarm as the orchestration engine and deploy our pets application.