Up until this point, we used Docker Machine to create servers locally and join them into a cluster. The intention was to teach you the fundamentals of creating and operating a Swarm cluster without spending money on hosting providers. Now that we reached the point where you are comfortable with how Docker Swarm Mode works, the time has come to move to "real" servers. We’ll continue being "cheap" by using small instances that are either free or very inexpensive, and create just enough servers to demonstrate the process. The goal of the chapters that follow will be to walk you through a few setups, compare them, and choose the one we'll apply to our production. The only things you should change are VM instance types and the number of servers. Everything else can be the same as in the examples we’ll work with.
We already saw how to accomplish fault tolerance by utilizing Docker Swarm as the service scheduler. The chapters that follow will try to achieve the required speed and automation...