Firing up a docker swarm cluster with one command!

This project aims to be a simple and fast solution to create and configure a whole Docker Swarm cluster. without user interaction as several virtual machines.

VirtualBox, Vagrant and Ansible have to be installed before you can create the Docker swarm cluster

After installing the previous mentioned dependencies, the cluster can be started as follows.

# Start the creation and configuration of the cluster
vagrant up --provision
# Check the current state of the VMs
vagrant status

On each VM is a Docker Engine running in "Swarm Mode". There are Swarm Managers and Swarm Workers. The Managers have IPs like 10.0.10.{i} and hostnames like manager-{i} (where {i} is the number of the manager). The workers have IPs like 10.0.11.{i} and hostnames like worker-{i}. All nodes (managers and workers) are in the same IPv4 subnet 10.0.0.1/16.

To SSH into a node, use vagrant ssh. E.g. vagrant ssh manager-1

There is also Avahi/mDNS installed on every VM. Therefore hostnames like manager-{i}.local will be resolved to the correct IPs of the VM and can be used by the host system, if avahi is also installed on the host.

The swarm manager manager-1 is the leading manager of the swarm cluster. On this node is also Portainer, a management UI for Docker, installed. It can be reached at http://manager-1.local:9000/ through a webbroser. Portainer is also capable of managing the Docker swarm as a whole, instead of only the Docker engine on manager-1.

Bug reports, feature requests, pull requests and feedback in general is always welcome!

MIT
Copyright © 2017 Simon Schürg