The Problem: I only need the Starbound server to run maybe 8 hours a month. Paying $40 a month for this usage level is a waste of resources. I have considered merging this into my Docker host but for now I will keep them separate. Because of this I require a method to backup and destroy my Droplets, then restore them when they are needed.

Solution: In order to solve this, I'm leveraging DigitalOceans snapshots and destroying the server when the snapshot is complete. This will reduce my usage charges to what I actually use, and provide a versioned history of the server should something go wrong.

• Koalalorenzo's python-digitalocean
• If using Vagrant, requires Vagrant 1.8 for ansible_local provisioner.

It should be noted that this script only targets Droplets named starbound-server.

• From Windows (or using Vagrant in general): A Vagrantfile and provisioner have both been supplied should you decide not to directly install any necessary packages.

  1. Run vagrant up from this folder.
  2. At the time of writing, a bug exists that will cause ansible_local to fail, once vagrant up fails, run vagrant reload --provision and it should pass. (Github Issue)
  3. Run vagrant ssh and you will be able to run the script (located in do_starbound). Options for usage are listed below.

• Environment Variables: You may supply a DigitalOcean API Key at runtime or set it to an environment variable accessible to the script. Something like export DO_API_KEY=mykey in ~/.bashrc should allow the script to autoload the key for usage.

• Command line: There are currently 3 Command line options. --backup, --destroy, and --restore.

  • --backup: Run the snapshot against the starbound-server Droplet.
  • --destroy: This should be used along with --backup. If provided, the droplet will be destroyed after it is backed up (This provides the cost savings).
  • --restore: Run this one on its own, It will use the latest starbound-server snapshot to create a new starbound-server droplet.