2

I've been using Vagrant for some time and am now exploring Docker. I've always loved Vagrant for its simplicity. Right now, I'm trying to use the Docker provider within Vagrant. I'm using the following Vagrantfile:

Vagrant.configure("2") do |config|
  config.vm.provider "docker" do |d|
    d.image = "fgrehm/vagrant-ubuntu:precise"
  end
end

My understanding is that I can just run vagrant up. I can then run the Docker container using vagrant docker-run -- <command>.

So far so good. What makes Docker so awesome is the fact that you can mess around and commit changes. I do not understand is how to incorporate this in my workflow when using the Docker provider for Vagrant. E.g. how do I run docker commit to commit the state of the container? I'd expect some kind of vagrant docker-commit, but this does not exist?


Edit: In hindsight, I think this is not the way you should be using Vagrant/Docker. Though it is claimed that both tools complement each other, my opinion is that they do not (yet) play really well together. Right now, we're using Dockerfiles to build our images. Additionally, we made a set of bash scripts to launch a container.

Community
  • 1
  • 1
Martijn
  • 5,491
  • 4
  • 33
  • 41

2 Answers2

1

Try to specify "has_ssh" to true, and ask Vagrant to use port 22 instead of 2222:

ENV["VAGRANT_DEFAULT_PROVIDER"] = "docker"

Vagrant.configure("2") do |config|
  config.vm.provider "docker" do |d, o|
    d.image = "fgrehm/vagrant-ubuntu:precise"
    d.has_ssh = true
    o.ssh.port = 22
  end
end

Then use vagrant up; vagrant ssh to access it.

Gea-Suan Lin
  • 598
  • 7
  • 14
  • The problem with that is you need a running SSH-Service. Unless your container is only an SSH-Service this will break the "One process in one container" guideline. In addition you'll need a supervisor to run SSHD + Something. In other words: Most likely you are building something like a mini VM and not a Docker container for one application which was the initial intention. On the other side in history technologies were often used in different ways than the intention initially was. – Yser Nov 26 '14 at 15:27
0

Another option if you have same client version of docker at your host-host machine as docker server started inside boot2docker VM. In that case you can set

export DOCKER_HOST=tcp://:4243 (or 2375 depending on docker version)

and then access all docker features running local client:

docker ps -a
Anaderi
  • 775
  • 5
  • 8