Comment 2 for bug 1693847

If we consider kolla to be the project for container image building, probably a virtualenv is much better if you don't want to pollute your environment.

Vagrant really shines when we want to setup a development environment for the deployment tools.

So I'm thinking we should do it the other way around: have the vagrant environment in kolla-ansible (or any other deployment tool that consumes kolla images) and have it depend on kolla so we retain the ability to build images. I believe this was the original intention after the repo split - the doc was moved to kolla-ansible - but the code was duplicated, making the vagrant env effectively broken in both projects.