This is a plugin designed to help integrate Chef-Zero into a Vagrant run, similar to Berkshelf. Chef-Zero will be started on the host machine and populated with the specified data. When the Vagrant machine is destroyed Chef-Zero will be killed.

Note: Only NIX systems supported at this time due to a call to lsof to find running Chef Zero servers.

vagrant plugin install vagrant-chef-zero

Currently only tested on NIX systems, though I believe most NIX only commands have been removed.

As for Vagrant providers, it has only been tested with VirtualBox. It should be possible to add support for other providers by changing certain assumptions, such as always running Chef-Zero on the host machine.

Just set your Vagrant provisioner to :chef_client. If you wish to use :chef_client without using vagrant-chef-zero just set config.chef_zero.enabled = false inside your Vagrantfile configuration block.

Inside of your Vagrant file you will have access to the following configuration options:

config.chef_zero.nodes = "../foobar/nodes"
config.chef_zero.environments = "../foobar/environments/baz.json"
config.chef_zero.data_bags = "../foobar/data_bags"
config.chef_zero.roles = "../foobar/roles/*.json"
config.chef_zero.cookbooks = "spec/fixtures/cookbooks"

Alternatively, you can use chef_repo_path and it will attempt to intelligently find the appropriate sub directories:

config.chef_zero.chef_repo_path = "../foobar/my_repo/"
# This implies
# config.chef_zero.nodes = "../foobar/nodes"
# config.chef_zero.environments = "../foobar/environments"
# config.chef_zero.data_bags = "../foobar/data_bags"
# config.chef_zero.roles = "../foobar/roles"
# config.chef_zero.cookbooks = "../foobar/coobooks"

# Failure to find any of these is ok, it will just be ignored.

As Vagrant is booting up, vagrant-chef-zero will search each specified location for files to upload to Chef-Zero. The upload will be done via Ridley APIs.

Check out the included Vagrantfile for example usage.

These are uploaded via the Ridley gem. It is the same backend that Berkshelf uses, though I am sure my usage is not as complete. It expects a path to a cookbooks/ directory, or an array of paths to individual cookbooks. If you omit this field, you can have Berkshelf upload cookbooks as usual (It will find the Chef-Zero URL automatically).

Currently only JSON files are supported as we do not have the Chef libraries to serialize .rb files.

When Chef-Zero starts, it will write out zero-knife.rb to your current directory. This will be a valid Knife configuration file with the following fields

chef_server_url "foo:bar"
node_name "baz"
client_key "/path/to/tmp/key.pem"

If you wish to manipulate the Chef Zero server manually, you can pass this config file to Knife with -c .zero-knife.rb

If for some reason you are unable to have Vagrant destroy the Chef-Zero server, you can find the PID it by running lsof -i tcp:#{port} where port is 4000 by default.

None! vagrant-chef-zero performs some ruby black magic to generate a fake RSA key and patch it in to the chef-client configuration. You should not need to worry about providing validation_key_path.

You don't need to specify one! If no url is specified, Vagrant-Chef-Zero will find your local IPv4 address and bind to it, on port 4000. If you specify a Chef Server url, Vagrant-Chef-Zero will try to parse it and bind appropriately.

Not required. As Chef-Zero does no authentication we can fake this. If it is left unspecified we will use a default value of dummy-validator.

• Tom Duffield (@tduffield)
• Ben Dean (@b-dean)
• Jesse Nelson (@spheromak)
• Mark Cornick (@markcornick)
• Jesse Adams (@jesseadams)
• Andrew Havens (@andrewhavens)
• Bao Nguyen (@sysbot)
• Miguel Ferreira (@miguelaferreira)
• Timothy van Zadelhoff (@j00p34)