Configuration Version
Configuration versions are the mechanism by which Vagrant 1.1+ is able to remain backwards compatible with Vagrant 1.0.x Vagrantfiles, while introducing dramatically new features and configuration options.
If you run vagrant init
today, the Vagrantfile will be in roughly the
following format:
Vagrant.configure("2") do |config| # ...end
The "2"
in the first line above represents the version of the configuration
object config
that will be used for configuration for that block (the
section between the do
and the end
). This object can be very
different from version to version.
Currently, there are only two supported versions: "1" and "2". Version 1 represents the configuration from Vagrant 1.0.x. "2" represents the configuration for 1.1+ leading up to 2.0.x.
When loading Vagrantfiles, Vagrant uses the proper configuration object for each version, and properly merges them, just like any other configuration.
The important thing to understand as a general user of Vagrant is that
within a single configuration section, only a single version can be used.
You cannot use the new config.vm.provider
configurations in a version 1
configuration section. Likewise, config.vm.forward_port
will not work
in a version 2 configuration section (it was renamed).
If you want, you can mix and match multiple configuration versions in the same Vagrantfile. This is useful if you found some useful configuration snippet or something that you want to use. Example:
Vagrant.configure("1") do |config| # v1 configs...end Vagrant.configure("2") do |config| # v2 configs...end
What is Vagrant::Config.run
? You may see this in Vagrantfiles. This was actually how Vagrant 1.0.x
did configuration. In Vagrant 1.1+, this is synonymous with Vagrant.configure("1")
.