Parse / extract server settings from your Capfile
One interesting thing that has evolved from our use of Capistrano is the configuration files have become the de-facto documentation hub for a project's server connection details. (We do maintain inventory data elsewhere, but for the developer in the trenches, config/deploy/prod.rb is the first place to look). A question arose: How to parse the settings out of these files?
Because they are in Ruby, the config files are actually little programs. They can't be parsed as such, only evaluated by the Ruby interpreter. Fortunately this is easy to do:
cap_info.rb
#!/usr/bin/env ruby
# run in a cap-managed project to get pertinent variable info
# output in YAML
require 'rubygems'
require 'capistrano/configuration'
require 'pp'
config = Capistrano::Configuration.new
config.load("Capfile")
config.logger.level = 1 # -v
for stage in config.stages do
stage_config = Capistrano::Configuration.new
stage_config.load("Capfile")
stage_config.logger.level = 1 # -v
stage_config.find_and_execute_task(stage)
puts "#{stage}:"
puts " gateway: #{stage_config[:gateway]}"
puts " user: #{stage_config[:ssh_options][:user]}"
puts " deploy_to: #{stage_config[:deploy_to]}"
puts " roles:"
for role in stage_config.roles do
puts " #{role[0]}:"
for server in role[1].servers do
puts " - #{server}"
end
end
endOutput
prod:
gateway: gateway.example.com
user: deploy
deploy_to: /var/www/sites/virtual/example
roles:
db:
- web01
web:
- web01
- web02
- web03
- web04What now?
Cap gives you a rich set of tools. Even one-off tasks are easy to accomplish with cap shell, cap invoke, or by loading extra Ruby tasks with cap -f Capfile -f extra.rb. However, a common interchange format can be useful when integrating with non-Ruby tools.
Add new comment