Re: Different configuration for dev/test/prod

On Dec 19, 7:29 am, Gregory Seidman <gsslist
+capistr...-dNXPQ6k9rNiG6BJUYyje5axOck334EZe@xxxxxxxxxxxxxxxx> wrote:
> I'm trying to figure out the best approach to having different
> configurations for when I'm deploying to a development machine (just a
> single mongrel running from a directory in my home directory), a production
> machine (service script via sudo to start mongrel cluster, deployed to a
> directory under /var/www, maybe multiple machines), or a test machine (same
> as prod, but a different machine).
>
> I don't think roles are the right approach. I could have different config
> files and load the right file based on a -S variable or with a -F flag, but
> that's kind of ugly.
>
> What I'd like is to have a dev, test, and prod namespace in which all of
> the deploy namespace is included, with overrides as necessary. Then I'd cap
> dev:deploy or cap test:deploy:restart or the like. Is there a way to import
> one namespace into another?
>
> --Greg

I have staging environment, not development, but same should work:

task :staging do
  role :app, 'dev.machine'
  role :db, 'dev.machine'
  set :mongrel_port, '3000'
  set :mongrel_ip, '0.0.0.0'
  set :rails_env, 'staging'
  common_stuff
end

task :production do
  role :app, 'production.app.machine'
  role :db, 'production.db.machine'
  set :mongrel_port, '9000'
  set :mongrel_ip, '127.0.0.1'
  set :rails_env, 'production'
  common_stuff
end

task :common_stuff do
  namespace :deploy do
    task :start, :roles => :app do
      run "mongrel_rails start -d -p #{ mongrel_port } -a
#{ mongrel_ip } -e #{ rails_env }"
    end
    # ...
  end
end

Run like:
  $ cap production deploy:start
  $ cap staging deploy:restart
  $ etc