Puma auto startup workflow

# in config/puma.rb
app = "cpp" # App-specific
root = "/home/deployer/apps/#{app}"

workers  5
threads  1, 1
#preload_app!

rackup      DefaultRackup
#port        9292
environment ENV['RACK_ENV'] || 'production'
daemonize   true

# directory "#{root}/current"
pidfile "#{root}/puma/puma.pid"
stdout_redirect "#{root}/puma/puma.log", "#{root}/puma/puma_error.log"
bind "unix:/tmp/puma.socket"
#ssl_bind '127.0.0.1', '9292', { key: "/etc/ssl/server.key", cert: "/etc/ssl/cpp.chained.crt" }
activate_control_app

on_worker_boot do
  # worker specific setup
  # ActiveRecord::Base.establish_connection(config)
end

============================================

# Key insight at http://dev.mensfeld.pl/2014/02/puma-jungle-script-fully-working-with-rvm-and-pumactl/

1. Deploy with mina as usual, get puma to run with `mina boot` and `mina deploy`
2. Make the three files as described here https://github.com/puma/puma/tree/master/tools/jungle/init.d
3. Populate the files thusly
  # /etc/init.d/puma | https://gist.github.com/Epigene/c7fa821d1babd429c7be # NB put real secret in export SECRET_KEY_BASE=32560bb
  # /usr/local/bin/run-puma | https://gist.github.com/Epigene/637a295bcea88bcedca3
  # /etc/puma.conf | line for every app, if one app, the line is "/home/deployer/apps/<app>/current,deployer"
3. Add puma folder in project /tmp