Added QUIT for unicorn master (old)

[prognostikos]

Don't know if this is useful for the main project, but I've added the ability for the restart task to automatically send QUIT to the old master process after sending USR2 to create a new master. This should close #20

[rauchy]

+1 for this. Helped me get things working.

[dkubb]

This commit seems to be causing problems for me when I have a unicorn process running and I execute "cap deploy".

What seems to be happening is the USR2 signal is sent, a new process starts and it dies trying to bind to a unix socket with an Errno::EADDRINUSE (Address already in use) error, then the QUIT signals is sent to the original process. In the end nothing is running.

I've experimented with long delays and infinite tries for the listen command in my unicorn config, which does not have any effect.

The only thing that seems to work is if I use the commit before this (747daa7), and add the following to the unicorn config's before_hook block:

old_pid = "#{server.config[:pid]}.oldbin"                                         
if File.exists?(old_pid) && old_pid != server.pid                                 
  begin                                                                           
    sig = worker.nr.succ >= server.worker_processes ? :QUIT : :TTOU               
    Process.kill(sig, File.read(old_pid).to_i)                                    
  rescue Errno::ENOENT, Errno::ESRCH                                              
  end                                                                             
end

Note that I have experimented with adding and removing this code with this latest commit, but neither have worked.

[sosedoff]

Im aware of this, working on refactoring. Thanks for the feedback.