Run script files in background (as daemons)

[viperey]

Hi.

For some reason I'm not fully understanding the usage of the library for some cases. My use case is the following:

• I have a main process (called scheduler), this scheduler knows the location of a script that performs a certain task.
• My goal is to call such script as a completely separated script file (using the :script argument).
• The scheduler want to run several times (virtually infinitely) the script file.
• I do not need to keep track of the new(s) workers that get created, just create and run them as soon as I need them.

What's the better approach to achieve this behavior?

Example code:

require 'daemons'
require_relative 'parser.rb'

def get_config(index = 0)
  name = 'worker' + rand(0..Time.now.to_i).to_s
  script_argv_array = [
      '--my_named_arg', 'my_named_arg_value'
  ]
  daemon_argv_array = ['start', '--']
  argv_array = daemon_argv_array.concat script_argv_array
  {
      :app_name => name,
      :ARGV => argv_array,
      :dir => '/tmp',
      :dir_mode => :normal,
      :force => true,
      :mode => :load,
      :multiple => true,
      :name => name
  }
end

1.times do |index|
  config = get_config(index)
  puts("config=#{config}")
  begin
    process = Daemons.run(script = __dir__ + '/infinite_loop.rb', options = config)
    puts("process=#{process}")
  rescue Exception => e
    puts(e.to_s)
    puts(e.message)
    puts(e.backtrace)
    break
  end
  puts("Done")
end

puts("Scheduler is done")

I.e. This example code would throw an exception and child processes won't get PPID=1

Thanks in advance and congratulations for the library :)

[thuehlinger]

For your application it would probalby be easiest if you called your script just using e.g. system() and then within the script require 'daemons.rb'and then call Daemons.daemonize(). This way, the child script is detached from the parent process.

[viperey]

Oh, cool then it seems it was using the wrong design approach. Thanks for the advice!