how to have a variable that will available to particular scheduled task whenever it runs again after some time?

Question

I am starting many scheduled jobs at the same time using rufus scheduler. I want a variable in that code that will be available to only that scheduled job, when it runs again after some time.

how can i maintain a variable for for each scheduled job?

Answer 1

there multiple ways to do that. I'll detail some of those, going from ugly to somehow elegant.

Solutions tested against Ruby 1.9.3p392 on Debian GNU/Linux 7 with rufus-scheduler 2.0.23 (https://rubygems.org/gems/rufus-scheduler).

You could use a global variable:

require 'rufus-scheduler'

scheduler = Rufus::Scheduler.start_new

scheduler.every '1s' do
  $counter0 ||= 0
  $counter0 += 1
  p [ :c0, $counter0 ]
end
scheduler.every '1s' do
  $counter1 ||= 0
  $counter1 += 1
  p [ :c0, $counter1 ]
end

scheduler.join

Or you could centralize the job variables in a single global variable (note: this isn't thread-safe):

require 'rufus-scheduler'

scheduler = Rufus::Scheduler.start_new

$job_vars = {}

scheduler.every '1s' do |job|
  ($job_vars[job.object_id] ||= {})['counter'] ||= 0
  $job_vars[job.object_id]['counter'] += 1
  p [ job.job_id, job.object_id, $job_vars[job.object_id]['counter'] ]
end
scheduler.every '1.5s' do |job|
  ($job_vars[job.object_id] ||= {})['counter'] ||= 0
  $job_vars[job.object_id]['counter'] += 1
  p [ job.job_id, job.object_id, $job_vars[job.object_id]['counter'] ]
end

scheduler.join

One step further, just for fun (but still not thread-safe):

require 'rufus-scheduler'

scheduler = Rufus::Scheduler.start_new

$job_vars = {}
job = lambda { |job|
  ($job_vars[job.object_id] ||= {})['counter'] ||= 0
  $job_vars[job.object_id]['counter'] += 1
  p [ job.job_id, job.object_id, $job_vars[job.object_id]['counter'] ]
}

scheduler.every '1s', &job
scheduler.every '1.5s', &job

scheduler.join

Finally, you could add a #vars to the Job class:

require 'rufus-scheduler'

class Rufus::Scheduler::Job
  def vars
    @vars ||= {}
  end
end

scheduler = Rufus::Scheduler.start_new

scheduler.every '1s' do |job|
  job.vars['counter'] = (job.vars['counter'] || 0) + 1
  p [ job.job_id, job.object_id, job.vars['counter'] ]
end
scheduler.every '1.5s' do |job|
  job.vars['counter'] = (job.vars['counter'] || 0) + 1
  p [ job.job_id, job.object_id, job.vars['counter'] ]
end

scheduler.join

This is the solution I prefer. I intend to add a similar bag of variables to Job in rufus-scheduler 3.0 (https://github.com/jmettraux/rufus-scheduler).

You could also put the variable somewhere else and use the job_id / job.object_id as a key to retrieve it (as the first snips of code do).

I hope this will help.