Rufus Scheduler doesn't run with daemonized puma 2.0.0.b4

[concept47]

I'm new to puma and I love it!
(Went from 500-600M of memory usage to 150M ... that's hard to beat)

I got 1.6.3 working using an Apache proxy yesterday and it seemed to run just fine with this gem https://github.com/jmettraux/rufus-scheduler. My rufus scheduler tasks got scheduled and run just fine.

To take advantage of the daemonize option in the new version, I switched to 2.0.0.b4, but in that instance, Rufus scheduler doesn't seem to run at all, it starts up but then does nothing. Does just fine when I run it from ssh in the fore-ground.

Any ideas why it would work this way.
Thanks!

[evanphx]

Do the workers start up? What config did you use to test it?

[concept47]

Here is the schedule.rb I'm running from my config/initializers directory (I've snipped out most of the jobs except a printout of "I"VE Started" that I'm using to test this. Its the same one I've used for years, and it works just fine when puma isn't daemonized.

I also think it might be related to this mongrel bug with rufus scheduler https://groups.google.com/forum/?fromgroups=#!topic/rufus-ruby/8ToQAH8-Rv0

require 'rufus/scheduler'
##check for console or rake before proceding, this stops all of the scheduling from running when we're at the console or running a rake task
if !defined?(IRB) && !defined?(Rake)
   #setup ---------------------------------------------------------------
    scheduler = Rufus::Scheduler::PlainScheduler.start_new

    logfile = File.open(Rails.root.join('log', 'scheduler.log'), 'a')
    logfile.sync = true
    $schedule_logger = Logger.new(logfile, 2, 1024000)
    $schedule_logger.info "------------------------------------------------------------------------"
    $schedule_logger.info "------------------------------------------------------------------------"
    $schedule_logger.info "[Scheduler] started at #{Time.now.strftime('%B %d, %Y at %I:%M %p')}\n\n"

    def scheduler.handle_exception (job, exception)
        $schedule_logger.error "------------------------------------------------------------------------"
        $schedule_logger.error "[Scheduler] Exception at #{Time.now.strftime('%B %d, %Y at %I:%M %p')}"
        $schedule_logger.error "job #{job.job_id} running every #{job.t} caught exception"
        $schedule_logger.error "'#{exception}'"
        $schedule_logger.error "#{exception.backtrace.join("\n")}\n\n"
   end

#schedules ------------------------------------------------------------
## FOR DEBUGGING STUFF
    scheduler.in '5s' do ## same logic here, put it as far away from start as we can
        $schedule_logger.info "I STARTED!!! #{Time.now.strftime('%B %d, %Y at %I:%M %p')}"
    end

"[Scheduler] started at ... " never prints out to the log

[concept47]

Hey Evan, is there a way I can try this out?

[evanphx]

Use puma from git in your Gemfile, then you can try it out.

[concept47]

IT WORKS!!!! WOOHOOOO!!!!

[evanphx]

WOOO!

[concept47]

Hi Evan,
Somehow this stopped working in 2.2.1+ :\

[evanphx]

@concept47 please open a new ticket.

[concept47]

Sorry about that. it was fixed in 2.3.2 so not need.
Thanks Evan!

[gamead]

It looks like Rufus Scheduler doesn't run with daemonized puma again,I'm in version 2.8.1,any help?

[djoos]

FYI: same issue thing on 2.9.0.

Temp workaround: omit the --daemon flag and use nohup instead? #sneaky

HTH!

[puzanov]

Still not working with Puma 2.9.1
Rufus version is 3.0.7

[puzanov]

Rufus 3.0.9 is also not working with Puma 2.9.1
Please help!

[puzanov]

Anyone?

[evanphx]

@puzanov What configuration are you running puma in? Are you using workers and preload?

[evanphx]

The problem is, mostly, that Rufus scheduler starts up threads at booted and they don't carry over properly.

[puzanov]

How to fix that?

On Thu, Oct 16, 2014 at 10:03 PM, Evan Phoenix notifications@github.com
wrote:

The problem is, mostly, that Rufus scheduler starts up threads at booted
and they don't carry over properly.

—
Reply to this email directly or view it on GitHub
#183 (comment).