beanstalkd

A revolutionary idea: why not use a job queue system for your job queue?

Most people use delayed job aka a database or resque/sidekiq for queuing. But why hack your way around a database or a "smarter memcache" to do a simple queue when there are other solutions that were build for this?

Beanstalkd is one of them. It's small. It's fast. It's awesome. And I show you why.

Transcript

1. submit
2. submit
3. @bitboxer @bitcrowd Bodo Tasche
4. Queues!
5. create table 'job_queue';
6. gem install delayed_job
8. beanstalkd
9. Beanstalkd is a simple, fast work queue.
10. brew install beanstalkd
11. https://kr.github.io/beanstalkd/
12. beanstalkd -p 11300
13. gem install beaneater
14. @beanstalk = Beaneater::Pool.new( [’localhost:11300’] )
15. @tube = @beanstalk.tubes['email'] @tube.put '{"key" : "foo"}', :pri => 5 @tube.put '{"key" : "bar"}', :delay => 3
16. @beanstalk.jobs.register('email') do |job| json = Oj.load(job.body) puts "job value is #{json["key"]}!" end ! @beanstalk.jobs.process!
17. @beanstalk.jobs.register( 'email', :retry_on => [SomeError], :max_retries => 2 )
18. @tube.put '{"key" : "bar"}', :ttr => 120
19. @beanstalk.jobs.register('email') do |job| # Long running stuff job.touch end
20. • ready • delayed • buried
21. job = @tube.peek(:buried)
22. job.kick
23. job.delete
24. Beaneater.configure do |config| config.default_put_delay = 0 config.default_put_pri = 65536 config.default_put_ttr = 120 end
25. @beanstalk.stats
26. @beanstalk.tubes['email'].stats
27. https://github.com/kr/beanstalkd/wiki/Tools
28. beanstalkd
29. @bitsofberlin http://bitsofberlin.org