Multiple beanstalkd servers issue

[coffeebite]

Hey guys ... sorry if this sounds like a super noob question.

So I am using backburner + beanstalkd.

Beanstalk settings:

• 3 servers
• persistence on

Worker settings:

• 2 different servers
• 50 workers per server (ie. 50 different "rake backburner:work" processes, managed by god).
• Simple workers
• 8 CPUs per worker box

We consistently see a few thousand jobs in "ready" state (using beanstalkd_view) and that number does not come down no matter what I do. I have tried the following

• increasing the # of beanstalk servers
• increasing the # of workers per server. In fact, I decreased it to one server with 25 workers and the rate of completion of the jobs was the same.

The number of reserved jobs is always 1 or 0 no matter how many workers we have. Also, often, it shows all my workers are in "waiting" state.

Am I misunderstanding how backburner is supposed to work? Can I only have one worker per tube or something?

UPDATE

If I use just one beanstalkd server, things are fine.

[contentfree]

@coffeebite Since Beaneater (and therefore Backburner) no longer supports a server pool, I'm going to close this issue.

[nesquena]

Yep, pool was always half-baked anyways, makes sense.

[coffeebite]

Sorry to reopen this, but can we update the docs to reflect this? I still see this which makes me feel that I can use multiple queue servers:

[nesquena]

@coffeebite I think I have adjusted the readme now accordingly, thanks.