| 2012-05-06 21:31:46 utc | kubicek | jmettraux: i still don't get one thing. i want to decouple launcher from workers as much as possible. i can have multiple separate launchers without workers which share the same database. I can have multiple workers as well which share their load. |
| 2012-05-06 21:31:54 utc | kubicek | jmettraux: what i don't get is how to make specialized workers (with only one or few participants implemented) |
| 2012-05-06 21:32:22 utc | jmettraux | kubicek: hello |
| 2012-05-06 21:32:29 utc | kubicek | i can use daemon-kit with ruote client |
| 2012-05-06 21:32:43 utc | kubicek | is there some other solution? |
| 2012-05-06 21:32:47 utc | kubicek | jmettraux: hello :)) |
| 2012-05-06 21:32:49 utc | jmettraux | so that would be workers that only accept work for some participants |
| 2012-05-06 21:33:28 utc | jmettraux | there is something for that, let me dig the thing |
| 2012-05-06 21:35:00 utc | jmettraux | http://ruote.rubyforge.org/implementing_participants.html#accept won't help much in that case |
| 2012-05-06 21:35:52 utc | kubicek | yes, specialized workers |
| 2012-05-06 21:36:24 utc | jmettraux | I guess that, most of the time, others have been using AMQP between ruote and the actual work |
| 2012-05-06 21:37:18 utc | kubicek | ok. i was feared this is too complex solution for my current task, but since i already have played with daemon-kit and ruote and amqp, i'll go this was |
| 2012-05-06 21:37:31 utc | kubicek | i was only wondering if i'm overseeing something |
| 2012-05-06 21:37:33 utc | kubicek | thank you |
| 2012-05-06 21:37:35 utc | jmettraux | ok, workers perform workflow work and participant work, by default they do both indiscriminately |
| 2012-05-06 21:37:53 utc | jmettraux | you could use resque or sidekick instead of amqp |
| 2012-05-06 21:39:12 utc | jmettraux | but you've given me an idea |
| 2012-05-06 21:42:07 utc | jmettraux | if you want that flexibility in ruote right now |
| 2012-05-06 21:42:27 utc | jmettraux | you could subclass Ruote::Worker |
| 2012-05-06 21:47:52 utc | jmettraux | it could look like that: https://gist.github.com/2624645 |
| 2012-05-06 21:48:42 utc | kubicek | coool! thanks a lot! i will try it :) |
| 2012-05-06 21:50:02 utc | jmettraux | https://gist.github.com/2624654 will only process participant related work |
| 2012-05-06 21:51:22 utc | jmettraux | I forgot a "not" in the first gist: https://gist.github.com/2624664 |
| 2012-05-06 21:51:44 utc | kubicek | i know :) |
| 2012-05-06 21:53:07 utc | kubicek | cool. it will be nice night :) |
| 2012-05-06 21:55:24 utc | jmettraux | night! |
| 2012-05-06 22:25:59 utc | kubicek | it works really nice |
| 2012-05-06 22:40:11 utc | jmettraux | ah great |