| 2011-02-10 07:31:57 utc | jmettraux | kennethkalmer: hello, http://www.quora.com/What-are-the-advantages-Riak-has-over-CouchBase/answer/Justin-Sheehy |
| 2011-02-10 08:55:11 utc | jmettraux | florianhanke: hello, congrats on the geek talk |
| 2011-02-10 09:59:34 utc | florianhanke | He jmettraux, thanks a lot! :) |
| 2011-02-10 13:42:02 utc | riddley | jmettraux, ! |
| 2011-02-10 13:42:12 utc | jmettraux | riddley: hello ! |
| 2011-02-10 13:42:14 utc | riddley | :) |
| 2011-02-10 13:42:21 utc | riddley | our time zones make it difficult to catch you |
| 2011-02-10 13:42:34 utc | jmettraux | :-) how are you doing ? |
| 2011-02-10 13:42:39 utc | riddley | I have some questions is you have a moment. I'm doing well! You? |
| 2011-02-10 13:42:54 utc | jmettraux | doing fine, ready for your questions |
| 2011-02-10 13:45:19 utc | riddley | would you say that ruby 1.8 is just memory-leaky and there's nothing to be done about it? I ask because we've been trying to write a daemon with ruby-ee (which is 1.8) and we could not stop it from leaking (badly!). At one point we pared it down to something this simple: https://gist.github.com/820524 and even that leaked memory really quickly. We ended up landing on 1.9.2 which seems to have solved it, but I feel like it must have been us not knowing som |
| 2011-02-10 13:45:19 utc | riddley | e common pitfall or something. |
| 2011-02-10 13:46:08 utc | jmettraux | whoah, I'm not that good with leaks |
| 2011-02-10 13:46:26 utc | jmettraux | how does it behave with a vanilla japanese 1.8.7 ? |
| 2011-02-10 13:47:16 utc | riddley | we didn't try that 1.8.7... we went from RHEL5's 1.8.5 -> ruby-ee 1.8.7 -> jruby (800M res!? are you kidding?!) and then landed on 1.9.2 which seems to have working GC. |
| 2011-02-10 13:47:41 utc | riddley | our real code would last around 12 hours and then it would get killed becuase the 2G VM was OOM. |
| 2011-02-10 13:47:54 utc | jmettraux | ouch |
| 2011-02-10 13:48:42 utc | jmettraux | you could replace the loop { sleep 10 } with @scheduler.join |
| 2011-02-10 13:48:42 utc | riddley | well thanks for listening, you're the wisest Rubyist I know, so I figured maybe we just didn't know that you never use X when doing Y or something :) |
| 2011-02-10 13:48:50 utc | riddley | aah ok! |
| 2011-02-10 13:48:52 utc | riddley | good tip :) |
| 2011-02-10 13:50:06 utc | jmettraux | but that wouldn't solve the issue |
| 2011-02-10 13:50:14 utc | jmettraux | how was the 1.8.5 performing ? |
| 2011-02-10 13:50:34 utc | riddley | about the same as ruby-ee in terms of memory... shocked us. |
| 2011-02-10 13:51:28 utc | jmettraux | is this really that script that goes ballistic or is the real script a bit more "interesting" ? |
| 2011-02-10 13:52:10 utc | riddley | that script does leak and somewhat quickly, but the real script is more interesting and dramatic in its leaks in 1.8 |
| 2011-02-10 13:52:28 utc | riddley | its code is here: https://github.com/riddley/yaketystats/tree/ruby_collector/collector please be kind :) we're not traditional rubyists |
| 2011-02-10 13:53:21 utc | jmettraux | riddley: ok ok, I'll be kind if you stop thinking of me as a "wise rubyist" ;-) |
| 2011-02-10 13:53:58 utc | riddley | hehe but you're my only hope! :) |
| 2011-02-10 13:54:15 utc | jmettraux | riddley : where is the script ? (BTW, I love the concept of "plugout") |
| 2011-02-10 13:54:55 utc | riddley | plugouts are written in non-ruby langs... bin/collector calls lib/boot.rb which creates an instance of lib/collector.rb |
| 2011-02-10 13:56:10 utc | riddley | I have a meeting in 4 minutes, I have to go, I'm sorry! Thanks so much for your time. I really appreciate it. If you have time to look over the code and give some feedback that would be amazing. If not, that's cool too. Thanks again. |
| 2011-02-10 13:56:28 utc | jmettraux | look at perftools |
| 2011-02-10 13:56:33 utc | jmettraux | ObjectSpace |
| 2011-02-10 14:01:26 utc | hallelujah | Hello :) |
| 2011-02-10 14:01:43 utc | jmettraux | hallelujah: hello |
| 2011-02-10 14:01:49 utc | hallelujah | hello jmettraux |
| 2011-02-10 14:02:15 utc | hallelujah | I think I have found a bug with daemon-kit |
| 2011-02-10 14:02:24 utc | jmettraux | riddley: if you could mail me some more details, I'll try to help you |
| 2011-02-10 14:02:35 utc | hallelujah | with ruote deployment |
| 2011-02-10 14:02:50 utc | jmettraux | woah, not sure if I can help |
| 2011-02-10 14:03:04 utc | jmettraux | I'm clueless when it comes to daemon-kit |
| 2011-02-10 14:03:14 utc | hallelujah | kennethkalmer is not gone |
| 2011-02-10 14:03:34 utc | hallelujah | hmm maybe it is a amqp issue but i am not sure |
| 2011-02-10 14:03:40 utc | jmettraux | he's not here right now |
| 2011-02-10 14:04:38 utc | jmettraux | if you can summarize your issue, even better "package" it in a script with some explanations, we can surely sort it out with you |
| 2011-02-10 14:04:59 utc | hallelujah | Ok will have to do that |
| 2011-02-10 14:05:06 utc | hallelujah | and paste you that gist |
| 2011-02-10 14:05:30 utc | jmettraux | if there is no reply here, please use the mailing list |
| 2011-02-10 14:22:32 utc | hallelujah | jmettraux: Ok thanks, after googling it is an amqp bug ... due to rabbitmq upgrade |
| 2011-02-10 14:45:42 utc | jmettraux | hallelujah: well done ! |
| 2011-02-10 14:46:36 utc | hallelujah | I've created a ticket here : https://github.com/kennethkalmer/daemon-kit/issues/issue/34 |
| 2011-02-10 14:47:06 utc | hallelujah | jmettraux: thanks for your help |
| 2011-02-10 14:47:16 utc | jmettraux | I did nothing ;-) |
| 2011-02-10 14:47:55 utc | jmettraux | hallelujah: are you based in France ? |
| 2011-02-10 14:48:05 utc | hallelujah | yes i am |
| 2011-02-10 14:48:41 utc | hallelujah | So are you planing a trip here ? |
| 2011-02-10 14:48:48 utc | jmettraux | if it's not too much to ask, what are you [planning to] use(ing) ruote for ? |
| 2011-02-10 14:49:15 utc | jmettraux | hallelujah: no sorry, no trips for me in France, too many strikes |
| 2011-02-10 14:49:41 utc | hallelujah | :-) First of all, just to keep an eye with a workflow engine written in ruby |
| 2011-02-10 14:49:49 utc | hallelujah | on* (sorry) |
| 2011-02-10 14:50:32 utc | hallelujah | I have some distributed and very complicated tasks with my projects |
| 2011-02-10 14:51:04 utc | hallelujah | There so much scripts that depends eahc other |
| 2011-02-10 14:51:13 utc | hallelujah | but I would like an abstraction |
| 2011-02-10 14:52:12 utc | jmettraux | may I say "orchestration of scripts" ? |
| 2011-02-10 14:52:18 utc | hallelujah | Yep |
| 2011-02-10 14:52:43 utc | hallelujah | Just an example |
| 2011-02-10 14:52:58 utc | hallelujah | I have dozens of gzipped xml files |
| 2011-02-10 14:53:10 utc | hallelujah | those weight ~ 1 Go per file |
| 2011-02-10 14:53:28 utc | hallelujah | I have to cut, parse etc ... |
| 2011-02-10 14:54:11 utc | hallelujah | some scripts will copy on servers |
| 2011-02-10 14:54:49 utc | hallelujah | after copied, the server will read and feed a database |
| 2011-02-10 14:54:57 utc | hallelujah | on each 4-5 servers |
| 2011-02-10 14:55:02 utc | hallelujah | and so on ... |
| 2011-02-10 14:55:31 utc | hallelujah | I am looking for a way to tell which server to launch which scripts |
| 2011-02-10 14:55:53 utc | jmettraux | ah, the imperative part |
| 2011-02-10 14:57:30 utc | jmettraux | thanks ! |
| 2011-02-10 15:01:00 utc | hallelujah | jmettraux, do you have some knowledge of ruote-amqp ? |
| 2011-02-10 15:01:30 utc | jmettraux | I am the maintainer by default |
| 2011-02-10 15:01:46 utc | hallelujah | ah ok :) |
| 2011-02-10 15:02:21 utc | jmettraux | https://github.com/kennethkalmer/ruote-amqp/commits/ruote2.1/ |
| 2011-02-10 15:02:41 utc | hallelujah | I don't understand why it is mandatory to set :reply_queue on amqp participant in the process_definition |
| 2011-02-10 15:02:58 utc | jmettraux | me neither, I hate this |
| 2011-02-10 15:03:19 utc | hallelujah | It would be much nicer to set it on registration |
| 2011-02-10 15:03:38 utc | hallelujah | by default and then we can override this on definition |
| 2011-02-10 15:05:05 utc | jmettraux | wait, isn't it already like that ? https://github.com/kennethkalmer/ruote-amqp/blob/master/lib/ruote-amqp/participant.rb#L207-210 |
| 2011-02-10 15:05:53 utc | jmettraux | well, this is master |
| 2011-02-10 15:05:56 utc | hallelujah | no ... it is't |
| 2011-02-10 15:06:05 utc | hallelujah | even with master |
| 2011-02-10 15:06:31 utc | hallelujah | ah uh |
| 2011-02-10 15:06:43 utc | hallelujah | ok found the problem, it is that daemon-kit stuff again |
| 2011-02-10 15:07:54 utc | jmettraux | https://github.com/kennethkalmer/ruote-amqp/blob/master/spec/participant_spec.rb#L112-113 |
| 2011-02-10 15:08:07 utc | jmettraux | https://github.com/kennethkalmer/ruote-amqp/blob/master/spec/participant_spec.rb#L136-138 |
| 2011-02-10 15:10:31 utc | hallelujah | Yes, I have to add another ticket on daemon-kit again :) |
| 2011-02-10 15:10:42 utc | jmettraux | sorry about that |
| 2011-02-10 15:11:42 utc | hallelujah | Yes!!! My scripts work again |
| 2011-02-10 15:12:00 utc | hallelujah | I am amazed by your work guys :) |
| 2011-02-10 15:12:09 utc | jmettraux | well done ! |
| 2011-02-10 15:12:16 utc | jmettraux | it took time ;-) |
| 2011-02-10 15:12:25 utc | jmettraux | (I mean, it took time to build it) |
| 2011-02-10 15:12:26 utc | hallelujah | hey but it worth it |
| 2011-02-10 15:13:35 utc | hallelujah | A friend of mine would like to implement it in his jobs |
| 2011-02-10 15:13:44 utc | jmettraux | oh nice |
| 2011-02-10 15:14:02 utc | jmettraux | sell your knowledge as a consultant ;-) |
| 2011-02-10 15:14:29 utc | hallelujah | "free" consultant :p |
| 2011-02-10 15:14:48 utc | hallelujah | ok thank you for your listening, help ... |
| 2011-02-10 15:15:00 utc | hallelujah | I have to go to a meeting |
| 2011-02-10 15:15:06 utc | jmettraux | I have to go to bed |
| 2011-02-10 15:15:19 utc | jmettraux | you're welcome, à la prochaine + |
| 2011-02-10 15:15:22 utc | hallelujah | :) |
| 2011-02-10 15:15:27 utc | hallelujah | good night |
| 2011-02-10 15:15:37 utc | hallelujah | ?? |