| 2013-06-27 10:19:03 utc | lbt | ruote is really quite powerful |
| 2013-06-27 10:19:27 utc | lbt | haha - I think maybe I'll email this chat to jmettraux |
| 2013-06-27 10:19:37 utc | edulix | lbt: so, can you tell me a bit more about how to do this federation thing? will it require to launch new ruote instances or something like that? have you got references/examples? |
| 2013-06-27 10:20:05 utc | edulix | lbt: feel free to do so hehe open channel =) |
| 2013-06-27 10:20:29 utc | lbt | yeah - john checks ruote_logger which was offline |
| 2013-06-27 10:20:55 utc | lbt | I think you should start by getting a ruote instance up with AMQP |
| 2013-06-27 10:21:22 utc | lbt | I wish we had better examples there - I find the ones we have to be a little opaque |
| 2013-06-27 10:21:31 utc | lbt | sadly I'm not a ruby guy at heart |
| 2013-06-27 10:22:13 utc | lbt | so the idioms make the code fabulously brief .. but a little ... "what now?" |
| 2013-06-27 10:23:13 utc | edulix | lbt: hehe yes |
| 2013-06-27 10:23:27 utc | edulix | lbt: I'm not a ruby guy either, I haven't used it much |
| 2013-06-27 10:23:48 utc | lbt | ok - so 95% of my ruote code is in python |
| 2013-06-27 10:24:07 utc | edulix | lbt: how's that possible? I'm more a python guy too haha |
| 2013-06-27 10:24:13 utc | lbt | I use ruote to handle the messaging/process definition |
| 2013-06-27 10:24:26 utc | lbt | then the actual "do stuff" is python participants |
| 2013-06-27 10:24:40 utc | lbt | they listen on amqp, get a json message, decode, reply |
| 2013-06-27 10:24:45 utc | lbt | wait |
| 2013-06-27 10:25:02 utc | lbt | they listen on amqp, get a json message, decode, disconnect |
| 2013-06-27 10:25:04 utc | lbt | then they work |
| 2013-06-27 10:25:11 utc | edulix | hehe I see |
| 2013-06-27 10:25:14 utc | lbt | later they connect to amqp and send a reply |
| 2013-06-27 10:25:33 utc | lbt | later can be weeks or almost instant |
| 2013-06-27 10:26:08 utc | lbt | so I have python client libraries |
| 2013-06-27 10:26:22 utc | edulix | https://github.com/agoraciudadana/election-orchestra/blob/master/public_api.py btw this is an example of the code I had in mind for the distributed election tally thing |
| 2013-06-27 10:26:26 utc | lbt | we also use supervisor to manage them |
| 2013-06-27 10:26:53 utc | edulix | I use also supervisor in my servers =) what's your project about, is this open source? |
| 2013-06-27 10:27:46 utc | lbt | did you ever hear of MeeGo ? |
| 2013-06-27 10:29:32 utc | lbt | ACTION assumes a quick google is happening :) |
| 2013-06-27 10:30:00 utc | edulix | hehe |
| 2013-06-27 10:30:02 utc | lbt | so meego was a collaboration nokia/intel ... it died. We picked up the parts at www.merproject.org |
| 2013-06-27 10:30:21 utc | lbt | we are now making an entire mobile oriented distro |
| 2013-06-27 10:30:23 utc | edulix | lbt: of course I have heard of meego, I'm a kde developer (well I haven't done much in some time) |
| 2013-06-27 10:30:44 utc | lbt | ah - so Mer is all about Qt on mobile |
| 2013-06-27 10:31:00 utc | lbt | and you may have heard of the Nokia N9 ... and possibly of Jolla |
| 2013-06-27 10:31:43 utc | lbt | so we use ruote to handle our build/QA processes |
| 2013-06-27 10:31:46 utc | edulix | sure |
| 2013-06-27 10:31:59 utc | lbt | yes all open source |
| 2013-06-27 10:32:02 utc | edulix | yeah, I read something of that in the website of ruote, that got my attention |
| 2013-06-27 10:32:19 utc | edulix | nice, where's the code? having gone deep into kde source code, I'm not afraid :-P |
| 2013-06-27 10:32:20 utc | lbt | ACTION is going to akademy too |
| 2013-06-27 10:32:45 utc | edulix | if you're going then we will even have the chance to see each other in there =) |
| 2013-06-27 10:33:13 utc | lbt | https://conf.kde.org/en/Akademy2013/public/schedule/2013-07-13 11:45 ish "Mer & Qt : What MeeGo Should Have Been!" |
| 2013-06-27 10:33:30 utc | lbt | yep - please come and say hi |
| 2013-06-27 10:34:42 utc | lbt | https://github.com/MeeGoIntegration/ |
| 2013-06-27 10:34:57 utc | lbt | https://github.com/MeeGoIntegration/python-ruote-amqp |
| 2013-06-27 10:36:40 utc | lbt | so this stuff is developed for use in our project - so not a huge amount of "how to get started" |
| 2013-06-27 10:37:00 utc | edulix | yeah, no worries =) |
| 2013-06-27 10:37:25 utc | edulix | lbt: this setup is already kind of federated or this is something you want to do in the future? |
| 2013-06-27 10:38:14 utc | lbt | I'll permit myself to say kind of |
| 2013-06-27 10:38:36 utc | lbt | the key thing is that a remote system can launch a process over amqp |
| 2013-06-27 10:39:07 utc | lbt | and currently it is a way for a participant to say "hey, I'm here" |
| 2013-06-27 10:39:21 utc | lbt | but that's the key requirement for federation |
| 2013-06-27 10:39:42 utc | lbt | after all - a remote federated ruote is simply a big participant |
| 2013-06-27 10:40:22 utc | edulix | nice, I'll then take a look at all this stuff, you've been really helpful |
| 2013-06-27 10:40:39 utc | lbt | np - I like to evangalise good technology and ruote is really good |
| 2013-06-27 10:41:08 utc | lbt | you can find me on #mer too |
| 2013-06-27 10:41:42 utc | edulix | nice, I'll be on the mer talk in akademy for sure =) |
| 2013-06-27 10:42:03 utc | lbt | if you go ahead there's a lot of discussion about signing, keys and boundaries in messages which I've had thoughts on but no push to implement yet |
| 2013-06-27 10:42:21 utc | lbt | ok ... l8r then - yell when you get stuck :) |
| 2013-06-27 10:43:26 utc | edulix | nice, those are things that I have to talk too |
| 2013-06-27 10:43:29 utc | edulix | =) |
| 2013-06-27 10:58:31 utc | jmestral | lbt: thanks for the beginning of the chat |
| 2013-06-27 10:59:07 utc | lbt | hey - no fair - you can't go changing nick on me! :D |
| 2013-06-27 10:59:14 utc | jmestral | I was just checking if ruote_logger was active and happened to see half of it |
| 2013-06-27 10:59:28 utc | jmestral | sorry ;-) In Europe right now |
| 2013-06-27 10:59:34 utc | jmestral | not the usual client |
| 2013-06-27 10:59:38 utc | lbt | *g* |
| 2013-06-27 11:04:21 utc | jmestral | lbt: thanks for helping out, much appreciated |
| 2013-06-27 11:06:11 utc | lbt | np - I tried to get started on a simple amqp install/get started recently but hit some ruby/amqp issues and work just got too heavy - hopefully can get back to it in a few weeks |
| 2013-06-27 11:07:34 utc | jmestral | :-) |
| 2013-06-27 11:17:46 utc | jmestral | merphone ftw! |
| 2013-06-27 11:18:13 utc | lbt | oh yeah - so close now :) |
| 2013-06-27 11:24:11 utc | hartog | ow, look, conversation! |
| 2013-06-27 11:24:31 utc | lbt | :P |
| 2013-06-27 11:27:25 utc | hartog | I have little to add on the conversation; other then that I strongly believed in a centralized orchestrator with decentralized services |
| 2013-06-27 11:28:29 utc | jmestral | what do you believe now? |
| 2013-06-27 11:29:01 utc | hartog | ehr ... make that 'believe' |
| 2013-06-27 11:32:25 utc | jmestral | have a nice day! |
| 2013-06-27 11:32:27 utc | jmestral | _quit |
| 2013-06-27 11:32:32 utc | jmestral | \quit |
| 2013-06-27 11:32:40 utc | jmestral | argh |
| 2013-06-27 17:48:55 utc | edulix | hello |
| 2013-06-27 17:49:30 utc | jmestral | hello edulix |
| 2013-06-27 17:50:20 utc | edulix | what does "centralized orchestrator with decentralized services" mean? :-P |
| 2013-06-27 17:50:41 utc | jmestral | I think it's very clear |
| 2013-06-27 17:51:49 utc | jmestral | centralized and decentralized might even be dropped |
| 2013-06-27 17:52:15 utc | jmestral | orchestration tends to be centralized, else it's coordination |
| 2013-06-27 17:52:44 utc | edulix | I like that concept, coordination, then |
| 2013-06-27 17:52:59 utc | jmestral | or should I say choreography |
| 2013-06-27 17:53:10 utc | edulix | that's better yet =) |
| 2013-06-27 17:53:32 utc | edulix | is there a way in ruote to limit the number of simultaneous jobs of a kind in an instance? |
| 2013-06-27 17:53:54 utc | jmestral | yes |
| 2013-06-27 17:54:03 utc | edulix | like, in a queue? |
| 2013-06-27 17:54:04 utc | jmestral | it's limited to the number of workers |
| 2013-06-27 17:55:02 utc | edulix | nice, but the number of workers can be fine tuned for specific "tasks"? like say, "tallying" |
| 2013-06-27 17:55:11 utc | jmestral | (sorry, my answer might be fuzzy and I don't know what you mean by "job") |
| 2013-06-27 17:56:12 utc | jmestral | edulix: I saw your work on frestq, why don't you just go on using it? |
| 2013-06-27 17:56:14 utc | edulix | jmestral: for me a job is something that needs to be done, the "processing of a workitem by a participant", I think it's called |
| 2013-06-27 17:57:06 utc | edulix | jmestral: oh, because I like standing on the shoulder of others, less code to maintain, more eyeballs etc =) I haven't decided yet though |
| 2013-06-27 17:57:16 utc | jmestral | ruote hands workitems to participants |
| 2013-06-27 17:57:37 utc | jmestral | the meego guys have placed AMQP between their "real" participants and ruote |
| 2013-06-27 17:58:21 utc | jmestral | this "indirection level" can be used to determine what jobs should be done |
| 2013-06-27 17:58:52 utc | jmestral | other people have built other systems |
| 2013-06-27 17:59:13 utc | jmestral | pushing workitems to queue/inboxes and having other systems picking up work |
| 2013-06-27 17:59:45 utc | jmestral | the joys of composing an application |
| 2013-06-27 18:00:03 utc | edulix | jmestral: that seems more similar to what I was doing with frestq, because it works with "queues" |
| 2013-06-27 18:00:40 utc | jmestral | tasklists, worklists, inboxes, queues, ... |
| 2013-06-27 18:00:56 utc | jmestral | ruote doesn't care |
| 2013-06-27 18:01:04 utc | jmestral | orthogonal |
| 2013-06-27 18:01:38 utc | edulix | hi chiradeep |
| 2013-06-27 18:03:22 utc | jmestral | ruote could be have a frestq participant and push tasks to it |