ruote tmp/log_2012-07-31.html

2012-07-31 06:50:01 utc hartog jmettraux: good $localtime!
2012-07-31 06:50:24 utc hartog I have some more work for you to review ;) https://github.com/jmettraux/ruote/pull/57#commits-pushed-09128ca
2012-07-31 07:55:54 utc jmettraux hartog: answering your email now, I'll look at the new version immediately after
2012-07-31 07:56:35 utc hartog jmettraux: there's nothing new in the new version - just forgot to remove the workitem and updated the docs a little bit
2012-07-31 07:58:50 utc jmettraux hartog: thanks for the help with Jordi, that's much appreciated
2012-07-31 08:09:48 utc hartog np; glad to be of service
2012-07-31 08:12:17 utc jmettraux I have to leave soon
2012-07-31 08:14:19 utc jmettraux it's starting to look ready, the return; rescue; return is a bit heavy, but well
2012-07-31 08:16:06 utc jmettraux ok, I'll be back later...
2012-07-31 08:19:52 utc hartog l8r!
2012-07-31 08:57:00 utc marco__ good morning
2012-07-31 08:58:30 utc hartog good morning!
2012-07-31 08:59:51 utc marco__ hi hartog :-) nice to find you on this channel, seems you're deeply involved into ruote dev meanwhile?
2012-07-31 09:00:20 utc hartog marco__: im becoming one of the campers ;)
2012-07-31 09:00:53 utc marco__ I'm a bit frustrated currently to be honest with our implementation ..
2012-07-31 09:02:27 utc marco__ I have problems to run ruote multithreaded with postgres -> results in 'to many connection errors' .. currently all participants connecting to postgres run with do_not_thread
2012-07-31 09:04:29 utc hartog john is away now; perhaps he can help with that when he returns
2012-07-31 09:04:30 utc marco__ anyway, I will write to the mailing-list later ..
2012-07-31 09:06:19 utc marco__ my second problem is that the workers are too slow, even with 3 processes, when the engine hits some load, It took far to long to process important workitems in time .. how many ruote-workers do you run usually?
2012-07-31 09:06:59 utc marco__ is there a way to 'reserve' a worker for specific participants ?
2012-07-31 09:09:35 utc hartog ai; I cant help you there!
2012-07-31 09:12:00 utc hartog btw; 'reserving' a worker for specific participants could be done using ruote-amqp
2012-07-31 09:12:31 utc hartog then you could dispatch the workitem to the resource-intensive participant on the other end of your cluster
2012-07-31 09:13:12 utc hartog but perhaps john has an easier solution :)
2012-07-31 09:13:54 utc marco__ ok, thats an option, did you use ruote-amqp . . maybe we can meet for a drink in KA after my holiday ?
2012-07-31 09:14:52 utc hartog i'm using something ruote-amqp'ish (and waiting for the rework branch to get finished ;))
2012-07-31 09:15:23 utc hartog sure! it has been to long since I left durlach and ventured near a starbucks ;)
2012-07-31 09:18:31 utc marco__ haha :-) great! say hello to Martin, Di and Jiri!
2012-07-31 09:19:28 utc marco__ if you want we can meet today or tomorrow afterwork .. or after my holiday (16.8.)
2012-07-31 09:21:21 utc hartog then after your holiday
2012-07-31 10:50:33 utc hartog marco__: did you look at http://www.pgpool.net/?
2012-07-31 10:50:53 utc hartog (although; be warned: https://github.com/rails/rails/issues/1627)
2012-07-31 11:50:48 utc jmettraux hartog: hello, your pull request is final? No more changes?
2012-07-31 11:51:54 utc hartog I only have the acts-like-participant question still open
2012-07-31 11:52:24 utc hartog it seems a matter of taste
2012-07-31 11:53:09 utc jmettraux I'm sorry, I can't remember what question it is
2012-07-31 11:54:13 utc jmettraux could you please enlighten me?
2012-07-31 12:08:46 utc hartog sry; was distracted
2012-07-31 12:09:34 utc hartog https://github.com/jmettraux/ruote/pull/57#issuecomment-7390881 - The SillyObserver code sample (and the text surrounding)
2012-07-31 12:18:10 utc jmettraux aah
2012-07-31 12:18:12 utc jmettraux well
2012-07-31 12:19:06 utc jmettraux one way to do it would be to provide two sugar-coated observers
2012-07-31 12:19:42 utc jmettraux the interesting thing with that SillyObserver is that it can be very lazy, we only compute/grab stuff like workitems when they are demanded
2012-07-31 12:20:13 utc jmettraux somehow, may I delegate that to you?
2012-07-31 12:20:27 utc hartog sure np
2012-07-31 12:20:34 utc jmettraux please explore the possibilities, I'll just pull in when you think it's ripe
2012-07-31 12:20:42 utc jmettraux and we can iterate
2012-07-31 12:20:47 utc jmettraux ripe++
2012-07-31 12:21:15 utc jmettraux when the iteration frequency drops, we'll document it (website wise, hopefully rdoc should follow)
2012-07-31 12:22:16 utc hartog i think it is ripe as-is and that the lazy observer could well be done in a 2nd pull request
2012-07-31 12:26:41 utc jmettraux ok, https://github.com/jmettraux/ruote/pull/57 pulled in, many thanks!
2012-07-31 12:27:17 utc hartog /cc: Thank you sir! For beeing such a good sparring partner (and teaching me some Ruby tricks!)
2012-07-31 12:28:08 utc jmettraux :-) yeah, I love the "let's put in back in perspective" triggers
2012-07-31 12:49:30 utc jmettraux escaping to bed now, thanks for the hard work!
2012-07-31 14:07:25 utc hartog time to go home now