[14:25:57] moooorning [14:26:04] morning! [14:26:09] ottomata, i am not sure if openjdk 7 is a good idea [14:26:24] i don't think it is compatible [14:26:32] i think its ok, i've seen places saying newer hadoops are fine with it [14:26:55] cdh 4.3 (which we are not running) has an explicit check for openjdk 7 and will abort [14:27:02] OpenJDK is handy to have on a development system as it has more source for you to step into when debugging something. OpenJDK and Sun JDK mainly differ in (native?) rendering/AWT/Swing code, which is not relevant for any http://wiki.apache.org/hadoop/MapReduce Jobs that aren't creating images as part of their work. [14:27:03] whaaa [14:27:04] ?? [14:27:10] is it just 7 or openjdk? [14:27:15] http://wiki.apache.org/hadoop/HadoopJavaVersions [14:27:22] i belief openjdk 7 [14:27:29] so openjdk 6 maybe ok? [14:27:36] maybe, not sure [14:27:43] because that hadoop page says openjdk is fine [14:27:46] doesn't say which version though [14:28:08] hm Note*: OpenJDK6 has some open bugs w.r.t handling of generics (https://bugs.launchpad.net/ubuntu/+source/openjdk-6/+bug/611284, https://bugs.launchpad.net/ubuntu/+source/openjdk-6/+bug/716959), so OpenJDK cannot be used to compile hadoop mapreduce code in branch-0.23 and beyond, please use other JDKs. [14:28:11] but does that mean 6? [14:28:12] or 7? [14:28:14] what crap [14:28:14] hm [14:28:57] wouldn't it be better to make a separate mingle card for this? and do the migration in one swoop? having different jdk's on the same cluster also sounds troublesome [14:33:25] it would be a prereq for the reinstall [14:33:33] i dunnooooo [14:33:40] ops just wanted me to do this [14:33:43] if we absolutely can't [14:33:48] then we can argue to staty on oracle [14:33:50] but sigh [14:40:58] but we had the exact same discussion back in august and we said we would do the upgrade once CDH would support it [14:41:19] why is this a prereq for reinstalll? [14:43:43] Probably unrelated: but people on the kafka mailing lists are having problems with OpenJDK and Kafka under load. Works with Oracle. [14:44:21] totally related! thanks Snaps [14:46:08] Snaps [14:46:08] Probably unrelated: but people on the kafka mailing lists are having problems with OpenJDK and Kafka under load. Works with Oracle. [14:46:27] i had thought I'd seen that openjdk 7 is supported, but its not really [14:46:35] people try it, but then report problems, and it is not recommended [14:46:42] oop [15:28:34] that sux re: openjdk [15:37:25] raiiin [15:37:42] be back in a bit, gotta drop car off to get it looked at, its being real funky [16:19:56] milimetric: can you have a look at https://github.com/wikimedia/kraken/commit/a06ea8031fd6b32d867670309e8d0c96f54ecb8d and let me know what you think? it's a bit urgent , it's the W0 stuff [16:20:06] ok, looking now [16:20:30] the config options are from https://wikimediafoundation.org/wiki/Mobile_partnerships [16:22:45] instead of default / legacy, drdee, I'd use something explicit [16:28:14] drdee: I am done looking it over [16:28:21] I made a comment on the one line that's dangerous [16:28:47] ty [16:28:53] I think the rest is ok but obviously running the tests will tell the whole story [16:29:14] so the other side note is that we have to seriously talk about the way forward for this kind of logic [16:29:23] about indexOf [16:29:25] we can't just keep piling up code in Java [16:29:44] (not without a clear architecture and way forward that is) [16:29:52] yeah, indexOf? [16:29:52] i checked all the filenames and this will work [16:30:26] yes, we need to think about the architecture, once you are done with UMAPI :)D [16:30:27] :D [16:31:42] back [16:33:22] oh I know it works for the current filenames, but it's very fragile [16:33:27] if someone adds a new carrier... [16:37:58] right but ottomata is consistent in naming the files [16:38:13] and this is for the soon to be deprecated method [16:42:41] whatchyall talking about? [16:44:13] ottomata:where is the umapi htaccess credentials stuff? [16:44:59] the pw? [16:45:06] yyup [16:45:07] what's the lin [16:45:08] k [16:45:11] in puppet private [16:45:17] wasn't it on office wiki [16:45:18] ? [16:45:35] oh yeah! [16:45:35] https://office.wikimedia.org/wiki/User:Aotto/umapi [16:45:40] thx [17:01:19] ottomata; scrum [17:02:31] sorry, here! [17:19:00] average, are you around? [18:22:36] apparently there are other people who think about celery and nose: http://www.youtube.com/watch?v=XOmcew2DtTE [18:32:28] hahahaha [18:34:46] * drdee hears an evil grin [18:37:21] ottomata and drdee: what do you think about creating a separate zero dashboard and data and code repo? [18:37:38] right now it runs on the gp.wmflabs.org dashboard [18:38:23] and so the code for generating it lives in the same place as the code for generating the rest of the stuff on that dashboard [18:39:22] or do we want to create a part of the kraken repo dedicated to limnification, so to speak [18:45:23] either way [18:45:27] fine with me [18:51:02] hey erosen, I'm hanging out [18:51:54] milimetric: word, talking to jonathan morgan in office for a moment [18:52:08] cool, no prob [19:56:39] hello [19:56:43] ottomata: hi [19:56:44] drdee: hi [19:56:45] milimetric: hi [19:56:46] erosen: hi [19:56:54] hiiii! [19:57:02] ottomata: I'll get my headset and do a call with you if you want [19:57:02] we can hang out [19:57:02] evening average [19:57:40] coool! [19:57:41] ok 1 sec [20:02:31] average: [20:02:31] https://plus.google.com/hangouts/_/fffbf9af564db0f8e4222731f74a7050870538dd?authuser=1 [20:03:42] ottomata: not allowed [20:03:56] hm, try now average [20:04:00] tryin