[08:27:39] New patchset: Rfaulk; "add. pop method to broker." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73392 [08:27:39] New patchset: Rfaulk; "mod. Use umapi_broker_context to pull new requests from queue." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73393 [08:27:58] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73392 [08:28:10] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73393 [08:32:36] New patchset: Rfaulk; "rm. request_queue dep." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73394 [08:32:59] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73394 [13:04:40] yo qchris! [13:05:15] yoooo ottomat [13:05:21] yooooooooooooo milimetric [13:05:28] yooooooooooooooooooooooooooo average [13:05:37] yoooooooooooooooooooooooooooooooooooooooooooooo erosen [13:33:00] New patchset: Milimetric; "added explicit mysql-python dependency" [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73402 [13:33:08] Change merged: Milimetric; [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73402 [13:46:00] there is no start of the day like pushing some code ;) [13:53:08] pushing code would be a great start to the day [13:53:13] self-merging code makes me feel like a circus animal jumping through hoops [13:56:48] milimetric: gerrit knows best :) [13:56:59] hahaha! [13:57:07] no matter where I am in the IRC universe [13:57:15] if I bad-mouth gerrit, Yuvi is here :) [13:57:42] it's just that i'm a bad developer [13:57:48] so when I try to fix stuff, I make like 200 changes [13:57:55] i'm a crazy fast developer [13:58:07] and it evens out in the end, i get stuff done [13:58:13] but gerrit affects people like me more [13:58:42] the above may not be entirely accurate :) [14:00:37] you can always ask me to blindly hit the +2 button :P [14:04:18] milimetric: :D I've started using dependent commits so I can push multiple commits to gerrit at a time, and do the pointless button pushing later at once [14:04:36] huh [14:04:37] milimetric: when I use gerrit directly at all, that is. I still mostly push to github and then have the bot pull to gerrit [14:04:43] what are dependent commits? [14:05:58] milimetric: step 1. have multiple commits in the same branch, step 2. do a git review, step 3. git review asks you if you are sure 4. say yes 5. the different commits you have on the branch are pushed as indepdendent patches on gerrit, but are marked as 'dependents' 6. you can later go review them [14:07:57] hm, not sure how well that works with multiple people pushing to the same repo [14:08:00] but I'll give it a try [14:08:10] thanks YuviPanda :) [14:08:51] milimetric: no that works pretty neatly, I think. Point is instead of 'make commit, go to browser, +2, come back, pull, make another commit', it becomes 'make commit, make commit, make commit, make commit, go to browser, +2 all the things' [14:08:52] so [14:09:11] yeah, i agree, that saves a lot of time [14:09:28] i'm just skeptical that gerrit will be ok with this :) [14:09:59] milimetric: it's a feature! :) [14:10:07] hehe [14:10:09] they advertise it! and stuff. [14:10:17] and it works fairly well [14:10:18] so try it out :) [14:10:20] cool, ok [14:10:28] mobile web uses it all the time [14:13:33] hey drdee the labsdb setup was pretty simple [14:13:49] the databases are now configured and I think except for a bunch of configuration cleanup we're done with that card [14:20:33] awesome! [16:01:05] Snaps, hangout? [16:01:20] yep [16:01:32] cool [16:03:12] Snaps: you have the link? [16:03:35] nope :) [16:03:41] https://plus.google.com/hangouts/_/fa3606e78bbeb125c0a4f1c9d6a8c74cb8c4ad54 [16:14:18] New patchset: Milimetric; "cleaned up navigation and added /metrics page" [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73428 [16:14:36] Change merged: Milimetric; [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73428 [16:15:26] drdee, tell me what you think: http://wikimetrics.instance-proxy.wmflabs.org/metrics/ [16:15:34] it's just a rough draft but that's all automatically generated [16:17:46] k [16:37:47] very cool milimetric! [16:37:51] really like it [16:38:00] k [16:38:06] onto cohorts then [18:00:55] New patchset: Stefan.petrea; "Initial Debian packaging" [analytics/dclass] (debian) - https://gerrit.wikimedia.org/r/68711 [18:08:40] New patchset: Milimetric; "implemented cohorts page" [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73442 [18:08:58] Change merged: Milimetric; [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73442 [18:09:38] niCE! milimetric [18:09:52] i just published to labs [18:09:58] you should be able to navigate around the site now [18:11:14] hit http://wikimetrics.instance-proxy.wmflabs.org/demo/create/cohorts/ to get it to create that sample cohort set for you [18:11:24] i already saw it ;) [18:12:49] uh... server seems brokeded now [18:13:01] i'm looking into it [18:13:31] create new job fails: Error! [object Object] [18:16:00] right, celery's not running [18:16:08] but i'm confused... [18:16:09] oh! [18:16:16] that might be what's killing my process :) [18:16:19] it's waiting on celery for you [18:16:26] hm, this is not good [18:16:34] hm.... [18:16:44] we should definitely test this thoroughly [18:23:13] hm, it'd be nice if somebody knew how to configure upstart [18:35:24] cool, so I got upstart working - it runs the wikimetrics celery queue on startup [18:35:31] seems to be working fine but the jobs aren't getting queued [18:35:44] I'm getting a little winded :) Might wind down for the day [18:36:08] work on some CSS or something :) [18:37:42] Soft [18:37:50] Argh stupid autocorrect [18:37:57] Aight [18:39:00] oh you callin me soft? [18:39:12] aite [19:15:57] New patchset: Stefan.petrea; "Initial Debian packaging" [analytics/dclass] (debian) - https://gerrit.wikimedia.org/r/68711 [19:18:09] New patchset: Stefan.petrea; "Initial Debian packaging" [analytics/dclass] (debian) - https://gerrit.wikimedia.org/r/68711 [19:25:26] New patchset: Stefan.petrea; "Initial Debian packaging" [analytics/dclass] (debian) - https://gerrit.wikimedia.org/r/68711 [19:31:30] milimetric: 504 Gateway Time-out on the labs instance when accessing http://wikimetrics.instance-proxy.wmflabs.org/jobs/create/ [19:48:46] thanks drdee I know, I'm not sure why that happens once in a while [19:48:52] Will have to ask Andrew [19:48:54] ok [19:56:23] could this be helpful: http://stackoverflow.com/questions/17174559/how-to-track-down-a-python-django-uwsgi-nginx-timeout ? [19:57:49] maybe :) [19:58:07] thanks for looking, I'll track this down eventually, I just hate wasting time on server poop [19:58:17] ok [20:29:05] drdee: can I ask something ? [20:29:14] always! [20:29:22] drdee: is kraken eventually going to be split into multiple deb packages ? [20:29:58] i don't think we will build deb's for the kraken jars [20:30:03] and for cdh4 we already have the debs [20:30:18] ok [20:40:09] New patchset: Milimetric; "fixed redirect after creating a job" [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73516 [20:40:09] New patchset: Milimetric; "got job refreshing smoothed a bit" [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73517 [20:40:09] New patchset: Milimetric; "fixed the current_user usage in Job.__init__" [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73518 [20:40:51] Change merged: Milimetric; [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73516 [20:41:01] Change merged: Milimetric; [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73517 [20:41:17] Change merged: Milimetric; [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73518 [20:41:27] omg that was awful :) [20:41:37] sorry for the spam [20:41:49] i think i'm gonna stick with not that YuviPanda [20:42:07] milimetric: not dependent commits you mean? [20:42:08] heh [20:42:08] ok [20:44:47] it doesn't really save me any work, I still have to approve all of them separately [20:44:48] and it's just that much more annoying for my teammates. [20:44:54] yea [20:45:14] but thanks though, it's interesting that it works [20:48:42] heh, drdee, I think maybe sorta creating a job works? [20:48:42] it still randomly does this bad gateway thing [20:48:54] and one of my jobs seems to be stuck in "pending" [20:49:06] but good enough for the end of a long Friday :) [20:49:18] have a good weekend everyone! laters :) [20:50:39] New review: Faidon; "(figured out a plan on IRC)" [analytics/dclass] (debian) C: -1; - https://gerrit.wikimedia.org/r/68711 [21:09:59] average: can you email me what you and faidon decided regarding the way forward of the dclass debianization [21:29:09] drdee: summary => There's a new package planned called libdclass-java (along with the other ones, libdclass0 , libdclass0-dev , libdclass0-jni, libdclass-data) [21:30:02] this package will contain a .jar with the .class files for the JNI wrapper [21:31:39] due to gerrit complications we have to pull the bits and pieces, drop the gerrit patchset, redo it in different chagsets, one for the upstream changes we need (.so will be loaded from disk and not from .jar ; Makefile.am will compile the upstream JNI and also compile the .java into a .jar which will be used later in the debian branch ) [21:32:03] new changeset about the debian branch which will basically do what we did so far, except in addition we have the aforementioned libdclass-java [21:32:30] and, there is a member of Faidon's team , Brandon, who will be reviewing the Makefile.am [21:33:26] (yay gerrit) [21:33:53] another topic discussed was the new multi-arch feature that's required for Debian packages [21:33:56] that's my understanding of the current situation