[00:05:51] (CR) Milimetric: [C: 2 V: 2] Remove updates on graph edits-monthly-5plus-editors [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/109790 (owner: Jdlrobson) [00:33:48] tnegrin: yt? [01:08:31] Ironholds: what's the bug #? [01:09:55] tnegrin, not thrown it in yet, text above refers to the one you resolved :) [01:10:00] I've replicated the second one, though [01:10:05] about to put it in BZ [01:10:17] great - lmk [14:53:49] Snaps: hiiii [14:53:53] i am looking at kafkatee, wooo [14:54:01] do you mind if I change tabs to spaces in your conf.example file? [14:54:10] to avoid this? [14:54:10] http://cl.ly/image/1K143D1w0N2x [15:24:09] ottomata: do it! [15:24:22] mmk, also q [15:24:32] did you mean to set the commented out configs to not their defaults? [15:24:39] # Default: largest [15:24:39] #kafka.topic.auto.offset.reset = smallest [15:27:07] yes, it is an example file, if it would all be defaults it wouldnt be very fun :) [15:31:40] haha, ok cool [15:31:44] juuust checkin' [18:55:54] (PS2) Ottomata: Initial kafkatee [analytics/kafkatee] - https://gerrit.wikimedia.org/r/109505 (owner: Edenhill) [19:02:24] something is going on with my google hangout , give me a sec and i will be on batcave [19:42:54] milimetric, do we have any browser stats currently? I assume http://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm as bad as it was will no longer be updated since we don't use squid anymore [19:48:18] jgonera: squid is just a legacy name for 'all webrequest logs' now [19:48:36] ottomata, so this page will still work? [19:48:39] i believe so [19:50:13] yes, jgonera, ottomata is right as far as I know [21:10:06] (PS1) Milimetric: Handle history file and output directory properly [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/109959 [21:11:20] (CR) Jdlrobson: [C: 1] "Will test and +2 when i get a chance but code looks fine" [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/109959 (owner: Milimetric) [21:23:15] (CR) Jdlrobson: [C: 2 V: 2] Handle history file and output directory properly [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/109959 (owner: Milimetric) [21:47:13] Ironholds: how's hadoop treating you? [21:47:17] i see maybe a failed hive query there [21:56:55] ottomata, creepy ;p. what query? [21:57:51] i can't see the whole thing [21:58:04] SELECT dt, ip, ... [22:03:31] aha [22:03:42] wait... [22:03:46] no, that query worked fine [22:04:07] done via a system() call to R that did idle after it completed, though. [22:04:33] which could produce odd-lookin' outcomes at your end, I guess? [22:05:43] I'll experiment with having it just kill itself [22:14:13] hmm, it worked? i see it failed in the job interface [22:14:30] oh [22:14:30] actually [22:14:30] no i don't [22:14:36] just some of the maps failed [22:14:36] hm [22:14:52] hm maybe it submitted new attemps for them [22:23:36] DarTar & Ironholds: Any experience extracting information about flaggedrevs from MySQL? [22:23:55] oh boy, a long time ago [22:24:16] I'd need to do some archaeology of my own scripts [22:25:22] (CR) Ottomata: Initial kafkatee (8 comments) [analytics/kafkatee] - https://gerrit.wikimedia.org/r/109505 (owner: Edenhill) [22:25:30] Gotcha. I'm just trying to figure out if there is a good way to sample the current number of revisions needing review. [22:25:44] Or better yet, a time series :) [22:26:45] brb [22:32:47] halfak: it's straightforward to get flagged rev status via the API but probably not for the analysis you want to run [22:34:20] halfak: 6 flaggedrevs* tables per DB, I'm sure there's plenty of documentation to look up ;) [22:34:51] ugh, blobs [22:38:30] Yeah. There's some flaggedrevs_stats and flaggedrevs_statistics, but it looks like the don't have new data and some of the stats are just constant values. [22:39:01] Anyway, thanks for looking at it with me. If I get anywhere, I'll share some docs.