[01:10:02] requested thoughts from any available analyst; so, I'm gonna do a writeup of our hive/[language researchers use] integration options [01:10:06] officewiki, wikitech or mediawiki? [01:10:11] wikitech feels like the best location to me, but. [01:18:47] Could use Meta-Wiki. [01:18:51] But wikitech is probably best. [01:19:12] Unless it's private, in which case the option is clearly officewiki. Not sure what would be private, though. [09:27:19] (CR) Nuria: [WIP] Run recurring reports using the scheduler (1 comment) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [10:05:56] (CR) Nuria: Add scheduler mode to wikimetrics (1 comment) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111623 (owner: Milimetric) [13:57:09] (PS20) Milimetric: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 [14:00:11] ls [19:26:00] hi drdee! :) [19:26:20] yo! [20:18:09] DarTar: whenever you have 5 min, let's chat about the internships [20:19:05] lzia: later this afternoon? I want to continue writing up reports on Meta in preparation for Wednesday [20:19:14] sure. no rush for it. [20:19:19] just ping me when you have time [20:40:54] halfak, milimetric: it just occurred to me that we haven't considered submitting a Wikimania talk proposal on metrics standardization / EEVS [20:41:00] how do you guys feel about this? [20:42:05] i feel like we should submit that! [20:42:27] Sounds like a good idea to me. [20:42:48] it doesn't commit us to anything in specific (other than an update on how far we got, research and development-wise) [20:43:20] cool :) I'll start a stub then [20:50:51] I should probably do a presentation one of these years, but: effort. [21:45:36] does anybody have a clue about the purpose of log_action="newusers" (not log_type) in dewiki? Felipe brought this up here (and I just saw his note a moment ago :( ) https://meta.wikimedia.org/wiki/Research_talk:Newly_registered_user [21:50:22] there's a total of 85,711 entries in this log (I haven't done any further analysis on this) [21:53:36] I saw that come up for a short period in (I believe) 2006. Let me check. [21:53:39] DarTar: ^ [21:54:06] Here's my note: New users appear in logging table as log_type="newusers" AND log_action="newusers" between 20050907221649 and 20060421232859. [21:54:10] See https://meta.wikimedia.org/wiki/Research:Wiki_archaeology [21:54:49] oh [21:55:24] I should really subscribe to diffs of that page [21:55:28] checking it out [21:55:57] Looks like the same business in dewiki. [21:56:01] Pastebin coming [21:56:13] http://pastebin.com/rhVANppb [21:56:28] ^ Between 200509 and 200604, the record appear. [21:56:32] halfak: let me understand, are we saying that during that window legit new user registrations use that log instead of log_action='create' ? [21:57:09] I never looked into it carefully, but I assumed that was the case. [21:58:44] Checking the user table now. [21:58:51] on dewiki log_action = 'create' starts on 20060417043326 [21:58:54] so yes, it definitely makes sense [21:58:59] Goddamn lack of user_registration date [21:59:54] I've got a new pastebin that suggests that log_action = "newusers" represents the entire user table during that timespan. [22:00:03] http://pastebin.com/1u4y68q6 [22:00:28] alright, thanks for the quick check [22:00:44] So, you'll notice that the records are limited (due to user_registration missing), but they nearly match up to the counts from logging. What little difference there is could be attributed to the seconds around a month switch. [22:01:12] yeah [22:01:42] * halfak can't wait to just build a canonical user_registration table for all wikis once we get them loaded onto our UBER DB SERVER. [22:01:46] there are also differences of a few seconds between centralauth and local DBs [22:01:49] FUN [22:01:56] amen to that [22:01:57] Yup. [22:02:24] fuzzy transactions [22:02:31] The timestamp of the registration record in the logging table is sometimes a couple seconds off of the user_registation date in the same DB. [22:02:38] bbl, off to talk to his tobyness [22:02:44] hasta luego [23:29:14] Hey folks. That query that is blocking replication on s1-analytics is now a zombie? Anyone got an axe? It's still blocking replication in it's undead state. [23:29:27] ottomata, I'm hoping you've got one. [23:29:38] I tried springle, but his side of the planet isn't warm yet. [23:34:02] helloo? [23:34:09] i have been summonnnneeed [23:34:45] ergghhh now i have to figure out how to get into that node [23:35:00] any idea what its real hostname is? [23:35:26] ottomata: s1-analytics-slave.eqiad.wmnet ? [23:35:43] ah! -slave [23:35:43] thanks [23:35:47] np [23:36:53] FYI: springle turned around. [23:36:53] halfak: it is busy rolling back: ---TRANSACTION 8A2A845D6, ACTIVE 616248 sec rollback [23:36:54] not much we can do except wait [23:36:54] OK> Thanks for your help [23:38:39] yeah, i was just about to quote some article that I found that said the same thing [23:39:50] k grabbing some dinnah, laters!