[00:00:40] hmm I just tried that new thanks feature, but I don't even know if it worked - did you get a thanks? [00:00:53] nothing in RC or logs [00:03:10] kaldari, I think it's broken [00:09:38] Did the link switch from a Thanks link to just 'Thanked' text? [00:09:45] yes [00:10:23] looks like it worked, I got 2 thanks notifications from you [00:11:27] I thanked myself using my bot account but no thanks have come through yet [00:11:50] but all this stuff should be logged somewhere IMO (the actions) [00:12:17] it's logged in the eventlogging server only [00:12:50] currently [00:13:00] Susan's point about abuse checking etc - onwiki logs might be a good idea when we deploy elsewhere (i.e. enwiki) [00:13:11] true [00:14:38] Thehelpfulone: You should get an error message if the thanks fails, although it's pretty hard for it to fail since it's so simple :) [00:15:22] If you have Firebug installed (or something equiv) you can look at the API result [00:15:33] and see if it actually worked or not [00:16:08] You should also make sure you don't have that notification category turned off in your prefs [00:16:53] bah, I don't know when I disabled that :/ [00:18:20] hmm kaldari I can thank someone for an edit more than once if I refresh the diff, not sure if that's intended behaviour or if you're keeping it simple without validation right now [00:18:33] it would make it very easy to spam someone with notifications ;) [00:22:00] Legoktm changed group membership for User:Thehelpfulone from autoreview, editor, sysop, transwiki to autoreview, editor, sysop, transwiki, ipblock-exempt: MZMcBride is slow; https://mediawiki.wikipedia.org/wiki/Special:Log/rights [00:22:07] :> [00:22:15] So many notifications. [00:23:12] https://meta.wikimedia.org/wiki/Talk:Requests_for_comment/User_site_behavior_collection#Further_thoughts [00:24:09] Thehelpfulone: Looks like legoktm beat me. [00:24:11] I had to do a call. [00:24:18] And post to Meta-Wiki. [00:24:51] * Susan NP: "All Star" by Smash Mouth from "Astro Lounge" [00:33:11] gn8 folks [00:34:39] Thehelpfulone: That's a bug that doesn't exist in bugzilla yet, but I'm aware of it [00:35:08] feel free to file a bug on it [00:38:51] [[Tech]]; MZMcBride; /* Chick Skin to Vector Skin - Text full page width */ +reply; https://meta.wikimedia.org/w/index.php?diff=5379806&oldid=5378578&rcid=4056585 [00:42:33] [[Tech]]; MZMcBride; /* Myskin */ +reply; https://meta.wikimedia.org/w/index.php?diff=5379813&oldid=5379806&rcid=4056592 [01:02:21] jdlrobson: james forester is on about you maybe having enabled mobile centralnotice? [01:02:31] mwalker: yeh... [01:02:59] all my patches are scheduled to go out thursday; but I can push them now if it's really a problem [01:03:10] there's only one banner running in the system right now [01:03:12] in poland [01:03:27] so if it is enabled it's bugging polish uesrs [01:03:34] mwalker: so the centralNotice div is not present [01:03:45] so in theory it should be benign [01:03:54] but createBanner function is running [01:04:07] i need to rethink how we do alpha/beta/stable RL targetting [01:05:18] annoyingly this didn't get merged - https://gerrit.wikimedia.org/r/#/c/56703/ - so it's also saying device=desktop :( [01:05:43] oh; well; even if I pushed my changes you'd still get banners [01:39:25] anyone here familiar with irc.wikimedia.org ? [01:40:53] Betacommand: somewhat [01:40:58] what's up? [01:41:12] Ryan_Lane: do you know if it limits the number of connections per IP address? [01:41:23] I don't believe so [01:41:36] I'd have to look at the config, which is long and ugly [01:41:41] why? are you having issues? [01:41:58] Ryan_Lane: the toolserver just changed IP addresses and since then there have been multiple users having issues [01:42:10] should be open to the world [01:42:13] hm [01:42:24] maybe our own network has some special properties, though [01:43:02] Ryan_Lane: there might just be a limit of 20 connections per IP which almost noone ever reaches, except the toolserver :P [01:43:27] number_per_ip = 30000; [01:43:34] grr [01:43:35] number_per_ip_global = 30000; [01:43:47] not sure then [01:43:51] hm [01:43:54] that's for "users" [01:44:05] default_max_clients = 10000; [01:44:18] restricted... [01:44:19] number_per_ip = 100; [01:44:19] The Toolserver was having networking issues, I think. [01:44:23] max_number = 100; [01:44:28] My IRC bots were freaking out, but I thought it was all resolved. [01:44:50] Susan: I just rebooted several of mine and its come back [01:45:09] anons are "users" though [01:45:15] so, 30k [01:52:24] looks like it is tide to freenode actually [01:53:50] irc.wikimedia.org isn't part of freenode... [01:54:11] Krenair: I know [01:54:28] Krenair: I was referring to the issues that Im having [01:55:11] You were having trouble connecting to irc.wikimedia.org [01:55:18] Then you said it's tied to freenode [01:56:29] Krenair: Im having issues with an IRC bot [01:59:34] Betacommand: maybe freenode whitelisted the old toolserver IP but hasn't the new one? [02:00:06] fwiw my bots are all running fine... [02:01:36] legoktm: Im not sure [02:03:01] i occassionally get notices like "at a conference and having trouble connecting, ask for help in #freenode" since im on a university network…i guess its possible they limit [02:30:17] I used to get those notices. [02:37:08] Susan: a long time ago, when we used to be friends? [02:43:39] When we were young. [03:47:40] Thehelpfulone: Added logging to Thanks extension: https://gerrit.wikimedia.org/r/#/c/58450/ [06:39:16] [[Tech]]; Optimist on the run; /* Position of categories in Vector */ new section; https://meta.wikimedia.org/w/index.php?diff=5380132&oldid=5379813&rcid=4056951 [10:04:54] any testwiki admins / importers / transwiki importers around? [10:59:02] YuviPanda: What's your username on wmf wikis? [10:59:08] Reedy: Yuvipanda [10:59:15] i can haz rights on testwiki? [10:59:24] (show/hide) 10:59, 10 April 2013 Reedy (Talk | contribs | block) changed group membership for User:Yuvipanda from (none) to administrator and importer [10:59:31] Reedy: <3 [12:31:08] anybody else getting DNS lookup failures for en.wikipedia.org? [12:47:21] not me [13:40:38] [[Tech]]; PiRSquared17; /* Position of categories in Vector */ +; https://meta.wikimedia.org/w/index.php?diff=5381754&oldid=5380132&rcid=4058645 [15:40:05] hey, anyone can me explain this private IP range edits? [https://de.wikipedia.org/wiki/Spezial:Beitr%C3%A4ge/10.64.0.127] [15:46:01] huh, cp1005.eqiad.wmnet. [15:46:09] must be a problem with XFF? [15:46:14] mark: ^ [15:46:22] (re Steef389) [15:50:01] strange [15:53:24] Already in XFF [15:53:24] '10.64.0.127', # cp1005, API [15:53:41] the box looked generally weird [15:53:44] i've restarted both squid instances [15:54:14] i suppose a checkuser could check what the headers where as received by apache? [15:54:35] were* [15:57:38] probably that was a bot editing from labs (or TS)? [16:16:58] addshore, ^ was that your bot? [16:17:11] yes [16:18:13] addshore, where does it run? [16:18:23] bots.wmflabs project [16:18:30] Nemo_bis, yeah it's in labs [16:18:34] Krenair, what did it do? [16:19:36] editid while logged out a wee bit :/ [16:19:37] Some squid forwarded its edits improperly and they showed as the host's private ip: https://de.wikipedia.org/wiki/Spezial:Beitr%C3%A4ge/10.64.0.127 [16:20:04] Krenair: just to let you know it isn't just on dewiki ;p [16:20:38] I know, I saw the conversation in -stewards :p [16:20:42] :) [16:25:59] addshore, I guess its logged in edits were getting that private IP, but this only became visible when it got accidentally logged out [16:27:02] a checkuser (or someone with db access) should be able to find out like jeremyb_ said [16:53:30] Wikimedia Language Engineering team office hours in a few minutes in #wikimedia-office. Wall of text: http://lists.wikimedia.org/pipermail/wikimedia-l/2013-April/125006.html [17:12:29] Reedy: Any chance you could look at this config change for me: https://gerrit.wikimedia.org/r/#/c/57649/ [18:09:55] LocalFile::loadExtraFromDB 10.64.16.24 2008 MySQL client ran out of memory (10.64.16.24) SELECT img_metadata FROM `image` WHERE img_name = 'Comptes_rendus_hebdomadaires_des_séances_de_l’Académie_des_sciences,_tome_106,_1888.djvu' LIMIT 1 [18:10:06] heh, either that was the last straw or there is lots of metadata :) [19:09:25] !log eqiad frack is now using clustered memcache and has site local configuration for DB, ActiveMQ, MemCacheD shtuff [19:09:32] Logged the message, Master [19:16:51] Aaron|home: how is the Redis job queue holding up? [19:17:26] seems to work, but I'd like to see the servers in ganglia [19:17:32] which is being done now [19:17:47] cool [19:19:27] Aaron|home: anomie: I don't know which of you to ask to look at bug 27320, but we'll need to figure that out soon [19:19:31] !b 27320 [19:19:31] https://bugzilla.wikimedia.org/27320 [19:20:04] this is the "Localization Update takes everything down with it" bug [19:20:14] nightly [19:21:02] I don't quite understand why it wasn't as much of a problem before, but it's pretty huge now [19:22:46] this is listed as a ResourceLoader thing, is that accurate? [19:25:22] robla- Seems to be. When the l10n update completes, it clears out the msg_resource and msg_resource_links tables, which seem to be RL. Then I guess the stampede to repopulate those tables brings down the database. [19:26:10] anomie: since this is RL, and since this is an area we anticipate you'll be doing more work on, is that something you can take on tomorrow? [19:26:31] what I think we can do for now is disable it until we have a fix [19:26:44] robla- Sure, I'll try to figure it out. [19:27:00] great, thanks! [19:27:02] I wonder how quick people will start complaining that resources are out of date. Or maybe they won't. [19:28:28] I think we'll just need to communicate well about it. [19:28:46] just spoke to Alolita since she just came back to her desk [19:29:10] is LU triggered from hume, or elsewhere? [19:30:35] It's triggered from a cron job somewhere. I don't know where. [19:31:16] Since the "l10nupdate" script isn't installed on hume, probably not there. [19:33:36] andre__: let me point out https://bugzilla.wikimedia.org/show_bug.cgi?id=47065 to you because it has me somewhat worried [19:36:10] Aaron|home: if the redis queue doesn't fall over this week, we would change all wikis and all jobs to it starting.... Monday? Wednesday? of next week? [19:38:03] chrismcmahon: thanks. Only recent change in WikiEditor itself was https://gerrit.wikimedia.org/r/#/c/58102/ yesterday [19:40:08] did that actually get backported and deployed yesterday? [19:41:06] guess so: april 8: 23:16 logmsgbot: catrope synchronized php-1.22wmf1/extensions/WikiEditor 'Deploying 17a90bc74e87341199a50ebe5348576e05be028a' [19:41:10] yeah [19:41:12] broke IE tests [19:41:17] andre__: what has me confused is that I only see it on test2wiki [19:41:20] andre__: have you see anything on the pumps about this? [19:42:12] robla: I emailed MaxSem about it last night after I saw chrismcmahon's note to me about it. [19:42:28] greg-g: why Max? [19:42:58] that looks like a Trevor/Roan thing [19:43:00] because he thought that my startup module changes broke it [19:43:21] right, chrismcmahon thought it aligned with the deploy [19:43:44] anywho, doesn't matter who, just need to figure this one out [19:43:44] ah, got it [19:43:46] greg-g: I still want to switch some real queues today [19:44:13] well, it did :-) but I don't see the problem on enwiki or mw.o, just test2. thus I worry. [19:44:22] chrismcmahon: right :/ [19:46:04] may be worth temporarily disabling any extensions on test2wiki that aren't also on enwiki to see if the problem goes away [19:46:52] Aaron|home: good deal. I'm going to send an email out to wikitech-l/engineering saying the switch over is in-progress. Should I give an eta (with an accuracy of a day or so) of when the queue will be fully switched over? [19:46:53] chrismcmahon: sounds like a problem we should mainly just keep an eye on rather than make an urgent priority. you can repro the problem manually, I assume [19:47:05] ideally monday [19:47:11] robla: yes, it's killing GuidedTour also [19:47:13] Aaron|home: perfect. [19:47:32] chrismcmahon: also only on test2wiki? [19:47:41] robla afaik yes [19:51:32] anomie: still looking for l10nupdate? [19:51:50] jeremyb_- I think I found it. It looks like it's run from the "l10nupdate" user's crontab on fenari [19:53:41] anomie: that matches site.pp [19:55:39] huh, the mediawiki module (which i guess is not in use?) would install the l10nupdate user everywhere it's used? [19:58:40] looks like the snapshot hosts probably have an l10nupdate [19:58:42] user [19:59:00] * jeremyb_ gives up trying to understand [20:01:34] nom time [20:15:12] kaldari, can you review the login change to core we were talking about? https://gerrit.wikimedia.org/r/#/c/55847/ [20:19:20] apergos: it works! \o/ https://ganglia.wikimedia.org/latest/graph_all_periods.php?c=Miscellaneous%20pmtpa&h=hume.wikimedia.org&v=823574&m=Global_JobQueue_length&r=hour&z=default&jr=&js=&st=1365625056&z=large [20:20:10] superm401: will try to look at it soon [20:20:18] kaldari, thank you. [20:27:05] it does? [20:27:58] huh I guess, when I looked earlier the fix didn't take but maybe it just took forever [20:30:29] apergos: yes, that's why I waited some hours before annoying you with pressing questions "when will we arrive?" "is something broken?" "did we take the right path?" :p [20:30:43] well I know the path was right cause [20:30:50] I checked it on hume [20:30:54] path metaphorically [20:30:59] but stafford has been a dog all day [20:31:08] enwiki's doesn't work though [20:31:09] I was trying an imitation of a child during car travel to seaside [20:31:11] partly due to a manifest issue but [20:31:25] jeremyb_: good news is that we can now ignore it :D [20:31:25] no enwiki's does not, and I noted that when disabling the jobs [20:31:33] orly [20:31:33] they just spewed instead. [20:31:47] yeah, 'no such table enwiki.job' or something [20:31:49] the query is wrong because the job table changed [20:32:04] since I didn't set em up and I dunn what the latest greatest setup is... [20:32:13] someone with a clue can fix that up [20:32:14] ...or rather nobody knows... [20:32:23] aka https://bugzilla.wikimedia.org/show_bug.cgi?id=41100#c5 [20:32:29] great [20:40:32] greg-g: no, no mention at all of the IE vs WikiEditor issue in about a dozen of VPs that I reads today. (sorry for late answer, had dinner so I just came back) [20:40:46] andre__: no problem! and thanks [21:01:02] wtf is going on with gerrit? it is crazy slow [21:01:24] and now is throwing errors at me occasionally... [21:02:07] ^demon: you around? [21:13:44] awjr: he was sick before (not too long ago...) [21:13:56] yeah i heard he's still out, thanks jeremyb_ [22:09:55] https://ganglia.wikimedia.org/latest/graph.php?me=Wikimedia&m=cpu_report&r=hour&s=by%20name&hc=4&mc=2&g=cpu_report&z=medium wow :P [23:04:04] Krinkle: it stands for Victory Form [23:04:09] (I kid.) [23:04:57] Vertical Form, I do believe. We decided all this "Agora" naming in core was not appropriate, and it should rather be something more functionally descriptive. [23:47:09] Hello, anyone with the New Contributor Survey on Wikipedia know how I can force it to appear? I'd like to test for a bug without registering an account... [23:53:39] Matthew_, I'm not quite sure what you're referring to... [23:53:56] apparently there's some sort of survey prompt given to new wiki signups [23:54:06] i'm tracking down someone to talk to about it [23:54:11] Krenair: The new contributor survey that new Wikipedia editors get. [23:54:14] jeremyb_: Thanks :) [23:56:00] asked in the ee channels [23:57:59] Krenair: Thanks :) [23:58:14] I just want to test a real quick thing with the "previous" and "next" buttons