[00:33:07] gn8 folks [01:09:22] 06 22:19:07 < jeremyb> hrmmm, 2012110610011273 looks to me like either a charset issue or more likely something with gzip ? [01:09:25] 06 22:19:38 < jeremyb> When I try to look at the wikipedia articles on Barack Obama and Mitt [01:09:29] 06 22:19:40 < jeremyb> Romney, both pages are unintelligible symbols. Any reason for this? I [01:09:32] 06 22:19:51 < jeremyb> (quoting the ticket. ticket has a screenshot) [01:09:39] idk if anyone actually investigated that [01:11:10] we got a second user reporting the same thing almost a full day later. (2012110710011682. this time about [[Hillary Rodham Clinton]] and they didn't give a screenshot) [01:11:23] maybe there was a 3rd, don't remember for sure [01:11:40] I checked [[WP:VPT]] and didn't see anything relevant [01:12:15] anyway, keep an eye out for reports of something like that and maybe someone knows of a relevant recent change? [01:12:43] bbl [01:16:51] gah, where did StevenW go so fast?! [01:19:27] hell [02:26:45] any devs in here? [02:27:12] some, maybe [02:27:19] !ask | Magog_the_Ogre [02:27:19] Magog_the_Ogre: Hi, how can we help you? Just ask your question. [02:27:39] if you head to https://commons.wikimedia.org/wiki/File:2012_General_Election_Results_by_County.png [02:28:01] you'll notice that my most recent additions aren't showing up on the pre-made thumbnail resolutions, even after purging cache [02:33:06] actually i see them when clicking the "other resolutions" links [02:33:32] that's the whole point [02:33:36] you have to click the other resolutions [02:33:42] or the image itself to let the browser render it [02:33:49] but the thumbs are NOT being renewed, even after purging [02:33:56] they're still stuck on the old versions, mutante [02:34:09] or, the thumbs that are displayed on that page [02:34:12] (e.g., 800px) [02:41:25] Magog_the_Ogre: i see what you mean, well, almost, the image on the File: page is not showing the latest revision, if i actually use a thumbnail in a page though or click any size thumbnail link it is current, so you should still be able to use it normal. While i dont know how to solve this currently, i can forward it [02:41:43] please do [02:41:48] ok [02:41:49] I filed this a while ago, and I think it was "fixed" [02:41:52] but apparently not ;) [02:42:01] notice if you click the 800px, it still doesn't work [02:42:05] how did you file it? [02:42:16] do you have some ticket or bug number handy? [02:49:26] I wish I could remember [02:49:44] is there a way to look up your own bugs in bugzilla, mutante [02:49:44] ? [02:57:12] Magog_the_Ogre: yes https://bugzilla.wikimedia.org/query.cgi?format=advanced ..by People [02:58:12] it says I've only filed two bugs, ever [02:58:21] which is obviously wrong [02:58:30] unless there is another bug reporting system [02:58:35] I have no idea why it's missing the others [02:58:55] or maybe I've filed less than I thought [02:58:56] hmm [02:59:20] Magog_the_Ogre: you may have to mark all products and components above. hmmm [03:00:12] AbsenteeLandlord: what a weird component name [03:00:26] or try: Summary: contains ... [03:01:51] well, I know I've had this conversation in here before [03:01:51] or your email ? [03:01:51] I guess I never filed a bug on it [03:02:00] ok [03:02:49] ha! [03:02:52] here's one: https://bugzilla.wikimedia.org/show_bug.cgi?id=36048 [03:03:23] apparently the search function on bugzilla is developmentally disabled [03:06:51] Magog_the_Ogre: i get 4 bugs reported by you [03:07:14] by gmail found others than those 4 ;) [03:07:19] I'll have to file a bug on bugzilla lulz [03:08:41] did you select all Status' and Resolutions? [03:09:08] feel free to do so (bug on bugzilla;) though [03:09:12] I tried at least [03:09:21] nah I'll decline [03:09:27] too much else to do that is important ;) [03:09:32] and on the thumbnails, wanna just reopen the one you found then? [03:09:42] well the one I found specifies an SVG [03:09:46] this is a PNG [03:09:49] so it's not an SVG specific one [03:11:23] Bug 31680 sounds right at quick glance [03:11:31] and it is reopenend already [03:13:16] or 41130 [03:21:09] Magog_the_Ogre: looks fixed to me now , heh [03:21:21] yeah me too :( [03:21:27] which is why I only bothered with this bug once [03:21:31] it always fixes in like an hour [03:21:37] and then the devs are like "meh, works for me" [03:27:00] heh, i see. well, just report it anyways to 41130, and if it is just pasting what you just said [04:13:27] Magog_the_Ogre: mutante: the simplest way to do that search is stick "ALL rep:emailaddresshere" in the box at the top right corner of all pages. that "emailaddresshere" can be a substring or the whole thing [04:14:35] (and you can then hit the edit this search link to see what you would have needed to do in the advanced form to get the same results) [04:16:26] aha, thanks! [04:23:11] mutante: any recent changes to to text squids as far as you know? [04:23:25] or maybe apache conf? [04:23:39] see me in here ~3 hrs back [04:27:01] my best wild guess is that it's a broken browser. but I know we (and lots of other sites) have some hacks to not ever send gzip for certain combos of headers for certain UA strings. (even if they say they can accept it). i'm just wondering if maybe there's either 1) a new broken browser that needs to be added to the blacklist or 2) a serverside change that made us lose that blacklist somehow or somewhere. i.e. maybe it's even just missing for on [04:30:14] anyway, we can at least see if more reports come in [04:30:51] jeremyb: sorry, not aware of changes related to browser blacklists, recent apache conf changes, yes, but just stuff like redirects and new virtual hosts, adding a ServerAlias.. [04:31:11] yeah, that's not relevant [04:31:11] ;-/ [04:31:28] * jeremyb wants more hours in the day [08:10:08] frrwiki is down because of a server config error [08:11:19] Reedy: could it caused by your extension update from Nov 2th? [08:12:21] http://frr.wikipedia.org/wiki/Wikipedia:Hoodsid [08:43:29] Reedy: edits on frrwiki stopped just after you wmf3 switch at 19:51 yesterday [08:45:05] Merlissimo: He'll see your note (as will other ops) when they arrive. What is frr? [08:45:35] spoen in the north of germany [08:45:50] Ah. [08:46:00] nordfriesisch [08:46:23] perhaps you know the isle Sylt [08:47:04] at the frontier to denmark [08:48:21] I just looked it up, actually. [08:48:22] :-) [09:10:35] Please, can someone review and deploy changes for this bug tag as "critical": https://bugzilla.wikimedia.org/show_bug.cgi?id=41872 [09:14:55] Tim-away: ^ if your back tonught and no one else has [09:29:53] Tpt, try nagging someone of ops [09:30:13] I +1 it, but I can't merge there [09:48:25] Platonides:Thanks. [12:50:20] hashar: any idea why https://gerrit.wikimedia.org/r/#/c/32361/ is failing? [12:50:25] in jenkins [12:52:48] aude: syntax error in a PHP file :-] [12:52:53] oooh [12:52:54] ok [12:53:04] if you click on the link jenkins-job gave [12:53:10] https://integration.mediawiki.org/ci/job/MediaWiki-GIT-Fetching/7307/ [12:53:36] the is a failing downstream job : MediaWiki-Universal-Linter#6611 [12:53:42] (there is a red bubble next to it) [12:53:42] ah, nice [12:53:51] related console: https://integration.mediawiki.org/ci/job/MediaWiki-Universal-Linter/6611/console [12:54:06] of course the message in Gerrit is a bit cryptic :-] That will be improved "soon" (tm) [12:55:13] it's helpful once i know where to look [19:04:49] Gerrit's down! [19:04:58] Gerrit's up! [19:06:06] Exercise [19:06:40] oooh..fundraiser banners [19:06:54] blue:) [19:07:19] <^demon|lunch> marktraceur, Reedy: I found the bug. Gerrit is singing Head Shoudlers Knees and Toes to itself. [19:07:37] <^demon|lunch> And every time it touches its toes, it goes down. [19:08:32] "Jenkins CI and Gerrit Code Review dance together" [19:09:33] Jenkins is aiding and abetting [19:09:34] <^demon|lunch> It's a very simple dance to do. [19:34:40] Reedy: how / why is wikidatawiki actually using the searchindex table? [19:38:15] AaronSchulz: do you know? ^^ [19:38:32] that I don't know [19:38:43] * AaronSchulz is work on job deduplication atm [19:38:56] *working [19:39:15] cool [19:39:19] i want to truncate the table [19:41:03] AaronSchulz: how do you feel about stopping memcached multiwrites? [19:41:35] I deployed the profiling last night [19:43:08] <^demon|lunch> binasher: Insofar as most wikis are using it. But for the structured data (which it's actually for), it sucks at searching that. [19:43:08] there are still about 25k gets/sec falling through to the old one (out of ~100k/sec) [19:43:12] <^demon|lunch> Which is why the push for Solr. [19:43:36] binasher: I wonder if those are just keys that are not in either though [19:43:42] ^demon|lunch: none of our wikis should be using the searchindex table [19:43:44] like those djvu metadata file keys [19:43:56] AaronSchulz: that is likely [19:44:20] <^demon|lunch> binasher: Oohhh, the table itself? I misread. [19:44:23] the drop off (from warm up) looks mostly done [19:44:24] <^demon|lunch> I have no clue... [19:44:38] binasher: so I'd say you can turn it off now [19:45:45] looks like mc hitrate on the old setup over a very long timeframe is 86% [19:55:47] yeah, wikidatawiki is the only wiki using the searchindex table [19:56:45] ah, it doesn't have the MWSearch extension [19:56:52] can we change that? [20:14:07] AaronSchulz: ^^ [20:15:25] binasher: poke Reedy maybe [20:15:42] Reedy: poke [20:15:45] or csteipp [20:15:54] Hmm? [20:16:03] why doesn't wikidatawiki have the MWSearch extension installed? [20:16:47] That actually doesn't sound familiar [20:17:14] I sec reviewed the wikidata specific extensions, but I'm pretty sure MWSearch was not on my list... [20:17:33] I do remember a conversation about lucene, but I'm not sure where it was left [20:18:08] csteipp: all of our wikis are supposed to have it enabled [20:18:25] Oh! Sorry.. [20:18:41] ^demon|busy handled most of the configs for actaul deployment. [20:19:20] <^demon|busy> csteipp, binasher: Whoops! [20:19:22] <^demon|busy> On it. [20:19:26] thanks! [20:21:44] <^demon|busy> I don't know why it's not being included, I don't see any conditional for it. [20:23:25] <^demon|busy> Oh, herp-derp. [20:23:26] <^demon|busy> Nevermind. [20:25:22] i was trying to figure that out and failed [20:25:32] \o/ greets to Bug 14901 - Email notification mistakes log action for new page creation [20:25:47] <^demon|busy> I'm not sure why Reedy did this. [20:25:59] it jsut reached the third position in most reported oben bugs rank https://bugzilla.wikimedia.org/duplicates.cgi [20:29:04] AaronSchulz: oh, mysql-multiwrite is still going to ObjectCache::newMemcached + mysql [20:30:42] it's still ok to take it out of the multwrite cache in mc.php [20:31:28] binasher: I guess we should also add pecl mc to mysql-multiwrite soon [20:31:37] <^demon|busy> binasher: Enabled it for wikidatawiki. [20:32:27] <^demon|busy> binasher: Is there an index yet, or do we have to wait for something? [20:32:38] that's a good question [20:32:51] notpeter: do you know if wikidata got indexed? [20:32:54] <^demon|busy> I'm not getting any search results. [20:33:56] well the update hooks were consumed by the mysql table index and not mwsearch [20:34:08] I think there is a script to rebuild the index [20:34:47] it didn't [20:34:55] and we have to wait for search to actually function [20:35:31] * binasher is itching to truncate table searchindex [20:35:35] it's my second favorite command [20:35:51] it's no drop, but it ain't bad [20:35:52] <^demon|busy> Well, it won't get any larger now. [20:36:07] you dn't have a truncate throw pillow, though [20:36:09] <^demon|busy> binasher: Can we please hang onto it until lucene catches up? [20:36:10] just sayin' [20:36:45] AaronSchulz: should we just make Objectcache::newMemcached return MemcachedPeclBagOStuff? [20:37:12] not really, few places are hardcoded to use that function [20:37:28] it stays that way for b/c [20:38:11] ^demon|busy: sure, but the table won't get queried with MWSearch enabled, will it? [20:38:22] <^demon|busy> No, it won't. [20:38:30] <^demon|busy> Shouldn't. [20:38:49] <^demon|busy> I'm shooting in the dark here--I've never futzed with search. [20:39:35] it sounds like we need to leave it enabled to get it index, so it'll have to stay broken for a bit [20:39:56] <^demon|busy> Life goes on. [20:40:09] <^demon|busy> It's not like the mysql search of structured data was actually decent at all. [20:41:41] AaronSchulz: what do we need to replace 'factory' => 'ObjectCache::newMemcached' with in a multiwrite clause? [20:46:00] * AaronSchulz looks [20:46:09] ah, should just be "$wgObjectCaches['memcached-pecl']"? i'm wondering if immediately swapping to the new cluster would kill pc1 ;) [20:59:46] hey mlitn, are you deploying anything? [21:00:12] hi, no I'm not [21:00:24] mlitn: cool, thanks [21:20:18] Reedy, are you there? [21:22:23] Any ops from labs available? [21:23:49] RobH [21:24:40] !ops [21:24:54] jeblad_WMDE, you might get a better answer in #wikimedia-labs itself, what was it you were looking for in particular? [21:29:10] lol [21:29:29] jeblad_WMDE: yeah ask in #wikimedia-labs, as !_ops just pings the irc ops for this channel :D [22:11:56] mutante: what sort of dumps are you importing for wikivoyage? [22:12:03] I mean, SQL dumps or XML? [22:12:21] do you import all data in it or something gets lost [22:13:08] i.e. will everything you got end up in dumps.wikimedia.org eventually or not [22:14:07] Nemo_bis: sql dumps [22:14:34] Nemo_bis: oh, i dont think we talked about putting it on dumps yet.. ehmm [22:14:40] good question [22:14:53] that's automatical [22:15:03] Nemo_bis: the dump files i get have already been processed by others before i get them [22:15:10] oh, who? [22:15:31] csteipp and/or Hansm and/or mmullie [22:15:45] I suppose the wikivoyage e.V. admin has to remove private data of users who didn't agree to the transfer [22:15:54] exactly [22:15:57] do you know what they do on it? [22:16:03] because just some users have agreed to the transfer etc [22:16:03] (on WMF end) [22:16:34] oh also, have you seen what a lovely use case of wikistats? http://hexmode.com/2012/11/mediawiki-support/ :) [22:16:41] Nemo_bis: some had to be converted from postgre to mysql too [22:17:18] cool!:) yeah, hexmode upgrading campaign is nice:) [22:18:22] Nemo_bis: csteipp has more details on what else exactly [22:18:41] fixing page titles [22:18:52] something related to a custom namespace... a footer..ehmm [22:18:52] oh [22:19:21] well none of this should make us lose actual information [22:49:08] Nemo_bis: wikivoyage user account names must not be reregistered. Otherwise a sul-user could get all edits done my another user conenected to his account [22:49:55] Merlissimo: are you replying to something I said elsewhere? [22:50:26] I suppose the wikivoyage e.V. admin has to remove private data of users who didn't agree to the transfer [22:50:51] of usernames are not private data iirc [22:50:53] i just wanted to be sure that not simply all these user accounts are deleted before dump is created [22:51:13] they've all been renamed [22:51:16] AFAIK [22:51:38] I suppose they'll justr drop email and password (or only email?) [23:11:34] is it possible to get a private wiki as an import source on a public wiki? [23:13:32] Thehelpfulone: Private to edit, or private to view? I imagine the latter might be more difficult :) [23:14:12] to view, to import from officewiki -> foundationwiki - yeah I thought it would be a bit difficult too but one can wish! [23:14:54] Thehelpfulone: I think it's *possible*, but it would probably require some magic on the backend (like setting up API access for officewiki)....I'll let someone else continue answering [23:16:32] ori-l: spagewmf: you guys done with your deploy window for the day? [23:16:36] okay, do you have any idea if it's a fair bit of work on the backend side? [23:16:43] mwalker: not quite, have a fix-up to deploy in a few [23:16:51] mwalker: are you guys in a hurry? [23:17:05] nope; just wanting to make sure we're all coordinated [23:17:30] thanks -- is it ok then if we take a bit longer? [23:17:42] totally Ok [23:31:52] mwalker: done! thanks again [23:31:59] awesome possumes [23:35:17] ori-l: you have two undeployed commits on fenari [23:35:28] E3Experiments and EventLogging [23:35:34] pgehres: ugh, sec [23:43:58] paravoid: https://labsconsole.wikimedia.org is down [23:44:02] hmm, he is away [23:45:47] It's like 1am for him I think [23:47:20] No it's not [23:47:21] He's in SF [23:47:25] And has been for a few months :p [23:50:50] * Damianz notes to tag with gps tracker so as he can not care about his movements :D [23:51:32] !log updated payments cluster to 7d92a1a35f0 [23:51:40] Logged the message, Master [23:58:18] gn8 folks