[02:48:01] I feel like this shouldn't be possible. https://en.wikipedia.org/wiki/Special:Contributions/8.8.8.8 [02:48:27] or shouldn't be happening, at least. [02:49:34] QuelqueChoseRose: It is weird that something like this would happen. [02:50:21] seems more likely, to me, that it's a problem on our end than on Google's [02:50:26] QuelqueChoseRose, I have a feeling someone at Google is responsible for that edit :) [02:50:58] perhaps. i'd think you'd have to be pretty high-up to use a public DNS address as your IP...? [02:51:16] QuelqueChoseRose: I really doubt it's on our end, though. [02:51:41] That IP is used for Google Public DNS. [02:51:45] I don't know how Google manages it's internal staff permissions/access/etc., QuelqueChoseRose [02:51:59] There are some edits on Wikimedia projects from 127.0.0.1 [02:52:10] But those are from WMF sysadmins. [02:52:18] rastus_vernon: except for when they're not [02:52:22] like the past couple days [02:52:43] Is it possible in... some way to fake an IP address? To fake from where it comes? [02:52:48] I mean, I know it's not possible trivially. [02:53:29] rastus_vernon: we've had multiple bugs in recent months with IP edits being mistakenly attributed to WMF IPs. something with the XFF headers [02:53:42] Ah, so you mean it has been common recently? [02:54:14] rastus_vernon: well, with our IPs, yes. but I haven't heard of it happening with another company's IPs [02:54:23] hrmmmmm [02:54:27] though there was that trouble with Apple a few weeks back [02:54:36] Google Public DNS' IPs are only used for DNS resolution. [02:54:46] So clearly it isn't related to that. [02:54:49] Trouble with Apple? [02:54:49] If this is indeed a error on our end, then it is necessarily because that IP was chosen... randomly? [02:54:55] neither of you are checkuser, right? [02:55:10] Really, I can't find an explanation to that. [02:55:15] i'm not [02:55:28] Krenair: someone was creating a lot of test accounts on en.wp, and there was suspicion that it was someone working for APple [02:55:40] wat? [02:55:42] Either the software rarely but unpredictably decides to use a random IP, either it's really a Google employee playing around, or it's a problem in the internet infrastructure. [02:55:43] Apple? [02:55:46] All of those seem very, very unlikely. [02:55:58] why would apple want a load of test accounts? [02:56:05] Krenair: i'll find you the link, 1 sec [02:56:33] rastus_vernon: hold your horses. first we just need a checkuser. then i can maybe report to google depending on what they come up with [02:56:35] rastus_vernon, I think we can safely discard the random IP theory :) [02:56:38] * jeremyb goes to find one [02:56:53] Will a checkuser be able to share the results with us jeremyb? [02:56:59] Surely there's a IRC channel for checkusers. [02:57:03] Krenair: maybe... [02:57:07] We have a IRC channel for everything, right? [02:57:11] Krenair: we shall see :) [02:57:14] Coren: around? [02:57:29] I don't think he is QuelqueChoseRose [02:57:36] Krenair: If the checkuser cannot share the raw results with us, he can certainly tells us something about the nature of the problem. [02:57:39] he is marked as away as well [02:57:44] oh. oops lol [02:57:49] But I don't know what exactly a checkuser can do to help us. [02:58:04] What can they do other than finding the IP addresses that were used by a user? [02:58:05] uhhh, the rest of the XFF chain? if it exists? [02:58:08] QueenOfFrance: What's up? [02:58:15] lol, ok then [02:58:16] Coren: https://en.wikipedia.org/wiki/Special:Contributions/8.8.8.8 [02:58:25] Coren: you're talking to QuelqueChoseRose not QueenOfFrance [02:58:25] Coren: please check for extra XFF? [02:58:29] * QuelqueChoseRose changes his nick so people stop accidentally pinging Snowy [02:58:49] Checkuser is a tool I haven't really played around with much, am not 100% sure what they can actually see [02:59:20] Do IP address assignment mishaps happen? [02:59:50] ... dafu? From 8.8.8.8? [03:00:07] Coren: Yes. The IP address for Google Public DNS. One of them, anyway. [03:00:09] Krenair: this, btw: https://en.wikipedia.org/wiki/Wikipedia:Administrators%27_noticeboard/Archive254#Mass_creation_of_automated_.22Keychain.22_accounts_from_Apple [03:01:18] Coren: oh, wtf, i didn't realize it wasn't a recent edit. maybe it's too old to check [03:01:40] although we are now logging sampled XFF? [03:01:50] 8.8.8.8 is funny to say out loud. sounds like Éponine, kinda [03:01:58] (i mean sampled everything which includes full XFF) [03:02:45] not recent? it's under 2 months old... [03:03:56] Krenair: i thought we only keep ~2 months worth? [03:04:13] okay good, that means we can still get the info I hope? [03:04:14] There wouldn't have been anything useful to check, most likely; and 8.8.8.8 is used by google as an anycast address so I'm a little confused how a whole http post could have gone through. [03:04:16] I'm curious as to how Google got the IP address 8.8.8.8 in the first place, though. [03:05:05] rastus_vernon: Peering agreement with L3 [03:05:15] they reserved 8.8.8.* from level3, who control 8.*.*.* [03:05:28] Krenair: errr, 8.8.4.4? [03:05:56] Coren: ah. yeah, the existence of an edit at all is what confused me. does the Foundation have contacts at Google for this kind of thing? or is there any chance it's an MW glitch like what we've been seeing with the squids and localhost? [03:06:01] Coren: yeah, me too. but if it was really from a different IP which was whitelisted for XFF and sent "XFF: 8.8.8.8" that could be possible [03:06:02] oh, they have 8.8.4.* as well it seems, and probably a lot more [03:06:26] I think they have 8.8/16 [03:06:45] Coren: and i would expect that difference to be reflected in checkuser [03:07:37] Yeah, it's funky XFF. I think someone has a badly configured proxy. [03:08:46] Level3 has 4.2, at least. [03:09:25] So one of our trusted XFF servers is sending bad data Coren? [03:09:32] Which one is it? [03:09:54] I'm looking it up now. There's more than one intermediate. [03:10:28] * Jasper_Deng wonders if Coren thinks this has something to do with our new caching center [03:11:21] AS44444. "Websense Hosted Security"; I think their little proxy isn't as secure as they tout it. :-) [03:11:38] ugh, why is git.wm.o sending me HTTP 500 ISE again? [03:12:01] what in a list of repositories can possibly cause an internal server error? [03:13:26] 44444? [03:13:27] really? [03:13:54] Yeay coincidence. :-) [03:14:05] Coren: so there was actually CU data available? [03:14:24] Jasper_Deng: i doubt it's related to ulsfo [03:14:30] Krenair: Might have to do with libcurl? [03:14:32] What I don't get is that the message left there is completely innocuous -- not at all the kind of thing you'd expect someone to go through the trouble of faking their IP [03:14:41] rastus_vernon, wat? [03:14:59] No; that was actually from the UK and gotten through AMS. Not related to ulsfo at all. [03:15:01] I doubt the user who made the post is at fault here [03:15:18] Krenair: Try setting GIT_CURL_VERBOSE=1 [03:15:27] Krenair: It might give you useful information. [03:15:32] jeremyb: Yeah, it's 90 days. [03:15:37] ah, cool [03:15:38] rastus_vernon, I don't even... what are you talking about? [03:15:54] you can't set GIT_CURK_VERBOSE in a web browser, lol [03:16:00] Krenair: Well, git.wikimedia.org gives you an internal server error, right? [03:16:02] yeah [03:16:10] Oh. I thought you were cloning a repository or something and it gave you that error. [03:16:20] .... no git.wikimedia.org is a web interface rastus_vernon [03:17:36] Cannot repositories be cloned from it, though? [03:17:43] I don't know, never tried it [03:17:49] I usually only ever clone repositories over ssh from gerrit [03:17:52] gerrit.wikimedia.org is a web interface too, but repositories can be cloned from it. [03:18:09] And github.com is also a web interface... and yet repositories are cloned from it. [03:18:09] Coren, so, are we going to pull this server from trustedxff then? [03:18:16] git.wikimedia.org seems like the perfect place for cloning Git repositories. [03:18:35] and git was never *supposed* to be a web interface AFAIK. gerrit is canonical, gerrit came first [03:18:59] Krenair: I don't know if that single incident justifies it; I'm guessing it has to be a bug -- we might want to poke them first. [03:19:00] Krenair: I can confirm that git.wikimedia.org is used as a Git server. [03:19:16] Coren, can't you check what other IPs they've spoofed? [03:19:24] "To make an anonymous git clone of core MediaWiki you can clone from https://git.wikimedia.org/git/mediawiki/core.git" [03:19:32] From https://www.mediawiki.org/wiki/Gerrit [03:20:21] rastus_vernon: ok. still better to use gerrit. i think? gerrit is canonical, gerrit is what everyone else uses. [03:20:25] Krenair: For most IPs, it's not obvious whether they are good or not. 8.8.8.8 stands out because it's well known. I can look for odd patterns in the past 90 days, but it's hard to tell. [03:21:42] i guess it could be as simple as a misconfiguration in what XFF *that* proxy trusts? [03:22:02] Also you might want to tell Tim Starling, who seems to have added websense to trusted-hosts.txt in the first place [03:23:47] Krenair: Also, I can't actually look that up from checkuser; I'll have to go snoop in the DB itself, and I'm hesitant to do that without at least a nod from LCA. [03:24:01] (https://www.mediawiki.org/wiki/Special:Code/MediaWiki/78622) [03:24:03] ugh [03:28:42] i wonder how we ensure these IPs don't change hands after being whitelisted? [03:29:07] maybe we should have analytics run reports on unused whitelistings [07:05:26] anyone else having trouble connecting to bugzilla? [07:07:15] !log Bugzilla isn't loading [07:07:33] Logged the message, Master [07:29:26] p858snake|l / tto: fixed [07:30:46] thanks for the heads up [08:13:42] yay! it works [08:13:44] thanks ori-l [12:55:23] deletion takes rather long, moving files even longer. I sometimes even get timeouts from simple file editing. ..... [12:56:59] apegos: ping [12:57:04] *apergos [12:57:15] !tabcompletion | Steinsplitter [12:57:15] Steinsplitter: Instead of manually typing another person's nickname in IRC, you can type the beginning of their name and hit TAB on your keyboard to get it autocompleted, like on the command line. This avoids misspellings. Just make sure it's the right person! [12:57:16] :) [12:59:07] https://commons.wikimedia.org/wiki/Commons:Administrators%27_noticeboard#DelReqHandler_problems [12:59:38] + Special:Nuke is broken. completly unusable.... atm. [13:03:41] the File:* thing is broken :/ [13:05:40] "the File:* thing"? [13:06:12] (sorry, i am not able to help with whatever's broken, just curious) [13:19:34] yes, see link [15:43:55] thumbnails for https://en.wikipedia.org/wiki/File:National_map_of_eastern_provinces_of_German_Reich_based_on_official_census_of_1910.jpg aren't working [16:02:37] jackmcbarn: me too but the error msg is empty... [16:02:50] maybe needs to wait for a weekday (tomorrow) [16:02:57] (is 500 ISE) [16:03:05]

Error generating thumbnail

Error creating thumbnail:

[16:03:13] apergos: ^ [16:03:36] https://upload.wikimedia.org/wikipedia/en/thumb/3/37/National_map_of_eastern_provinces_of_German_Reich_based_on_official_census_of_1910.jpg/779px-National_map_of_eastern_provinces_of_German_Reich_based_on_official_census_of_1910.jpg [16:03:44] https://upload.wikimedia.org/wikipedia/en/thumb/3/37/National_map_of_eastern_provinces_of_German_Reich_based_on_official_census_of_1910.jpg/120px-National_map_of_eastern_provinces_of_German_Reich_based_on_official_census_of_1910.jpg [16:38:32] jeremyb: the media thing is broken [16:38:45] huh? [16:38:46] problems when moving/deleting files to.... :/ [16:38:56] wow [16:38:59] https://commons.wikimedia.org/wiki/Commons:Administrators%27_noticeboard#DelReqHandler_problems [16:39:12] you could be more specific than "media thing" next time? [16:39:14] kthx [16:41:32] > [16:41:33] Lot of problems with deleting/moving files in the last weeks. --Steinsplitter (talk) 13:02, 3 November 2013 (UTC) [16:42:08] so when did you first mention this here? [16:42:17] if it's been a few weeks? [16:42:24] yes, see bugs [16:46:03] well, deletion and move has rarely worked sine we've been used swift, AFAIK [16:46:26] seriously? [16:46:34] ever* ;) [16:47:00] you can expect things to be a little bumpier with multiwrite i guess. but i don't know so much about how it works [16:47:09] multiwrite is back on again now [16:47:42] Special:Nuke is broken (for files)... [16:47:54] hey jeremyb, happen to know where the supposed source code for http://tools.wmflabs.org/wm-bot/logs/ is? [16:48:05] (you're listed on https://meta.wikimedia.org/wiki/IRC/Publicly_logged_channels ) [16:48:15] MatmaRex: errrr, that's C#. i don't touch C# [16:48:20] petan: or if you're around… [16:48:34] jeremyb: i don't want you to fix it, i want you to let *me* fix it. :P [16:48:34] MatmaRex: (i only deal with configuration and restarting after e.g. netsplit) [16:48:38] okay [16:48:56] MatmaRex: so i can find the deployed version. idk about version control though. have to run soon anyway [16:49:02] https://meta.wikimedia.org/wiki/Wm-bot#Source [16:49:45] ah. thanks Nemo_bis [16:49:48] Steinsplitter: i see some bugs filed. most are not about speed and the one semi-related to speed (56487) is possibly only seen for a few hours period? [16:49:59] Steinsplitter: you should give bug #s... [16:50:04] (it would probably be a good idea to link the source from the logs page, too…) [16:50:13] Steinsplitter: as a general rule. not just "see bugs" [16:50:53] jeremyb: ... i dos not know the bug urls atm. [16:52:14] jeremyb: https://bugzilla.wikimedia.org/show_bug.cgi?id=53770 about moving files [16:52:27] Steinsplitter: that's not about speed [16:52:36] jeremyb: no, ther is no about speed [16:52:42] haha [16:53:13] coucou tout le monde [16:54:37] Steinsplitter: this is not helpful. you respond to a section on commons where everyone's talking about speed (or "slow") and you're not talking about speed? and you don't say what you are talking about? and then you link to it here and tell people to look at bugs but don't say which ones? [16:54:42] Steinsplitter: please stop doing that! [16:54:51] Steinsplitter: say what you mean, don't make me pull teeth [16:55:37] https://en.wiktionary.org/wiki/pull_teeth [16:57:10] jeremyb: this is admin noticeboard [16:57:15] and i can post ther what i want [16:58:23] i like to say ther is a gerneralproblem with the media thing [16:58:48] Steinsplitter: i'm making an honest attempt to make your reports to us better. they suck now. the alternative is just ignore you. i could do that if you prefer [16:59:49] Steinsplitter: part of this is surely a language barrier [17:00:08] Steinsplitter: but part of it is you give very few details and expect people to just figure out what you mean [17:00:10] wikimedia is multilingual ;-) [17:02:19] Steinsplitter: uhuh. you're writing in english. this is primarily an english channel. you used english at the link you linked too. if you prefer to use some other language you're welcome to (and then you'll be ignored by even more people because they don't have a clue what you're saying) but i think improving your use of the language you are actually using may be a worthy endeavor [17:02:35] anyway, not spilling any more words over this. you can believe me or not as you wish [17:04:25] You're rude to me ... [17:05:12] heh [17:05:16] Steinsplitter: no, he's being realistic [17:05:30] Steinsplitter: no one can fix your bug report if you don't say what the bug is [17:06:01] Steinsplitter: recommended reading: http://www.chiark.greenend.org.uk/~sgtatham/bugs.html [17:06:02] re rude: that's possible. maybe i'm even in a bad mood. who knows. but i think i gave some good advice [17:06:22] a lot of users has commented this bug. maby it is difficult to fix the bug. i dos not know. [17:06:46] and not oly i hav reported such problems ;-) [17:06:52] (and i don't intend to be mean. seriously, read that, and everyone will be happier) [17:07:01] (including you, finally having your bugs fixed :) ) [17:08:01] jeremyb: if you're still around, https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#Failed_action.3Dpurge might be related [17:08:28] jeremyb: apparently they managed to delete the files in the end, but it threw errors first (filenames are mentioned) [17:08:35] jeremyb: thanks aniway [17:09:10] MatmaRex: could compare timestamps against the other bug. i have to run. meeting in 50 mins and i'm not even ready to leave yet [17:09:44] vs. https://bugzilla.wikimedia.org/56487 [17:39:07] jeremyb: for when you're back, if you still want to investigate: [17:39:17] https://bugzilla.wikimedia.org/show_bug.cgi?id=54790 https://bugzilla.wikimedia.org/show_bug.cgi?id=53770 https://bugzilla.wikimedia.org/show_bug.cgi?id=38028 https://bugzilla.wikimedia.org/show_bug.cgi?id=53768 https://bugzilla.wikimedia.org/show_bug.cgi?id=49362 [18:11:41] usually on bawolff is able to shed light on such media storage bugs [18:11:44] *only [19:08:56] LeslieCarr: Every once in a while I notice some packetloss. Did you already fix the full ams-ix link? [19:15:31] Leslie's afk [19:22:52] Jasper_Deng: I know, at some point she'll see it. [19:23:13] But thanks for pointing it out :-) [20:11:34] sigh, will we ever have our Wikimania 2005 videos back https://archive.org/post/1003894/wayback-machine-doesnt-support-the-range-header-aka-wget-continue-doesnt-work [21:29:43] Krinkle's RTRC users here? [22:31:44] [[Tech]]; Mathonius; /* Luxotool/guc */ new section; https://meta.wikimedia.org/w/index.php?diff=6250374&oldid=6232210&rcid=4648886 [23:00:52] [[Tech]]; MZMcBride; /* Luxotool/guc */ +reply; https://meta.wikimedia.org/w/index.php?diff=6250552&oldid=6250374&rcid=4648933