[00:03:31] matanya here, what's up? [00:03:54] hi there spagewmf I tested the new login page [00:04:20] what is the checkbox "ignore the blacklist" ? [00:04:58] matanya, that should only show up if you're already logged in, and probably only if you're an admin. [00:05:39] I guessed that, but what does it mean? which blacklist? the local one? or the global one? [00:07:18] I'm not sure. Extension:TitleBlackList adds that checkbox, while there's also a separate GlobalBlocking extension. Checking more... [00:14:05] Susan: Can EdwardsBot post messages to multiple projects? [00:15:04] kaldari: Yes, but you'll need to do it from Meta-Wiki. :-) [00:15:20] kaldari: https://meta.wikimedia.org/wiki/GMD [00:16:16] matanya, I believe WMF wikis use http://meta.wikimedia.org/w/index.php?title=Title_blacklist&action=raw&tb_ver=1 as wgTitleBlacklistSources , so the title blacklist here is global. [00:16:34] Yes. [00:51:57] greg-g: I'm going to do a quick config change to disable WikiLove per bug 47457 [00:52:29] I assume no one else is deploying anything currently? [00:52:43] https://gerrit.wikimedia.org/r/#/c/60947/ [00:52:53] is the change [00:53:27] kaldari: I'm deploying VE right now [00:53:29] Will be done very soon [00:53:35] exciting :) [00:53:36] Actually it's done just now :) [00:53:40] Go ahead [00:53:43] thanks :) [01:21:07] spagewmf: errrr, = sounds wrong. [01:21:12] https://wikitech.wikimedia.org/w/index.php?title=Server_access_responsibilities&diff=0&oldid=68013 [01:21:43] probably fixing capitalization is sufficient [01:22:39] jeremyb, thanks for checking. ssh complained, I added = and that didn't help[, then changed capitalization. I have other lines "IdentitiesOnly = yes" so guessed at the =. [01:23:18] i think your IdentitiesOnly is wrong too then :-) [01:23:40] also, someone had earlier capitalized and now you reversed. not sure which is better but maybe you two should discuss [01:23:50] (err, that's about headings not the setting) [01:25:13] $ vagrant ssh-config 2>/dev/null | egrep ' no$' StrictHostKeyChecking no PasswordAuthentication no [01:25:20] grrr, someone ate my newlines [01:26:34] I Will Fight to the Death to Abolish Over-Capitalized Headings. They're nearly as distancing and unfriendly as passive voice has been found to be. [01:26:47] hehe [01:26:52] do you edit enwiki? :-) [01:27:34] yes, occasionally as skierpage [01:29:34] heh, > You're smart, you can figure out. [01:33:04] * jeremyb_ wonders if spagewmf wants to join planet wikimedia [01:36:01] spagewmf: Heading tweaks look good to me. [01:36:04] Sentence case yay. [01:36:49] I always get Skierpage and Skier Dude confused. [01:43:24] heh, maybe WMF got Skier Dude and me confused, thinking I'd made a beelyun edits [05:51:57] who deals with upload wizard? [05:52:23] uhhh. can't remember [05:52:32] matanya: what's wrong? [05:52:48] can't import from flickr [05:52:54] tell me more [05:52:56] Neil K. [05:53:01] Heh. [05:53:10] well i thought that wasn't a useful answer [05:53:15] so i refrained [05:54:03] jeremyb_: are you an admin at commons? [05:54:09] i isn't [05:54:34] so I can't show you :( [05:54:39] no one awake in the commons channel? [05:54:56] whym is awake [05:55:04] ohio! [05:55:48] matanya: you could take a screenshot or copy/paste to a pastebin or i guess a number of other htings [05:55:51] things* [05:56:27] nothing to show, I prees the button "get from flickr" and nothing happens [05:56:41] you could have said that :) [05:56:43] browser? [05:57:49] ff on ubuntu [06:08:42] weird, it works on chrom [06:08:43] e [06:10:09] matanya: hit ctrl-shift-k in firefox [06:10:20] matanya: make sure the JS button is enabled [06:10:41] matanya: reload the wizard and break it again and tell me what the error messge is [06:10:49] ok, a sec [06:17:00] jeremyb_: it is one of my add-ons [06:17:18] which? [06:17:28] checking [06:17:41] I suspect https-wnywhere [06:18:52] yes, approved [06:19:12] what's the issue? [06:19:40] not sure, when disabling it, it works [06:20:07] I think I know [06:20:19] well what's the msg? why did you suspect it? [06:20:42] the tool requests the http version of flickr, and https doesn't allow this [06:21:06] I suspected since every weiredness I see on wikimedia sites it related to that add-on [06:31:40] hehe [10:32:27] lunch time [11:32:58] hello there [11:35:30] anybody out there? -where do I report a possible problem with wikidata? [11:35:52] https://wikidata.org/wiki/Wikidata:Contact_the_development_team is not urgent [11:36:03] *if [11:36:09] Nemo_bis: you involved? [11:36:23] nope, don't look at me *runs* [11:37:56] So the very basic problem: I create a new article, while logged into a Wiki (Simple English), at some point I get "add links" (to edit the iw's in wikidata). While I am not logged in into WD, I am in SEWP. So what keeps the system from transmitting my login details to WD? [11:39:26] have you enabled cookies? :) [11:40:08] wikidata can read cookies from enwiki? [11:41:13] Nemo_bis: new iceweasel does not accept third-party cookies [11:41:47] btw. no cookie from enwiki is marked as secure when using https [11:44:12] i was wrong... when loggin in with https the centralauth_auth session gets the secure flag [11:44:25] but not host-only [11:45:05] anyway, when I do the same in dewiki, I don#t get the you need to be logged into wikidata thing. [11:46:21] so its definitely a problem. [11:58:26] eptalon: ah, yes, you're too ahead of time ;) there's a discussion on wikitech-l and a solution is being worked on [11:59:21] thx [11:59:39] so I can lean back and don#t have to worry any more [14:19:20] hello anyone here ? who an fiks a wikidata link : joining two wikidata-id to one wikidata -item [14:19:53] errr, i don't understand the question. but... go ask #wikimedia-wikidata [14:20:20] okay thanks for the channel-name [14:20:25] will do that... [15:26:09] anyone with server access have a second to run a query? I need to check if the toolservers database is corrupt [15:26:43] greg-g: just a heads-up: https://bugzilla.wikimedia.org/show_bug.cgi?id=47457 comments 43-45 imply that the problem is getting even worse [15:27:02] Betacommand: ask Reedy [15:27:12] "After WikiLove being disabled, this starts appearing on [[Special:Preferences]] [15:27:12] on zhwiki." what?! [15:27:37] Reedy: poke [15:28:46] andre__: yeah :/ saw those last night [15:28:54] ok [15:28:55] * greg-g is waiting for kaldari [15:29:04] sumanah, yeah, things are getting weird. [15:38:53] andre__: sumanah just chatted via voices with roblaAWAY and James_F|Away , we'll get Krinkle|detached to get on it asap [15:38:57] ok, thank you [15:42:25] thanks [15:44:15] andre__: my comment was about the previous situation, I referred to a 23rd April comment [15:44:45] maybe stryn would know if it happens again [15:51:54] Betacommand: ? [15:54:36] Reedy: can you run select count(*) from imagelinks where il_to = 'Flag_of_Seattle.svg'; on enwiki ? [15:55:25] 2 [15:55:29] Reedy: 26 15:26:09 < Betacommand> anyone with server access have a second to run a query? I need to check if the toolservers database is corrupt [15:55:44] jeremyb_: Yes, I can read thankyou :p [15:56:06] Reedy: you're welcome! [15:56:08] dam, that means the toolserver db is trashed [15:56:29] surprise surprise [15:56:33] Reedy: Im getting 68 [15:56:33] Betacommand: is that not the default? [15:56:59] jeremyb_: the DB isnt normally [15:57:20] Betacommand: huh. i thought it was like at least 1/4 of the time. (maybe more than half even) [15:57:41] jeremyb_: was 5% until very recently [15:59:32] jeremyb_: hopefully -labs will take care of these issues [16:00:31] Betacommand: sanitarium! [16:02:41] jeremyb_: WMDE wont spend the money the toolserver needs to help fix the issue [16:03:11] Betacommand: money? how do you figure? [16:03:45] jeremyb_: the toolserver needs more hardware [16:04:16] it was overloaded almost two years ago [16:04:28] errr, so overload causes corruption how? [16:04:47] since labs was announced all improvements where canceled [16:04:59] i could understand things being slow or lagged or disks full. i can't understand the data being wrong [16:05:10] so more hardware should not help [16:06:14] jeremyb_: during replication the servers end up dropping/loosing some commands [16:06:44] because they are severely over loaded [16:06:55] i think your statements don't hold water [16:07:08] citations or other people supporting you are welcome [16:07:45] jeremyb_: they are in the IRC/mailing list archives I really cannot be arsed to dig them up [16:07:58] ok, well i don't believe it [16:08:10] the toolserver needs several new servers and more rack space [16:08:33] which is certainly not relevant [16:09:20] there's a bug or misconfiguration somewhere which causes the data to be wrong. just fix that [16:09:51] and maybe use percona stuffs to detect incosistencies earlier than later [16:10:37] jeremyb_: its far more complex than that, Ive seen half a dozen different causes for db corruption over my time at the toolserver [16:11:13] Betacommand: ok, so rinse, repeat [16:11:28] Betacommand: something's really screwy if this issue can be solved by more hardware [16:12:11] (unless more hardware means you can use different software/config. e.g. maybe you'd just eliminate trainwreck or it's replacement altogether if you had enough HW) [16:12:19] jeremyb_: why do you say that? if a server is doing 250% more than it should it will not function correctly [16:12:49] Betacommand: elaborate [16:14:24] the current servers are old hardware and its not uncommon to see a single TS server replace 2 or more WMF servers [16:14:44] that kind of load means things will break [16:15:48] plus the number of toolserver users is increasing while the hardware sits in a cicra 2009 timeframe [16:16:30] well I think that doesn't add up [16:16:52] if you have a server with high load or iowait or CPU use, so some things will wait to be done [16:17:05] but when they are done the answer should be correct [16:17:20] at least at the micro level where a process is not relying on any other processes [16:17:35] e.g. `echo 1+1 | bc` should always be correct [16:17:42] regardless of what else is running [16:17:46] jeremyb_: this isnt a micro level [16:17:50] mwalker|away: heya, question on the new lua module hotness for doing the deployment calendar: when I 'archive' old entries at the bottom of the page (then, at the end of the month, on a monthy archive page) will that update the date of the deploy to os.date at the time of archival? ie: a deploy that happens this week, when archived, will have it's date incorrectly set to the wrong thing? [16:18:03] mwalker|away: that's at least how I read the lua module :/ [16:18:27] mwalker|away: sorry for not reviewing it until now [16:18:37] jeremyb_: add the stress workload and the fact that we have a very limited staff (and multiple OS environment ) things cannot be maintained effectively [16:18:43] Betacommand: well it's a level where mysql relies on no other process to do the replication (at least with vanilla replication. idk about this trainwreck business). [16:19:25] Betacommand: so it's not e.g. a web app where the app and DB might independantly both work fine but the web app has a timeout configured and doesn't wait long enough for the DB [16:19:45] Betacommand: anyway, i think your arguments don't hold water [16:21:00] Betacommand: https://gerrit.wikimedia.org/r/gitweb?p=operations%2Fpuppet.git&a=search&h=HEAD&st=grep&s=sanitarium [16:21:50] jeremyb_: I guess we will have to agree to disagree. the needed resources to keep the TS in fully functional condition just are not there, and this will continue to get worse as time goes on [16:22:28] i'm not saying it has resources to functional or useful [16:22:46] i'm saying resource saturation is no excuse for wrong data [16:24:08] jeremyb_: if you dont have the resources (be it network connection, servers, disk space, or man power) to maintain a system the system will start to degrade and collapse given enough time [16:25:36] i earlier claimed that none of high load, high CPU use, high iowait would explain the answers being wrong. late or slow or lagged, sure. wrong, no. [16:25:39] i stand by that [16:27:05] mwalker|away: oh, I think I see how it should be done.... [16:27:33] jeremyb_: how about frequent interruptions of replication? [16:28:27] Nemo_bis: ahhh, so idk exactly what you mean but you remind me of an issue I've discussed fairly extensively with nosy and some with asher too [16:29:15] Nemo_bis: sometimes the WMF clusters do some action which never makes it into the replication feed at all. or do a master switch and the TS keeps using a non-canonical master [16:30:17] Nemo_bis: one of my solutions for those cases was to have a poison pill of sorts where the WMF could do something that would break the toolserver replication at exactly the right place and would not break the WMF replicas [16:30:51] Nemo_bis: then send instructions to TS admins for what to do before skipping the poison pill [16:31:17] but this doesn't remove the need for communication which is the biggest problem [16:31:39] i think that can be fixed... [16:31:52] maybe [16:32:18] anyway, i don't think that's a terribly common cause of incostencies. well mostly. idk about the master switches [16:32:30] and now we're using mariadb and TS i guess is still mysql [16:32:38] inconsistencies* [16:32:56] mwalker|away: but that way I just tried isn't working as expected.... (the row isn't displaying with "|when=2013-04-29 11:00 PDT" instead of eg "|when={{#invoke:Deployment_schedule|dateOf|Thu}} 11:00 PDT" ) [16:39:58] * jeremyb_ has been summoned... [16:40:01] jeremyb_: https://wikitech.wikimedia.org/wiki/Module:Deployment_schedule [16:41:00] jeremyb_: see my last comment to mwalker, or, just see the broken deployments wiki page https://wikitech.wikimedia.org/wiki/Deployments (the first item for next monday doesn't show up because I've set it as a specific date) [16:41:13] in the "week of april 29th" section [16:44:38] greg-g: sorry -- just got in [16:44:52] so the problem is that the date isn't being parsed correctly? [16:47:10] mwalker: apparently [16:47:34] mwalker: check out the deployment calendar right now, seems broken (I only tried to edit the first entry for next monday in the "week of april 29th" section) [16:47:49] mwalker: saved it so jeremyb_ could see what i was doing [16:48:09] greg-g: > Sam / Aaron / RobLa MediaWiki general deployment window (1.22) [16:48:10] ? [16:48:22] it's at the bottom :) [16:48:25] it gets sorted by date [16:48:25] did i load the page too early maybe? [16:48:42] i'm not seeing an error... [16:48:55] ohhhhhhhh [16:49:00] whoa [16:49:07] ok, me goes back to editing [16:49:16] thanks mwalker ! [16:49:59] np -- someday someone (me or someone else) will figure out a way to make that subst automatically from the dateOf to an actual date [16:50:05] (and you, too, jeremyb_ ! ;) ) [16:50:58] greg-g: was it just broken in preview and worked on save? [16:53:40] jeremyb_: no, it worked, just not as I expected. I didn't realize it autosorted them, so, since it was the only one I explicitly set to next week (all others were this week) it went to the bottom when I expected it ot be on top [16:54:51] ahh [17:55:56] anomie: heya, so, our initial plan today was to have Krinkle|detached be on point for that slow user page bug, but he's not in today, I'm assigning the bug to you as you're now the best option to continue working on it :) [17:56:36] anomie: if you get stuck, feel free to ping RoanKattouw_away, James_F will warn him you might ;) (he's doing school stuff today, but "available") [17:57:35] greg-g: Oh, fun. (: [17:57:52] anomie: weee ;) [18:03:35] !log Upgraded fundraising civicrm install from a3db312836b6ac004694f831071b1c7d00560fbf to a65fbcec0dbb5bb79a97aca220a998e56d08b067 [18:03:43] Logged the message, Master [18:21:45] greg-g: Hrm. Since someone disabled WikiLove, I need to find a new way to reproduce the bug. [18:22:28] anomie: its still enabled on test.wp [18:22:35] right [18:30:07] how do I set $wgDefaultUserOptions for a particular WMF wiki? [18:30:58] how do I set $wgDefaultUserOptions for a particular WMF wiki? [18:31:31] no need to repeat yourself... [18:31:36] you could file a bug [18:31:40] which wiki? [18:31:48] do you have local consensus to make the change? [18:32:17] jeremyb_: repeating myself was a mistake though … accidentally pressed enter twice [18:32:51] ok :) [18:33:26] well, I know how to change CommonSettings.php … but I see no way to change $wgDefaultUserOptions for a particular wiki [18:33:41] as CommonSettings.php defines it for /all/ projects [18:33:47] if I see it correctly [18:33:58] you should try other files then :) [18:34:04] InitialiseSettings.php ? [18:34:19] anyway, which wiki and do you have consensus? [18:34:37] I find no $wgDefaultUserOptions for particular wikis there either :S [18:34:58] hevoy [18:35:20] so maybe you're the first... [18:35:31] in any case, consensus? [18:35:59] and we need to disable an option there which is enabled for all other wikivoyages …RTL text issues [18:36:39] discussion: [18:36:48] are you the only RTL wikivoygage thus far? [18:36:53] which option? [18:37:17] yep, the only RTL wikivoyage [18:37:29] see the problem here: http://he.wikivoyage.org/wiki/ויקימסע:קרון_המזנון#Bug_analyzed [18:44:45] jeremyb_: is this fixable? :) [18:49:26] Heinsten: i read that in a few hours. or maybe you can find someone else to look before then [18:49:50] jeremyb_: ty [18:50:03] Heinsten: bavakasha :) [18:55:50] Hallo jeremyb_ . [18:56:50] aharoni: Heinsten [18:56:55] Heinsten, jeremyb_ I think that just cleaning up the CSS is the right solution for that bug. [18:57:19] let me just try it on my wiki ;) [19:04:06] ori-l: heya, anything interesting from E3 happening next week? :) [19:04:23] ori-l: specifically, anything interesting being deployed? [19:07:15] spagewmf: ^^ you know of anything before I send out the deploy highlights email? [19:08:31] jeremyb_: would an entry like [19:08:34] # wgDefaultUserOptions @{ 'wgDefaultUserOptions['toc-floated'] => array( 'default' => true, 'hewikivoyage' => false, # @} end of wgDefaultUserOptions [19:08:46] work in InitialiseSettings.php? [19:10:02] aharoni: the CSS was already fixed [19:10:21] the TOC is already on the right side [19:10:28] the only remaining problem is toc-floated [19:10:45] Heinsten: I see that it was fixed in Common.css. It's better to fix it in the extension upstream. [19:10:57] So that the Arabic Wikivoyage would enjoy it, too. [19:11:11] wouldn't it be on the right side on /all/ wikis then? [19:11:18] I just tried removing all the rtl-specific styles and I think that it does the right thing. [19:11:28] No - CSSJanus flips it automatically. [19:11:33] ah [19:11:39] well [19:12:02] Heinsten: Just let me check one last time before I commit it: [19:12:18] aharoni: the CSS? [19:12:19] In English the floating TOC is supposed to be on the left, and in Hebrew the floating TOC is supposed to be on the right? [19:12:40] yes [19:12:44] OK, I'm committing. [19:13:04] but wait [19:13:31] toc-floated still needs to be disabled by default [19:13:47] at least on he [19:14:22] so would "# wgDefaultUserOptions @{ 'wgDefaultUserOptions['toc-floated'] => array( 'default' => true, 'hewikivoyage' => false, # @} end of wgDefaultUserOptions" in InitialiseSettings.php fix that? [19:14:49] Heinsten: why? [19:15:35] because disabling it moves the TOC to the top of a page where it should be [19:16:24] Heinsten: the community just doesn't want it to float by default? [19:16:29] yep [19:16:39] OK, no problem, but it would have to be a separate commit. [19:16:45] okay [19:16:53] the default settings are in a different Gerrit repo. [19:17:14] aharoni: could you check whether the code which I posted above would work? [19:17:18] I'm not sure [19:17:31] probably. [19:17:50] Heinsten: do you have a Gerrit account? [19:17:59] yes [19:18:05] I can do it [19:18:28] Heinsten: https://gerrit.wikimedia.org/r/#/c/61051/ [19:18:31] that's the CSS fix [19:18:41] very trivial [19:19:16] Jenkins failed it for some reason, but it's probably just a Jenkins bug. hashar, can you please check https://gerrit.wikimedia.org/r/#/c/61051/ ? [19:19:43] aharoni: jenkins is hosed atm [19:19:52] yeah jenkins restarting sorry [19:19:57] i will get it upgraded next week [19:20:01] No problem, thanks. [19:20:01] that will speed up restart [19:20:02] :( [19:21:05] Heinsten: What's your Gerrit name, so I'll assign you as a reviewer? [19:21:11] aharoni: so after the gerrit change is merged the other code can be removed from local Common.css, right? [19:21:22] Heinsten: yes. [19:21:35] Heinsten: more precisely, after it's deployed. [19:21:41] yeah [19:22:38] aharoni: my gerrit account? I dunno … either eikes or Vogone [19:23:40] Heinsten: apparently Vogone. got the Gerrit email? [19:23:51] let me check [19:25:13] * legoktm stares at Heinsten [19:25:20] aharoni: got it [19:25:44] legoktm: don't be faster than me again, please … [19:25:51] :P [19:29:10] Heinsten: another little commit for you: https://gerrit.wikimedia.org/r/#/c/61054/ [19:39:20] hi [19:39:28] is anyone any good with CSS and/or JS [19:43:59] Dragonfly6-7: yeah a bunch of people there should :-] [19:44:10] Dragonfly6-7: I guess the best is to ask your question [19:47:06] !g 61057 | greg-g [19:47:07] greg-g: https://gerrit.wikimedia.org/r/#q,61057,n,z [19:52:13] does anyone have a few minutes to help me with a complex but straightforward task or eight? [19:52:50] hashar - for instance, I'm trying to customize my css to get rid of those ugly "star" icons for watching articles [19:55:12] Also, I'd like to alter the relative font sizes used by various elements of the css [19:56:16] oh, and I need a proper "current version" link for histories, and a proper "my contributions" link [19:56:59] and how can I change the color of the separator bar? I mean, I know how to change hex codes, but what's the name of the separator bar element? [19:57:58] Dragonfly6-7: browsers usually have plugins to find out the name of the elements [19:58:05] Dragonfly6-7: which browser do you use? [19:58:34] Firefox [19:58:49] I have a highly customized JS and CSS [19:59:00] I'm working on replicating Classic within Vector (because Vector is ugly) [19:59:18] I've got a *lot* of it, but it's still missing some important stuff [19:59:41] and the biggest problem is that I know very little about CSS and JS, so I need other people to write the stuff for me [19:59:50] :-) [20:00:16] so maybe you could write the spec and find other devs to assist [20:00:22] wikitech-l mailing list might be a good start [20:00:52] I think it'd be easier to just make the changes here [20:00:56] quicker, anyway [20:02:22] would you mind taking a look? [20:03:19] I am awful at css :D [20:03:45] how about js? [20:12:33] can someone review this trivial fix for UploadWizard today: https://gerrit.wikimedia.org/r/#/c/60029 (so it gets in before the branch is cut) [20:20:20] having some issues with overwriting an image [20:20:22] http://en.wikipedia.org/wiki/File:Wncnlogo.png [20:20:32] StevenW: https://nl.wikipedia.org/wiki/Wikipedia:De_kroeg#New_account_creation_and_login <- ROFLMAO! No, we don't kill the messenger, we just torture him to death [20:20:56] Lirodon: ah, it looks like the logo thumbnail is not updating [20:20:56] Heh. [20:21:19] Lirodon: assuming that NBC 17 is the new logo ? [20:21:32] Lirodon: or is the WNCN with the little triangle the logo ? [20:21:47] multichill: It's okay. I don't expect Wikipedians to go crazy with joy that we're redesigning signup and login. It's not a big priority for them, obviously. [20:22:42] Also: the Dutch aren't as bad as the French, who spent half the time complaining that I posted in English. [20:23:12] Our village pump is called de Kroeg. That means something like a brown old bar. The people are acting like the obnoxious regulars in such a bar [20:23:26] Commenting on everything in a negative manner [20:24:44] StevenW - although people could become quite upset with the changes. [20:25:06] legoktm, WNCN one [20:25:11] see the article and it mentions why [20:25:21] Hi? [20:25:22] http://www.tvnewscheck.com/article/67114/wncn-emphasizes-local-in-new-branding [20:25:46] I'm not sure why you pinged me? [20:25:53] whoops [20:25:54] LeslieCarr, [20:25:56] :P [20:26:14] http://www.wncn.com/ see? [20:27:32] Lirodon: so what was the problem with the two previous uploads ? [20:27:45] greg-g: ping? [20:29:15] I thought re-uploading would work [20:29:19] but it didn't [20:29:41] besides, per NFCC, all but the correct version should be removed to begin with [20:32:16] i don't get what the problem was with the two previous uploads [20:32:35] i can't help if i don't know what the problem was [20:33:04] also, mediawiki shows all the previous versions, that's a feature [20:33:22] LeslieCarr, uploaded the new version. old one still came up, but in the size of the new one [20:33:42] tried re-uploading. still does it, but the previous revision now shows the one I had uploaded [20:34:32] ok, so unrevert this, then give me a traceroute [20:35:44] and preferably also take a screenshot so i can see what you're seeing [20:35:51] want to see if there's a purging problem [20:37:37] LeslieCarr, I think it is a purging problem [20:38:16] but i need the information, i need to see which location you're hitting [20:46:21] LeslieCarr, how do you do that on Windows? [20:48:19] http://www.mediacollege.com/internet/troubleshooter/traceroute.html [20:58:47] LeslieCarr, tried it with the full URL but Windows complained. trying just upload.wikimedia.org [20:59:45] legoktm, http://pastie.org/private/ozc5zonsp1bk7vttbyvb9w [20:59:48] whoops LeslieCarr [21:01:23] paravoid: here now, sorry, lunch out of office [21:01:34] no worries [21:02:07] do we need to do something to schedule Ceph for monday? [21:02:15] is it part of the general window? [21:02:29] it says TBD on the wikipag [21:03:31] paravoid: I assume there are mediawiki changes (just config?) that need to happen at the same time? [21:04:06] yes [21:04:48] just config afaik, AaronSchulz is the authoritative contact point though [21:05:04] paravoid: I think my answer is: if the the code that needs to be a part of MW Core is deployed during the normal MW window, but with Ceph turned off in config, then let's do that, and then turn it on an hour or so after the MW Core deploy [21:05:08] * greg-g nods [21:05:20] AaronSchulz: ^^ is that acceptable to you re Ceph deploy? [21:05:21] I don't think there's any code to be deployed [21:05:24] ok [21:05:33] but confirm with aaron [21:05:34] * greg-g nods [21:07:25] there is no new code [21:08:09] paravoid: let's schedule it for an hour after Reedy's done deploying wmf3 [21:17:37] greg-g: perfect, thanks! [21:18:06] paravoid: scheduled for 1pm PDT/20:00 UTC [21:18:10]