[00:04:44] Danny_B|backup: afaik their broken at the moment [00:04:48] poke reedy though [00:05:18] (the automated update part) [00:16:03] Aaron said it's indeed broken https://bugzilla.wikimedia.org/show_bug.cgi?id=42152 [00:18:30] It's even more broken than it possibly was before [00:19:08] I guess I should run it manually again in the meantime [00:20:02] it's running [01:16:15] Reedy: haven't forgotten the change re: zero-division in parser functions; i'll get to it tonight. [01:16:38] Woo ;) [01:16:51] The float one doesn't break the cases that broke last time at least :D [01:17:57] yay for regression tests [03:24:21] So I'm working on https://meta.wikimedia.org/wiki/Wikimedia_Blog/Drafts/Ops,_2011-2013#Draft_post and notpeter said, "our Puppet manifests for our databases were a file that just had a comment that said 'domas is a slacker'." (in early 2011) -- I'm poking aroundhttps://gerrit.wikimedia.org/r/gitweb?p=operations/puppet.git;a=tree;f=manifests;h=60b2e93dc9e9e13211a3306133e250080b6c2549;hb=HEAD and can't find that [03:25:13] where should I be looking? [03:28:43] or maybe that was before we moved all the manifests to the public repo in fall 2011, and we didn't import the history .... I think that must be what it is [03:28:55] I won't be able to access (or publicly link to) that old revision of the manifest [03:28:57] I know I've definitely seen it somewhere... [03:28:58] ding ding ding ding ding ;) [03:29:17] you should be looking at the non-public puppet SVN repo on sockpuppet/stafford [03:29:20] because those manifests used to like include passwords etc. :-) [03:29:33] well, there's no reason for me to do that unless I can link to it in the blog post [03:29:43] that's why I was looking [03:30:02] well if that's really all it says then it's not really a secret and you can probably publish it ;) [03:30:32] I'm already talking about it here in this public and logged channel so it's ok to say the words :-) but I just like linking to this kind of stuff where possible [03:30:33] 2009-11-03 23:04:01 thumbs domas is a slacker [03:30:33] 2009-11-03 23:04:18 ebergen I don't think he's even working [03:30:33] 2009-11-03 23:04:21 ebergen slacker [03:30:58] who's ebergen? [03:31:26] http://mysqllog.archivist.info/2009-11-03.txt [03:31:31] mysql logs apparently ;) [03:32:15] sumanah: I know I have seen it. But it might've been one of the other ops people pasting it somewhere... [03:33:35] It's no prob [03:33:48] Switched to branch 'production' [03:33:48] Your branch is behind 'origin/production' by 155 commits, and can be fast-forwarded. [03:33:48] reedy@ [03:33:49] lol [03:34:40] https://gerrit.wikimedia.org/r/gitweb?p=operations/puppet.git;a=history;hb=2e238bd8644b48bf8aab986968861ce876d19e49 [03:35:14] jeremyb: if you can point to the specific blog or diff..... [03:35:19] that will be unexpected & cool [03:35:48] sumanah: i still have to asusme it's something i don't have access to. actually this is the first i'm hearing that particular line [03:35:55] but really, I don't need it; being able to link to this specific fossil would be a nice to hanve [03:35:56] have* [03:36:20] yeah, I think the public doesn't have access to those incredibly early primitive manifests [03:36:23] that's okay [03:36:35] You might just need to ask the right person/people [03:36:36] I trust Peter [03:36:48] well, I don't really need special access [03:36:59] but thanks for the investigation [03:37:07] I'm meaning, someone from ops who's been around a while will probably have it somewhere ;) [03:37:17] sumanah: coming to the combined holiday party for a bunch of user groups in 3 days? [03:37:19] Also, is wikitech seemingly dead for anyone else? [03:37:25] it was, it seems up now [03:37:42] if I run into something later in the post that I *need* or really want to link to and feel like it ought to be publicly available then I'll nag people :-) [03:37:54] it's slow for me [03:38:02] It's not just you! http://wikitech.wikimedia.org looks down from here. [03:38:08] !log Wikitech is dead [03:38:10] :D [03:38:13] as in i'm still waiting for the first response [03:38:28] 09 02:25:34 -!- morebots [~morebots@wikitech.wikimedia.org] has quit [Ping timeout: 246 seconds] [03:39:40] lol [03:39:42] sumanah: it is dead, right? [03:40:07] http://www.downforeveryoneorjustme.com/wikitech.wikimedia.org [03:40:23] * sumanah skims https://blog.wikimedia.org/2011/10/28/tech-meetup-moves-wm-infrastructure-forward/ re mentions of the puppetization drive [03:40:34] jeremyb: so tell me about this party [03:42:15] sumanah: well details are on nylug.org; there's ~10 groups that have been involved in some way. puppet-nyc, NYLUG, devops meetup, and some others. i think a few lang specific meets and maybe also NYCBUG [03:42:48] it was on the Debian list too [03:43:49] hrmmmm, wikitech should be on watchmouse? ;P [03:44:50] I don't think I can make it, sadly, jeremyb - will you take on the onus of telling people that WMF is hiring? :-) [03:45:38] sumanah: i'm not so familiar with the current openings really [03:46:18] jeremyb: check out http://www.harihareswara.net/sumana/2012/12/01/0 for an overview of the most crucial ones, especially the storage engineer & search engineer ones [03:47:00] > get from deploying every two weeks to continuous integration, so everyone benefits faster from fresh code (release manager) [03:47:17] that doesn't make so much sense. at least as i understand the terms [03:47:24] do you mean continuous delivery? [03:47:50] I hear folks say CI synonymously with CD sometimes [03:47:56] huh [03:47:58] interesting [03:48:01] if you're more picky about that, then go ahead and substitute CD when you tell others [03:48:42] i was thinking CI is hashar's stuff and CD is pushing to prod and using it [03:48:53] ok [03:49:02] but i'm not necessarily right [03:49:28] anyway, that's a useful post ;) [03:49:45] Thanks! [04:03:08] sumanah: hrmmm, which ops people specialize in what. (central list) [04:04:13] so, there's some cases where you can say historically X has been the person to deal with Y. but they're not necessarily an expert in it and they're not really responsible for making it not break [04:05:44] that's happened a lot with some people. e.g. Tim Starling and ma rk [04:06:24] and i guess ariel too with some of the uploads stuff? although i'm less sure about that [04:06:38] Solaris! [04:06:52] and some things do have specific people responsible for them. e.g. labs [04:06:58] networks [04:07:09] Computers [04:10:05] so ops people I know a fair bit about what kind of stuff they do and some I seem to no very very little about. but i guess that's partly a function of people never showing up on IRC (or at least not in my channels). if i never see them then i (might) never hear about what they're doing [04:10:15] some ops* [04:10:22] to know* [04:17:13] i don't think i remember a time before nagios. but i can't say for certain if it was around 6 years ago [04:17:28] "a time before nagios" - like "The Land Before Time" [05:10:24] ok, https://gerrit.wikimedia.org/r/#/c/37353/1/wmf-config/InitialiseSettings.php has been merged, so we're on our way to true flickr upload to Commons yay [06:26:23] gah http://meta.wikimedia.org/wiki/Creating_wiki-torrents [06:26:47] didn't we get one of these 2 or 3 weeks ago on the foundation wiki feedback page? [06:30:46] jeremyb: Doesn't that user not know that bandwidth doesn't grow on trees? [06:31:12] torrents? [06:31:27] read the proposal [06:31:39] even more ridiculous than what I saw 3 weeks ago [06:31:43] are you referring to the torrents user or someone else? [06:31:50] do you has OTRS? [06:32:04] jeremyb: ? [06:32:14] ? [06:32:25] this has nothing to do w/ OTRS [06:32:35] that's what you think [06:32:38] do you have OTRS? [06:32:42] no [06:32:53] ok, i'll pastebin i guess [06:33:06] http://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28proposals%29#Creating_wiki-torrents [06:33:13] was original [06:33:57] I'm pretty sure that's never going to fly [06:34:14] for one thing the English makes it very hard to understand the specifics of it [06:36:37] hey, if someone wants to make a demo they can have at it [06:36:47] ticket is 2012112610000612 for anyone that does want to look [06:36:59] it's a stupid idea, but it's their time to waste :D [06:37:23] * jeremyb is rereading to see if there's anything identifiable i need to redact and then will pastebin and /msg the link [06:38:00] actually, i'm not even sure if i read the whole thing through the first time [06:43:33] Jasper_Deng_busy: you has /msg [06:46:13] Jasper_Deng_busy: i thought you said it had nothing to do with OTRS? :) [06:46:28] jeremyb: it has nothing /directly/ to do w/ OTRS [06:46:46] though the concept itself does, so it indirectly does [06:46:46] 09 06:46:07 Isn't this just the same as having gazillions of different datacenters? [06:46:49] 09 06:46:17 w/ some of them the rep masters? [06:47:47] so, he's talking about having 500MB be a higher usage node. (i guess that means stored data?). so most nodes would have less than that much [06:49:10] I guess it's even more decentralized than having many datacenters [12:05:06] !log job queue snapshot for bug 42614: en.wiki 900k still raising steadily, Commons 70k, fr.wiki 4.5M, Meta 2k, other tops 1k at most [12:05:14] Logged the message, Master [16:19:03] MaxSem: are you going to add {{#coordinates}} to all languages? https://en.wikipedia.org/w/index.php?title=Template:Coord&diff=526665541&oldid=368854805 [16:20:11] ah the extension is not installed yet [21:44:04] is anybody here, or is it still weekend in the States? [21:44:49] ah, turns out it's still weekend. i'll pester someone about https://bugzilla.wikimedia.org/show_bug.cgi?id=40124#c30 tomorrow, then. [21:44:57] It's still the weekend in the uk and since the states are 8 hours behind... it's still the weekend [21:45:09] Unless you're in china anyway [21:59:14] MatmaRex: errr, when is weekend in europe? [21:59:35] now [21:59:53] jeremyb: now, and for one more hour [22:00:02] (i mean, in CET) [22:00:06] one hour and seven minutes left in CET ;) [22:00:15] i don't think so [22:00:20] please reset your clock [22:00:22] -7 min, actually (evil watch :/) [22:00:34] 09 22:00:06 < Jan_eissfeldt> one hour and seven minutes left in CET ;) [22:00:36] that's UTC [22:01:02] maybe you forgot to wind it [22:01:35] I'm sitting in Munich right now and it is 23:01 locat time [22:01:38] jeremyb: it's 59 minutes left in CET for me now and i'm quite sure i'm in the right timezone. :P [22:01:43] cet is +1 [22:02:13] Yep [22:02:39] MatmaRex: right. but you (or Jan_eissfeldt) may have clocks that are a little fast or a little slow. like 7 mins worth [22:03:18] Never argue with mr ntp [22:03:50] Damianz: i do sometimes. but now only with jan's watch [22:04:30] jeremyb: ah, yes. my clock is automatically synchronised every full hour. [22:04:32] my evil watch has a balance wheel :/ [22:04:37] anyway, yes, we are 8 hrs behind UTC (or 9 hrs behind CET) in some parts. but some are only 5 (or 6) hrs off [22:04:37] jeremyb: it was synchronised four minutes ago [22:04:48] unless it managed to get out of sync in this time, which is not too unlikely [22:04:59] Jan_eissfeldt: i'm not familiar with this mechanism [22:05:02] IIRC us is betwean 5 and 9 hours depending where... damn timezones [22:05:03] MatmaRex: odd configuration [22:05:09] Cali is 8 though :D [22:05:36] Damianz: errr, you're counting hawaii? [22:05:44] jeremyb: that's because otherwise it would be about three minutes per hour too fast [22:05:45] where is this 9 you speak of? [22:05:50] jeremyb: there's something wrong with my laptop. [22:05:55] oh, hah [22:06:15] jeremyb: a bit more sophisticated than http://en.wikipedia.org/wiki/File:Pocket_Watch_Balance_Wheel.jpg [22:06:24] *click* [22:06:51] so, since i'm on windows, i decided to cook up a wicked command line thing that synces it all the time and stuff it in the scheduler [22:07:18] doing this, i discoered that the scheduler is as braindead as the command line [22:07:51] Krenair: space before protocol please!!! [22:08:14] what [22:08:23] > I think you wanthttps://www.mediawiki.org/wiki/API:Upload#Uploading_from_URL [22:08:32] jeremyb: Hawaii is totally us :D [22:08:37] please insert space before /http/ [22:08:42] jeremyb, there is a space.... [22:08:48] Damianz: ar eyou counting it? or just CONUS? [22:08:55] Krenair: where? [22:09:09] Between 'want' and 'https' [22:09:22] http://lists.wikimedia.org/pipermail/wikitech-l/2012-December/064957.html [22:09:24] CONUS excludes Alaska and Hawaii so it's no fun [22:09:34] Krenair: i think not [22:09:42] ... what. [22:09:48] But in my sent folder it has a space [22:09:56] * jeremyb chuckles [22:12:27] https://dl.dropbox.com/u/10971457/Wikimedia/Space%20between%20want%20and%20https.png [22:13:10] blame mailman [22:13:12] Krenair: ctrl-u ? [22:13:15] also shoot Krenair for top posting [22:13:16] :D [22:14:04] Posting underneath a reply that was at the top looks silly [22:14:48] * Damianz would delete everything under their reply since it's a mailing list so keeping full thread is just spammy [22:15:08] jeremyb, what does that do? [22:15:23] Krenair: message source? raw mime stuff? [22:16:03] Ugh wtf [22:16:08] Showing it as plain text shows there is no space [22:16:30] hah! [22:16:34] lol [22:16:39] Stupid Thunderbird [23:21:32] What's the correct forum for JPG orientation problems, and EXIF data display problems? Is here ok?