[00:50:32] !log [00:50:35] :) [00:50:37] ? [00:50:46] drdee, lol, what? ^ [00:51:08] not sure if it works in our channel [00:51:16] (yet) [00:51:16] oh ok [00:51:21] what's it do? [00:51:41] so it writes your log statement to a wiki page [00:51:54] so there is a public log if things that 'changed' [00:51:57] ooh [00:52:13] but isn't that the same as editing the page? [00:52:41] so if erosen asks when did you guys deploy filter X then actually he can look it up [00:52:57] yes but with the extra benefit that everybody in the channel immediately sees what you are doing [00:53:05] oh [00:53:11] and it's faster than editing a wiki page [00:53:19] my personal opinion on that type of information is good a git branching strategy [00:53:24] *a good [00:53:34] look in #wikimedia-operations for examples [00:53:53] is this for the zero-filters/ [00:53:58] or limn [00:54:00] he's saying in general [00:54:04] gotcha [00:54:05] to publish major events [00:54:10] with !log [00:54:19] I feel like it's duplicate work atm [00:54:23] but I can be persuaded otherwise [00:55:08] (duplicate because this type of information is synonymous with a git repository's tags on the master branch really) [00:55:31] so I'd be all for parsing that and making it publicly available [00:55:33] on a wiki [00:56:18] it's a standard practice used by all other teams except us [00:56:51] but pushing code is not synonymous with deployments or taking a machine off line [00:57:05] or upgrade a dependency [01:02:58] i see, taking machines offline is another thing. Though that should probably be mentioned on the lists as people are more likely to see that than look at our wiki or be in IRC [01:03:20] however, deployment should be synonymous with pushing code if you're doing it right. [01:03:23] CI and all that [01:03:39] but it does not cover 100% of the cases ;) [01:07:15] in theory you might be right but we are not there yet and it's easier to go through a public log then to have to grep through git commit messages [01:07:59] it should cover everything software related. For example, if some software that we depend on has to update, we should keep track of the new version in at least a README file, so that would warrant a commit. Hardware related stuff would be the only exception [01:08:06] yea, it's true that this is theory [01:08:20] but fake it 'till you make it. And I agree that we can do !log in the meantime [01:09:05] just important to keep the ideal solutions in mind [01:22:39] milimetric: this is the output: http://wikitech.wikimedia.7val.com/view/Server_admin_log [01:22:54] (which is actually not a wiki) [01:40:17] cool, let's get !log working here then [13:25:57] dschoon: busy? [13:26:12] very :) [13:26:17] metrics meeting in 4 hours [13:26:21] must jam! [13:26:22] sorry! [13:26:37] i will ping you after, if you are around? [13:27:26] sure, no problem. i think it should be a simple problem, i'll be around [15:35:39] Morning!!!!!! [15:35:48] Morning [15:38:17] Ping average_drifter [16:34:26] Ping ottomata [16:34:56] did Asher review the blog.wikimedia.org udp2log stuff? [16:45:53] dschoon: do you mind if I make an ssh identity file for the www user on kripke? [16:46:02] erosen: go for it [16:46:08] coo; [16:46:19] that way I can add handmade graphs to the repo [16:57:56] ping average_drifter [16:58:34] here [16:58:41] good day ! :) [16:58:57] so please tell me we can deploy the editors stuff for wikistats :) [16:59:55] can't say that yet [17:00:23] but I'm working on it [17:02:04] what has the progress been since yesterday? [17:02:43] still looking on 2) of Erik's list [17:04:17] what was 2) again? [17:04:34] "new report shows strange domains, like 2620-, 2001-, 2a01" [17:04:36] the swap of variables or the ip6 stuff? [17:04:37] right [17:05:05] it sounds to me an odd number of fields [17:05:20] that causes the ip6 address to be assigned to the url field [17:06:15] can I have a vhost please on stat1 that exposes one of my HOME directories ? [17:06:23] I need to generate stuff on there and see the result in my browser [17:06:43] I don't have root rights so I can't make the vhost myself [17:06:49] ottomata: hey Andrew [17:07:44] ottomata: I'm working on fixing a bug and I need to make wikistats reports along the way so I need to expose them from stat1. can I have like a vhost on stat1 that exposes my /home/spetrea/public ? [17:11:04] yesssuh! [17:11:11] :) [17:12:27] so i've done this for a couple of people already, mind if we call it /home/spetrea/www ? [17:12:31] just to keep this consistent? [17:12:36] no problem [17:12:39] www is good [17:14:42] there you go [17:14:42] http://stat1.wikimedia.org/spetrea/ [17:14:52] awesome [17:14:53] thanks [17:15:03] now I can test stuff easier [17:48:50] ottomata: want to check in about metrics meeting? [17:55:00] ja! [17:56:42] https://plus.google.com/hangouts/_/2e8127ccf7baae1df74153f25553c443bd351e90 [17:56:44] https://plus.google.com/hangouts/_/2e8127ccf7baae1df74153f25553c443bd351e90 [17:57:26] erosen^ [17:57:57] hey [18:05:49] oo, drdee [18:05:57] yooooo [18:06:14] can I say something about our desire to get regular edit data imports in as well [18:06:15] ? [18:07:23] totally, and say that we already are storing unsampled data for mobile and english (with the usual caveats) [18:08:54] what are the usual caveats? [18:12:33] milimetric, dschoon: quick web dev question [18:12:42] shoot [18:12:51] any way to make a file in global-dev.wmflabs.org/data/../..//.. [18:12:58] display as a page rather than be downloaded [18:13:06] should I just wrap it in html? [18:13:14] what kind of file [18:13:18] a tsv [18:13:34] you gotta mess with the response headers [18:13:43] interesting [18:13:47] yeah, it's easy [18:13:52] come downstairs [18:13:53] i'll show you [18:13:55] k [18:13:57] :) yea that'll be faster [18:14:00] hehe [18:14:02] thanks [18:14:03] we just need to change a thing in your server-wrapper [18:14:07] hmm [18:14:17] i'm actually not sure it is worth changing the server at this late date [18:14:23] sure [18:14:39] i'll come down anyway [18:14:42] you just need to add a ContentType header [18:14:46] text/plain [18:15:59] yeah, without that the browser handles it however it sees fit based on its system's file extensions [18:16:06] but text/plain forces it to render as text [18:21:16] you could probably just change the file ext. to .txt [18:24:30] i think it'd still download it [18:35:23] set up the server to set the mime type for .txt and .csv files [18:35:53] now you can see the perversion of the tunisia mobile users: http://global-dev.wmflabs.org/data/topk/orange_tunisia/x_counts/part-r-00000.txt [18:36:15] milimetric: another q if you have a sec [18:36:23] shoot [18:36:25] any ideas on how to set encoding? [18:36:41] charater-encoding on the html itself? [18:36:45] i think this file is utf-8 [18:36:52] or for serving that specific file [18:37:00] for that specific file [18:37:20] basically I want that link to display text in the browser as utf-8 [18:37:56] right, response has to have a Content-Encoding header set to utf-8 I think [18:38:01] gotcha [18:38:44] wait, you can't do that now right? [18:39:06] for reference: http://en.wikipedia.org/wiki/List_of_HTTP_header_fields [18:39:11] cool [18:40:39] heh, erosen that file's got a lot of porn and sex in it :) [18:40:47] yeah [18:40:49] i might not actually display it [18:40:56] i thought it was sort of funny [18:40:58] but on second though [18:40:59] it is [18:40:59] ... [18:41:09] oh whatever [18:41:15] yeah [18:41:16] no information is bad [18:41:21] hehe [18:41:30] exactly, it's the truth [18:41:38] wait so does that help? do you have control/desire to set character-encoding? [18:41:47] if not I can try to look up hacks [18:42:23] yeah i do [18:42:40] this is what david added: [18:42:41] app.use (req, res, next) -> [18:42:42] if /\.(csv|txt)$/.test req.url [18:42:43] res.header('Content-Type', 'text/plain') [18:42:44] next() [18:42:54] so I could probably just add another res.header call [18:43:21] yes, you can also try this in content-type: [18:43:21] text/html;charset=utf-8 [18:43:24] uh [18:43:31] text/plain;charset=utf-8 [18:43:36] cool [18:44:28] if not, res.header('Content-Encoding','utf-8') definitely [18:44:30] niiice [18:44:31] worked [18:44:37] the first version [18:44:43] http://global-dev.wmflabs.org/data/topk/orange_tunisia/x_counts/part-r-00000.txt [18:44:44] cool [18:44:46] yea [18:44:46] saw [18:45:29] Accueil means like home page btw [18:45:34] so they get into it right after that [18:45:37] :) [18:47:23] yup [18:56:38] metrics meeting: http://youtu.be/wB5m5AHoGno [18:56:45] A bunch of us are now in #wikimedia-office to talk about https://meta.wikimedia.org/wiki/Metrics_and_activities_meetings/2012-12-06 - come join and watch the live video stream :D [19:10:54] meeting will be available for playback afterwards right ? [19:10:57] I wanna watch it too [19:11:36] average_drifter: http://youtu.be/wB5m5AHoGno [19:11:52] not sure if there's a playback option [19:26:05] you can watch it after it's over [19:26:23] (or just watch what you missed) average_drifter ^ [19:40:11] nice job ottomata, erosen [19:40:19] +1 [19:53:00] thanks [19:58:47] thanks [20:05:04] drdee: just tried to grep, no entries with "blog.m" [20:05:43] can you directly listen to the udp port? [20:05:47] and grep that? [20:06:08] so the raw squid logs ? [20:06:37] got no access to the raw squid logs, I only have access to the logs piped through udp-filter [20:06:43] the stream actually.. [20:09:08] ottomata1: can/may I have access to raw squid logs ? [20:09:20] ottomata1: raw squid stream actually [20:09:41] is it not coming in on stat1 anymore? [20:10:00] oh thath is the same thing [20:10:09] ohhhhh [20:10:11] sorry, i understand [20:10:19] yeahhh, you want me to pipe that over too, eh? [20:10:41] if it's possible yes please [20:11:10] I need to grep for blog.wikimedia.org to see if it's in the raw stream [20:14:33] ok [20:14:33] from stat1 [20:14:34] sudo netcat -lu stat1.wikimedia.org 8520 [20:14:48] or you don't need sudo [20:15:13] ottomata1: has https://gerrit.wikimedia.org/r/#/c/37037/ been deployed ? [20:16:28] i don't know for sure, but I assume if asher merged it he deployed it [20:16:36] acutally, i think i can check... [20:17:44] yes please [20:21:03] i don't think its there yet, i'd rather let asher make it happen [20:21:05] will poke him [20:21:22] ok [20:21:57] cool [20:30:56] brb 20m, getting stuff [20:33:22] :) Erik M ftw [20:36:56] ja, i will be back in a min too [20:38:15] me too [20:38:26] ditto, that was long :) [20:40:18] suggestion that came to mind after endlessly hearing people talking about how they never know what's going on. [20:40:18] I think the definition of "knowing what's going on" might be biased these days by how we get information from facebook. So in that spirit, I propose a http://timeline.verite.co/ graph that all the teams can send information to with !log from IRC. This could be prominently displayed on a homepage somewhere and would provide an official record of "what's going on" [21:15:35] averagedrifter, I see what is wrong with puppet, trying to get a hold of asher to check it [21:19:50] ottomata: ok [21:22:38] * drdee keeps fingers crossed [21:28:06] Is there anyway groks stats can be fixed so as not to default to Wikipedia when navigating by month? [21:33:36] milimetric: +1 ^^ [21:34:23] purplepopple: what are you suggesting? [21:36:20] drdee: Um. Let me get an example. [21:37:05] http://stats.grok.se/en.n/201212/IPC_criticizes_US_Paralympic_coverage [21:37:25] If I change the month on the bottom to say use September... [21:37:35] http://stats.grok.se/en/201209/IPC_criticizes_US_Paralympic_coverage [21:37:44] The navigation bar gives me that. [21:37:46] Wrong project. [21:40:55] can you code in python? you might be able to supply the patch yourself ;) [21:41:26] Nope. [21:41:36] I'm pretty incompetent at coding. [21:41:37] I can write some python [21:41:40] :D [21:41:54] purplepopple: you already know python, you just don't know you know it [21:41:59] syntax is almost english [21:42:46] average_drifter, drdee: we need this reviewed before blog stuff happens [21:42:46] https://gerrit.wikimedia.org/r/#/c/37295/ [21:42:55] reading [21:43:22] ask notpeter / binasher ? [21:43:31] average_drifter: I'm sitting in a ski resort in Colorado. :) Not probably the time. :) [21:43:38] I can make minor changes to perl scripts but that is abotu it. [21:43:56] python is easier than perl :) [21:58:41] ottomata, regarding asana: mmmmmmm weird [21:59:09] drdee: any ideas on how to let ryan faulkner commit to my global dev dashboard [21:59:20] i think i need to add him to the analytics members [21:59:21] on gerrit [21:59:23] sure, [21:59:27] i'll do taht [21:59:29] hold on [21:59:31] at the moment I don't seem to have permission [21:59:34] thanks [22:00:07] does he have a gerrit account? [22:00:32] because his name does not show up in the auto suggest list [22:00:35] eah [22:00:43] rfaulkner [22:00:48] he is standing right next to me [22:00:52] got it done [22:00:59] try again [22:01:02] Rfaulk apparently [22:01:03] cool [22:01:12] sweet [22:01:14] it's there [22:01:18] now let's hope that fixes our problme [22:03:12] i think it will [22:04:26] robla and i are joining momentarily [22:04:33] he is performing AV setup [22:05:19] we got a hangout coming i guess? [22:06:28] https://plus.google.com/hangouts/_/07ce7bab94f46bb5555f8a1592b35e125e2c08b5 [22:06:57] ottomata ^^ [23:03:13] any of you boys ever played agricola? [23:03:16] i'm about to (re)learn how [23:15:20] aiight. i've been up since 4a today [23:15:20] i'm heading home :) [23:15:26] Alchimista: you still have a question? [23:17:15] ottomata, blgo traffic is streaming? [23:18:50] okay, heading out [23:18:52] ta~ [23:18:58] see y'all on tues! [23:19:01] laterz dschoon [23:19:07] i'll be in the office [23:19:11] (then) [23:19:18] yay! [23:24:13] ottomata: I've looked at the changes in the link https://gerrit.wikimedia.org/r/#/c/37037/1/manifests/misc/blogs.pp and the other file [23:24:40] ottomata: but I know little about Ruby or puppet so I can't review [23:25:23] oh no he just left you hanging man [23:25:34] everyone's tired average_drifter, we've had a long meeting-filled day [23:26:06] no he has a wacky internet connection [23:26:11] oh really? [23:26:24] although he might be off for dinner [23:26:25] never mind, *I'm* tired [23:26:28] nite everyone [23:26:30] :) [23:26:31] laterz! [23:26:56] average_drifter: AFAICT, https://gerrit.wikimedia.org/r/#/c/37295/ has been merged [23:27:01] so that might have fixed the problem [23:27:12] can you check one more time if you see blog traffic on udp2log? [23:32:06] drdee: checked, no blog output on url field [23:40:41] average_drifter, yeah, i pushed it through, but now there is another problem :( [23:40:46] i will fix it tomorrow, i gots to run [23:41:57] ottomata, what's the problem?