[00:06:45] ottomata, updated khadoop [00:06:49] cool, danke [03:03:29] drdee: I still have problems with many titles [03:03:31] drdee: you there ? [03:03:37] yo [03:03:39] drdee: I mean, some of them just don't have titles [03:03:42] which domains? [03:03:46] let me give some examples [03:03:58] http://nl.wikipedia.org/w/api.php?format=json&action=opensearch&search=bu&namespace=0&suggest= [03:04:24] ok so -o should output 5 fields [03:04:28] the last field is the title [03:04:34] what should I output if it doesn't have a title ? [03:04:56] but that url does not have /wiki/ [03:05:10] so you can skip it [03:05:26] alright [03:05:27] we only count urls that contain either /wiki/ or index.php= [04:02:19] drdee: within ~40 of the same output as filter [04:02:44] in total? [04:04:23] drdee: yea [04:04:53] and those 40, can you share those on a gist? [04:05:21] yes [04:05:25] but first, I need to mention [04:05:29] this is a comparison between [04:05:38] ./udp-filter -o -t and filter [04:05:51] that's fine [04:07:07] I'll gist these up [14:51:36] good morning! [14:52:43] mornign! [14:52:50] drdee is gonna have a babyyyyy! [14:53:02] hey, what did the mommy bullet say to the daddy bullet? [14:53:40] dunno [14:55:36] "We're gonna have a BB!" [14:58:13] wait it's official? [14:58:32] no, can't be [14:58:35] he wouldn't be on IRC! [15:03:48] no laughter? [15:04:02] cmoooonnnnnn good joake! [15:04:57] :D [15:05:17] i am tapering off soon [15:05:23] busting out some final emails [15:05:32] try to join scrum bot no promises [15:09:11] lol ottomata, BB [15:09:12] :) [15:09:50] hehehe [15:43:11] average_drifter:around? [15:53:57] q for him too, and maybe you drdee [15:54:04] i've got the change to udp-fitler for field delimiter [15:54:12] should I jsut commit it and let you guys deal with changelog and debianizing? [15:57:09] yeah just push it [16:06:29] ok pushed [16:06:34] oops, i shoulda git-reviewed it [16:06:37] i just pusheed [16:07:28] https://gerrit.wikimedia.org/r/gitweb?p=analytics/udp-filters.git;a=commitdiff;h=67dcf53f0ee4214e24f75528ecee6ba37e0e58ca [17:59:13] https://plus.google.com/hangouts/_/2e8127ccf7baae1df74153f25553c443bd351e90 [18:01:00] ottomata, erosen, average_drifter ^^ [18:28:15] Analyst syncup is https://plus.google.com/hangouts/_/33bb08418c094ead3577050db709b808c2007e86 [18:28:57] in case anyone (like, say, erosen) would like to alleviate the crushing loneliness [18:31:37] yeah [18:32:22] or not! [18:32:31] oh, i forgot to hit "join" [19:06:07] ok, actually going to post office now [19:07:21] oops just kidding [19:07:28] have to wait 30 minutes til pie is done [19:37:24] hey [19:37:33] I fell alseep and forgot to review [19:37:35] doing now [19:39:28] i committed the —field-delimiter change to udp-filter [19:39:32] you should pull and check that one out [19:43:17] afk for a few mins, going to post office (for real this time) [19:47:40] ottomata, i did a bit more reading and updated http://www.mediawiki.org/wiki/Analytics/Kraken/JMX_Monitoring [19:48:06] basically i think we should just go with graphite or ganglia. graphite seems shinier, mostly for the query API [20:08:51] ottomata: hello [20:08:53] ottomata: https://gist.github.com/7fa3c3eb24ceca675ffa [20:09:28] drdee: https://gerrit.wikimedia.org/r/#/c/33403/ [20:09:34] drdee: https://gerrit.wikimedia.org/r/#/c/33407/ [20:09:55] ottomata: fixed bugs, please have a look at the output [20:11:01] ok, these are the outputs [20:11:10] http://garage-coding.com/filters_output_fixing_bugs.zip [20:11:27] ottomata , drdee the input file was the one we used yesterday [20:21:33] average_drifter, drdee's having a baby [20:21:50] so he probably won't get to review [20:24:37] :) [20:43:15] ok soo waaa? [20:44:07] average_drifter, that's cool, i'm not going to get into reviewing any of the filter/collector webstats stuff [20:44:11] as you know way more about it than I do [20:44:23] should I just merge those? [20:45:14] ottomata: you can yes, tests are passing [20:45:20] cool [20:45:46] jenkinsbot doesn't like this one though [20:45:47] https://gerrit.wikimedia.org/r/#/c/33407/ [20:46:15] it doesn't like it because our ticket to operations in gerrit is not finished [20:46:25] ottomata: can you help me make the final changes for the ops please ? [20:46:47] ottomata: I don't know in what role I should put the package installations for openssl [20:46:51] actually libssl-dev [20:46:56] well, both [20:47:28] oh [20:47:28] hm [20:47:34] if we do that, then we'll have libanon built, libcidr files and then we can do continous integration for udp-filters as well [20:47:38] yeah, what's the link to that change again? [20:47:48] and jenkins bot will not complain anymore [20:47:50] moment [20:47:53] lookin for it [20:49:03] ottomata: https://gerrit.wikimedia.org/r/#/c/32192/ [20:51:12] ok, mind if I just create a new commit to do this? [20:52:49] ottomata: no problem [20:56:07] um so [20:56:11] are you sure this is the problem? [20:56:19] openssl and libssl-dev are both currently installed on gallium [20:56:45] average_drifter^ [20:57:15] are they ? [20:57:16] hmm [20:57:22] let me check another libanon build [20:57:56] 20:57:44 ./configure: line 11393: syntax error near unexpected token `OPENSSL,' [20:57:59] 20:57:44 ./configure: line 11393: `PKG_CHECK_MODULES(OPENSSL, openssl)' [20:58:02] https://integration.mediawiki.org/ci/job/libanon/20/console [20:58:19] I don't know why that happens [20:59:20] that says syntax error though, right? [20:59:27] hmmm [21:00:15] well libanon builds without problems on my machine [21:00:26] and I don't know why that's happening there [21:00:45] not having access to gallium, I can't figure out the problem [21:03:53] can we trigger a build now? [21:04:17] want to try it [21:04:22] average_drifter^ [21:04:30] ottomata: yea [21:04:38] how do we do that? [21:04:43] https://integration.mediawiki.org/ci/job/libanon/ [21:04:46] just hit build now [21:06:10] yeah? [21:06:11] can I see it? [21:06:47] the result ? [21:06:58] yeah, don't see the build running [21:07:05] did you press "build now ? [21:07:21] oh ha, thought you did... [21:07:22] uhhh [21:07:28] dno't see that link [21:08:15] ah, got it [21:08:16] had to log in [21:08:18] :) [21:12:11] https://gerrit.wikimedia.org/r/#/c/33466/ [21:12:13] :) [21:13:19] looking [21:13:43] sweeet :) looks good [21:16:19] that looks good, so pkg-config was missing [21:16:24] and pcpa [21:16:26] pcap [21:16:33] yup [21:16:57] so, both of your udp-filter changes are merged now [21:17:51] thanks ! [21:18:00] ottomata: how can I make libanon be part of analytics on jenkins ? [21:18:04] ottomata: currently they're not [21:18:06] ottomata: https://integration.mediawiki.org/ci/job/libanon/ [21:20:30] hmm, i have never used jenkins before! [21:21:16] so i unnooooo [21:38:13] ottomata: buildbot ? [21:38:36] buildbot? [21:38:49] i haven't used CI before, honestly [21:38:55] oh ok :) [21:57:31] hey ottomata, I was able to rewrite my own github kraken repo's history using the command dschoon gave and a force push, but I don't have a push access to the wmf-analytics repo. [21:57:47] *nod* [21:58:03] github is git in most respects, so all the tools basically work. [21:58:29] the thing is, `git push --force` will cause a huge mess for everybody who pulls afterward [21:58:46] i'm ok with recloning kraken [21:58:49] yeah. [21:59:06] i recommend that FIRST, everybody try cloning louisdang's fork [21:59:12] hmm ok [21:59:15] make sure that works, still builds/whatever [21:59:18] do it clean [21:59:57] cloned no prob [22:01:01] once everybody signs off, then ottomata you: [22:01:15] git remote add wmfa git@github.com:wmf-analytics/kraken.git to your fresh clone [22:01:49] and finally, after triple-checking that everyone knows: `git push --force wmfa --all` [22:01:51] i believe [22:02:02] you may want to test this with a dummy repo first :) [22:02:04] brb meeting [22:03:51] ottomata, I haven't merged your latest commit yet [22:05:22] ok, i'm not doing anything atm [22:57:56] ottomata: what was that library you made for processing generic csv? [22:57:59] files [23:11:54] bwerrrrrrr [23:11:58] pipeline [23:12:34] https://gerrit.wikimedia.org/r/#/admin/projects/analytics/reportcard/old-pipeline [23:30:13] danke