[00:05:03] milimetric: do you have shell access to stat1001? [00:05:15] i do not [00:05:24] oh, oh well [00:06:54] milimetric: yes you do :P [00:07:07] oh duh [00:07:11] i thought you said sudo [00:07:15] yes, i have shell [00:07:18] but not sudo [00:07:36] could you possibly copy the apache error log to stat1? [00:08:24] /var/log/apache2/error.log, i mean. you may need sudo, actually. hrm. [00:09:58] yep, needs sudo [00:10:27] bummer [00:10:29] for user metrics, i have to ask andrew to fetch it for me [00:10:30] thanks anyway [00:10:39] do you have a previous log? [00:10:43] sorry i'm powerless [00:10:52] no, just snippets relevant to user metrics [00:11:06] can you share those with me? [00:11:25] they're kinda random throughout my chat history, are you debugging umapi? [00:11:58] DarTar mentioned he was getting occasional 500s and I wanted to see if it was related to something I had been poking at [00:12:33] oh, all the snippets i have i've fixed myself [00:13:00] ah. are there no more errors, then? [00:13:13] there were two 500s - one was because we hadn't configured the tunnel to s6 and one was because it wasn't using "something".decode('utf8') properly [00:13:26] no others that i know of [00:13:45] oh, okay, not related then [00:13:49] thanks for that [00:13:57] but yea, that's a problem - we need to be able to see logs [00:14:46] yeah, we can specify in puppet that umapi logs should go somewhere world-readable [00:15:08] i'll ask andrew tomorrow [00:15:17] otto, that is [00:15:27] two andrew opsen in the channel :) [00:15:28] he's actually on vacation all week [00:15:41] oh, right [00:15:52] umapi logs should - but i think there's some config problem where they're still going to error.log [00:16:06] like error.metrics.log should be world readable [00:16:21] oh, do you have that? [00:23:01] average: Where is the native code for dclasswrapper ? [00:31:07] xyzram: moment [00:31:54] xyzram: https://github.com/wikimedia/dClass/blob/package/jni/dclass-wrapper.c [00:32:17] xyzram, need help? [00:32:58] in a hangout with millimetric [00:33:26] should i join? [00:34:50] https://plus.google.com/hangouts/_/2da993a9acec7936399e9d78d13bf7ec0c0afdbc [01:09:32] average: thanks. Looks there are 2 branches: master and package; which is the one we use ? [01:10:01] 1 sec [01:10:41] in https://github.com/wikimedia/kraken ? [01:11:03] https://github.com/wikimedia/dClass [01:11:11] ohhhhh [01:11:17] why are you looking into that? [01:11:26] I was asking about where the native code is for that class [01:11:39] because I saw it used in UserAgent I think. [01:12:21] you don't need that for this particular use case [01:12:44] that's used to recognize browsers based on useruagent strings [01:12:44] true, that was before our hangout session :-) [01:13:05] ok [01:15:50] xyzram: do you have a mingle account? [01:16:11] don't think so [01:19:25] ok [01:20:05] drdee: the perl regex starts "^(.*) CFNetwork ..." with a literal space before the 'C' but the user agent strings have a dot in that position. [01:20:20] They cannot match [01:20:30] feel free to improve ;) [01:22:21] I can fix it of course but I'm wondering if I'm starting with a bad model; I'll just ignore that model and write my own based on the samples. [01:22:33] ok [10:59:35] morning! [11:35:29] milimetric: morning ! :) [11:35:57] milimetric: very early bird [11:36:03] milimetric: how so ? [14:06:06] drdee: hi [14:06:30] I'll be out for 1h to get some stuff [14:06:57] drdee: what needs be done by the end of wednesday ? please let me know [14:17:21] yo [14:30:26] milimetric, average: morning [14:31:00] hi [14:31:05] i think average stepped out for a bit [14:31:09] kool [14:31:13] but GOOOOOOOD MRONING! [14:31:31] it's such a great weather here [14:31:41] went for my morning stroll [14:31:49] happy to hear. it's my favorite weather here but that means rain :) [14:32:27] :) [16:59:47] scrum guys [16:59:54] erosen; average; milimetric [17:00:15] * average is running towards laptop [17:01:11] hangout issues, one sec [17:31:56] average grooming [17:31:59] https://mingle.corp.wikimedia.org/projects/analytics/cards?favorite_id=758&view=%3EWIP+-+Feature+Analysis [18:28:38] New review: JGonera; "(1 comment)" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/60608 [18:28:46] New patchset: JGonera; "Generate JSON files in datasources using templates" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/60608 [18:29:00] Change merged: JGonera; [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/60608 [19:02:37] ori-l: https://plus.google.com/hangouts/_/2da993a9acec7936399e9d78d13bf7ec0c0afdbc [19:03:11] milimetric: give me 5 minutes, running a bit behind [19:03:14] sorry [19:03:53] no prob, evan and i await in pre-design euphoria :) [19:30:34] New patchset: JGonera; "Add debug argument (-d or --debug-folder)" [analytics/limn-mobile-data] (master) - https://gerrit.wikimedia.org/r/62664 [19:43:11] average: how is the progress with debianizing the latest version of udp-filters? [19:43:16] xyzram: need any help? [19:43:21] brb [19:47:27] drdee: the progress is good, but it's not finished [19:47:36] I will hover back the chan with an update in 30m [19:47:43] ok [20:22:30] drdee: have all changes of Ram been merged into master ? [20:22:53] xyzram: do you have any unmerged patchsets ? [20:22:57] xyzram: for udp-filters ? [20:28:49] ok, considering that yesterday we merged a couple of patchsets [20:28:58] I assume that now we have everything we need on master [20:39:27] average: that is correct [21:04:06] drdee: not yet, I'm poking around ... [21:05:28] cool, just let me know & don't spend too much time poking :) as you might awaken the kraken and be swallowed by it :D [21:05:57] That beast is dangerous, I take it :-) [21:16:37] drdee: I cloned the kraken repo to: https://github.com/xyzram/kraken.git [21:17:18] nice! [21:35:05] oh DarTar, I totally forgot about the meeting today [21:35:09] are you and Evan talking? [21:41:33] hey milimetric, Evan and I just had a long chat with adrian, I think we don't need to meet on UMAPI, I'll send you guys feedback on the latest fixes/code changes as requested by drdee - is that ok? [21:50:12] sure DarTar, that works for me [22:29:16] xyzram: you can now access the an boxes [22:29:27] however i need to get dinner first :( [22:29:35] but please add unit-tests and create new jars [22:29:43] then we can start looking at testing this [22:30:05] ah, thanks, will do. [22:30:46] You want a single PageviewType for all five apps we discussed yesterday ? [22:31:36] currently I have just a single regex so a single type makes sense. [22:31:52] but I can split them out if needed. [22:54:39] drdee , xyzram udp-filter 0.3.23 release is available here http://bit.ly/18UOU0K [22:56:05] these are the contents of the package http://bit.ly/10nabkG [22:56:13] it includes multiplexor and udp-filter [22:58:20] if there's anything else please let me know [23:08:57] average: ok, thanks. [23:16:25] drdee: amit wants to try out the maxmind ISP database as a replacement for manually curated cidr ranges [23:16:29] thoughts? [23:16:59] drdee: it costs $25 so at the very least it seems worth trying [23:21:11] I'm off for the day [23:21:19] later [23:24:44] laater [23:38:10] erosen: sounds good [23:40:02] xyzram: you don't need to worry about the PageviewType for now [23:42:28] That's what the unit tests seem to be using. [23:42:49] ok, so you can just use that