[01:09:34] [travis-ci] develop/92c0fbf (#109 by milimetric): The build passed. http://travis-ci.org/wikimedia/limn/builds/5331168 [02:03:03] eros [02:03:08] *erosen [02:03:16] hmm [02:03:22] drdee: hey [03:29:36] dschoon: hey David [03:29:45] uhm, it's late [03:30:04] I was wondering if I could catch Evan for a few minutes [03:30:37] i just got home [03:30:38] sorry :( [13:53:23] hi ottomata [13:59:19] hi! [13:59:20] morning [13:59:23] average_drifter [15:08:10] hey erosen [15:08:19] average_drifter: sup? [15:08:21] sent an e-mail :) [15:08:24] ya [15:08:26] hope you got it [15:08:30] did indeed [15:08:36] cool [15:08:44] sorry I didn't respond lasst nightl wasn't on e-mail muh [15:08:51] regarding the cidr_ranges.json thing [15:09:06] it spits out an error but it is more appropriately a warning [15:09:36] it should still run, it just checks whether that files exists and if it doesn't it means you can't do ip lookups [15:11:00] average_drifter: regarding details to be added to the diagram [15:11:23] I have been trying to figure out the best way to diagram the way I parse out the project [15:12:10] erosen: nice. if I can help you with anything graphviz-related I will try to [15:12:19] i'm good on that part [15:12:24] I just don't know how detailed I want to get [15:12:31] but I think I can do it relatively succintly [15:12:34] I'm making a new one now [15:12:38] i'll ping you when it's ready [15:12:41] cool [15:12:44] thank you [15:12:48] np [15:16:56] [travis-ci] develop/d09910c (#110 by dsc): The build passed. http://travis-ci.org/wikimedia/limn/builds/5341339 [15:32:47] average_drifter: https://github.com/wikimedia/metrics/tree/master/pageviews/embr_py [15:32:56] i've updated the diagram to show the parsing on domain that I do [15:33:19] i'm considering adding a node type which represents processing, rather than a decision step [15:35:35] feel free to :) [15:35:41] reading the diagram now [15:40:53] erosen: consider writing a node with regexes(or something that expresses intuitively what class of urls you are matching) [15:41:00] erosen: it would reduce the number of nodes [17:34:35] average_drifter: did the diagram clarify anything? I think I might have already mentioned it [17:37:42] 17:40 < average_drifter> erosen: consider writing a node with regexes(or something that expresses intuitively what class of urls you are matching) [17:37:45] 17:41 -!- erosen [~erosen@wikimedia/Evan-WMF] has quit [Read error: Connection reset by peer] [17:37:48] 17:41 < average_drifter> erosen: it would reduce the number of nodes [17:37:51] erosen: ^^ [17:38:01] erosen: I have added www.[projectname].m.wikipedia.org [17:38:04] erosen: to my code [17:38:20] erosen: I didn't get more accepted lines :| [17:38:30] hmm [17:38:48] u [17:38:52] morning [17:38:56] drdee: morning :) [17:39:06] yeah, I had held off on putting it into the graph as nodes because it seemed a little verbose [17:39:12] but I was worried that a regex would be misleading [17:39:31] erosen: doesn't have to be regex in the rigorous sense of the word [17:39:33] and i'm not sure it is possible to exactluy reproduce the functionality [17:41:21] erosen: the path in your graph yes,no,no,no,yes [17:41:26] erosen: is that correct ? [17:42:20] average_drifter what do you mean? [17:42:47] https://raw.github.com/embr/metrics/master/pageviews/embr_py/Pageview_definition.png [17:43:12] erosen: start from your first decision node, and follow this path yes,no,no,no,yes [17:43:29] this implies you don't have a language id [17:43:36] yeah [17:43:42] in that case it will still register as a page view [17:43:42] but you are still counting it, but what language id do you count it to? [17:43:57] i just say it is NA [17:44:01] or None in my case [17:44:05] average_drifter: && [17:44:06] ^^ [17:45:40] hallo [17:45:49] erosen: you had a qq yesterday? [17:45:54] ok [17:46:10] dschoon: ja, i am trying to set up limn-deploy on my machine [17:46:46] dschoon: and I keep getting an ssh error because it is trying to use my local username evanrosen rather than my username on the labs [17:46:58] I was curious if you know anything about the ssh internals of fabric [17:49:14] ah yes [17:49:17] absolutely [17:49:30] there's a big fat monkeypatch to make it work with a bastion gateway [17:49:32] erosen: ^^ [17:49:39] i noticed that file [17:49:47] drdee, are we still on for the demo this morning? or could we move that to wednesday to be part of our weekly sprint showcase/reflect/planning? :) [17:49:54] i looked into it a bit and couldn't quite figure it out [17:50:01] dschoon: ^^ [17:50:16] kraigparkinson i'm fine with moving it, though i'd also like to show off the much-improved viewport handling [17:50:48] erosen: lemee look real fast [17:50:52] drdee, is that time sensitive? [17:50:58] dschoon: thanks [17:51:01] kraigparkison: i thought we had moved that to wednesday [17:51:12] erosen: i imagine you've done this? pip install -U Fabric paramiko path.py [17:51:17] yup [17:51:32] i'll reproduce the error and show you the console messages [17:51:38] so let's demo all our stuff on wednesday, my 2 cents [17:52:00] dschoon: [kripke.pmtpa.wmflabs] Executing task 'deploy' [17:52:00] [kripke.pmtpa.wmflabs] Making Target Directories for Data... [17:52:01] [kripke.pmtpa.wmflabs] Login password for 'evanrosen': [17:52:15] er, maybe gchat would be better for the Wall o Text [17:52:31] dschoon: that is all, so I figured IRC was okay [17:53:00] heh [17:53:09] right [17:53:12] so, your login is erosen, isn't it? [17:53:17] ya [17:53:20] but evanrosen locally [17:53:25] and it's attempting to connect as evanrosen [17:53:26] hence the problem [17:53:27] so! [17:54:01] drdee, I'm happy with all demo stuff wednesday. Just noticed that the demo invitation is still in my calendar for post-scrum... [17:54:11] today [17:54:22] i deleted it 5 minutes ago [17:54:48] erosen: run fab -h [17:54:49] :) [17:55:43] of interest is --user, --ssh-config-path, -i, -A [17:55:48] got it [17:55:50] also i think --gateway is new [17:55:52] almost worked [17:55:53] i'll add a readme [17:56:11] thanks for the tip, I didn't think to check the help doh! [17:56:23] rtfm n00b! [17:57:55] dschoon: on a related note, gp-dev.wmflabs now points to kripke.wmflabs.org (i.e. old documentation about limn) [17:58:04] yay [17:58:06] oh [17:58:07] wait. [17:58:08] boo! [17:58:11] hehe [17:58:15] i'll fix that. [17:58:53] dschoon: thanks [18:00:42] dschoon: gp works! thanks again [18:00:42] MEETING TIME! [18:00:49] good point [18:00:54] dschoon: on my way down [18:01:04] https://plus.google.com/hangouts/_/2da993a9acec7936399e9d78d13bf7ec0c0afdbc [18:01:36] https://issues.apache.org/jira/browse/PIG-3214 [18:02:28] kraigparkinson are you around? [18:02:35] coming! :) [18:09:03] dschoon: they have a blocker right there :) [18:28:36] lunchtime, be back in a little while [18:32:16] http://detectify.com [18:32:57] aw [18:33:07] the free scan isn't automated [18:35:02] milimetric: http://ternjs.net/ [18:35:46] milimetric: also, when you get settled, i fixed the issues with viewport zoom last night [18:36:14] we decided you'll act as the customer for that -- once you sign off, we can close the card [19:08:42] off to lunch, see you in a bit [19:18:59] I'm stuck like Chuck [19:22:03] [travis-ci] master/cd997e9 (#75 by Diederik van Liere): The build has errored. http://travis-ci.org/wikimedia/kraken/builds/5350003 [19:27:03] dschoon [19:27:05] i have a 1on1 in 30 [19:27:07] and am scarfing [19:27:13] ja? [19:27:13] want to hangout and talk about oozie and stuff? [19:27:25] dschoon: hi :-] So I was wondering how you managed to have a terminal with split windows each having different background colors :-] [19:27:25] yeah. [19:27:31] dschoon: forgot to ask about it last week :-] [19:27:33] hashar: iterm2.com [19:27:39] it is fucking fantastic. [19:27:43] seconded [19:27:44] ottomata: yeah [19:27:45] its amazing [19:27:49] iterm2 [19:28:01] lemee grab a room so i don't annoy people [19:28:05] ping you in 1sec [19:28:07] dschoon: yououuuu thanks! ;-] [19:28:19] k, dschoon, i'm here [19:28:20] https://plus.google.com/hangouts/_/2da993a9acec7936399e9d78d13bf7ec0c0afdbc [19:37:35] erosen: I can deploy to 'mobile' as well, right? [19:37:41] as in fab mobile deploy.only_data [19:37:41] ? [19:37:45] instead of just fab mobile_dev? [19:37:47] hat should work [19:37:58] ok [19:37:59] it all depends on the configuration [19:38:09] * YuviPanda tries [19:39:03] looking at stages.py, seems like it should work [19:40:08] yeah it did :) [19:40:14] erosen: does limn let us tag arbitrary points in time? [19:40:28] so that we could tag a date and say, 'this blog post on this day', 'this campaign started on this day', etc? [19:40:30] hmm do you mean like a graph annotation? [19:40:40] there is some support for that, but I haven't used it myself [19:40:45] yeah. [19:40:49] I think milimetric built that in ro DarTar [19:40:57] s/ro/for/ [19:49:06] YuviPanda, you are welcome to deploy to mobile but be careful because that's the site we're using as "production" for the stuff Tomasz asked us to do [19:49:21] so if it goes down and we're not around to help, it could be bad [19:49:41] milimetric: isn't it also pulling from limn-mobile-data? [19:49:46] and erosen, YuviPanda, yes, annotations are supported [19:49:47] milimetric: sorry for giving a go ahead on something I didn't know much about... [19:50:09] I'm definitely using -dev for checking changes and pulling to mobile-reportcard only when I've verified it all works. [19:50:09] yes, but they're separate sites so that we can deploy to mobile only when we know everything's stable and happy [19:50:16] yep, that's cool [19:50:18] yeah, I understand :) [19:50:27] just saying that sometimes the deployment breaks because of SSH, labs, etc. [19:50:51] yeah, I hardly expect myself to need to deploy once we've it settled down [19:51:01] cool beans [19:51:10] yeah, especially when you guys get that stat1001 access [19:51:28] yeah [19:51:39] milimetric: ther'es a '3 day waiting period' before anyone gets access, so I'll get access on monday [19:51:57] oh and YuviPanda, I fixed limnpy and pushed the changes to your generate.py script [19:52:08] so about annotations, they live in an annotations node, here's an example: https://github.com/wikimedia/limn-data/blob/master/graphs/timeseries_example.json [19:52:10] it now allows you to put that 'mobile' directory in the datafiles path [19:52:11] erosen: yup, i'm happy to see that gross hack go :) [19:52:16] search for "annotation-group" on that page [19:52:44] cool, thanks erosen [19:53:17] milimetric: thanks for the annotation link, I've been wondering about taht [19:53:34] they're sort of ugly right now - you have been warned [19:55:04] \me is warned [19:55:16] * erosen …. [20:49:41] [travis-ci] develop/87fb722 (#111 by milimetric): The build passed. http://travis-ci.org/wikimedia/limn/builds/5352363 [20:54:24] [travis-ci] develop/b82e8a7 (#112 by milimetric): The build passed. http://travis-ci.org/wikimedia/limn/builds/5352500 [21:00:31] drdee, i'd rather not deploy filter today, its 4pm here and I need to leave right at 5 [21:02:05] I'm doing another run now [21:02:20] ottomata: did you get an10 ip stuff sorted out? [21:02:21] relaxed the discarded rules a bit more [21:02:26] *discarding [21:02:26] average_drifter: what is the status? [21:02:40] erosen: we'll find out in about 20-30m [21:02:42] yes [21:02:43] well [21:02:53] erosen, what do you mean? [21:02:56] it is back online [21:03:01] ottomata: that's all [21:03:03] aye [21:03:05] yup shoudl be back up [21:03:15] ottomata: i tried to use pig a while ago and couldn't [21:03:19] right yeah [21:03:21] sorry about that [21:03:24] no worries [21:03:25] it should be ok now [21:03:27] wasn't a big deal [21:05:21] kraigparkinson: on my way [21:05:35] drdee okey dokey [21:05:52] drdee: do you still have time for testing the pageviews pig script? [21:07:31] i can do a knowledge transfer this afternoon [21:08:17] i am on the 6th this afternoon [21:12:03] drdee: sounds good. i'm on 3 now, but I'll be back up in a bit [21:27:25] drdee, still there? i think I lost you. [21:27:33] trying to get back in [21:27:38] ok [21:28:53] whole hangouts seems to be unavailable to me [21:29:44] hmph... [21:29:48] I could restart it. [22:10:05] drdee: where u at? [22:10:31] drdee, i didn't deploy filter, i'm leaving like now, had meetings was working on other things [22:10:33] will do monday [22:20:41] don't see drdee erosen [22:20:48] i am here [22:20:50] hold on [22:20:51] hehe [22:21:06] oh :) i don't need you, evan did ^^ [22:33:12] hey everyone. just tried to edit a crontab of mine on stat1 and got a "no space left on device" error. df -h shows that /tmp is currently 100% full. Any idea why? [22:35:29] errr [22:48:58] J-Mo: /dev/mapper/stat1-root 14597932 5901940 7963628 43% / [22:49:15] are you sure the device is full error came form stat1? [22:50:26] erosen: /dev/mapper/stat1-tmp 50G 48G 0 100% /tmp [22:50:42] aah yeah [22:50:43] the tmp directory seems to be full. That's where my crontab lives. [22:50:45] i was just look at that [22:51:39] i can make some space [22:52:04] thanks :) I don't need much! [22:52:28] hmm [22:54:43] just deleted 11g but not difference yet [22:54:49] I'm compressing another 17G [22:55:07] man, my english really suffers on IRC [22:57:04] Thanks! IRC is it's own language anyway. [23:00:29] J-Mo: still tarring [23:00:36] kk [23:02:05] drdee: org.apache.pig.impl.logicalLayer.FrontendException: ERROR 1066: Unable to open iterator for alias COUNT [23:04:02] i'll swing by later [23:04:13] sounds good [23:04:32] kraigparkinson: on my way [23:04:54] average_drifter: do you have a folder / tar with all the maxmind databases in it? [23:05:36] i have a folder in my tmp directory and I want to free up some space, but I don't want to have to download all the files again [23:21:03] J-Mo: something weird is happening [23:21:11] can't seem to make spcea [23:22:22] cool. I'll work on something else. Thanks for looking into it. Ping me if you figure out the cause or find a solution? [23:22:59] will do [23:23:28] J-Mo: one thought: it's possible that cron itself doesn't use tmp, but one of the libraries in your code does [23:23:36] in fact it might even be gcat, if you are using that [23:24:18] jmorgan@stat1:~$ crontab -e [23:24:18] /tmp/crontab.ujygzK: No space left on device [23:24:18] Creation of temporary crontab file failed - aborting [23:24:35] ah [23:24:40] well that seems pretty straightforward [23:24:55] i'm just curious be that hasn't been an issue for my cron scripts... [23:24:56] the third line is new. in previous attempts it just gave the first line of output. [23:25:29] I'd be happy to move my crontab. this was just the default location it saved to when I created it. [23:25:45] * J-Mo doesn't know how to move his crontab >_< [23:26:31] hmm [23:26:52] what is you $EDITOR? [23:29:03] umm… for cron? Whichever one is default. I don't remember off hand. I just evoke it with -e and it opens. IIRC, looks like nano or pico. [23:29:28] yeah, i'm just wondering if it is actually your editor which is causing the issue [23:29:55] if you used a different editor, it might not create the temprary file [23:30:09] er, nvm [23:30:14] just tested and got hte same issue [23:30:38] I think you can still manually edit your crontab [23:30:46] it's not the best practice, but it should work [23:31:18] yeah, if I can find it. I don't actually see it in tmp! [23:31:31] well it is just putting the copy of the file in tmp [23:32:12] the real file probably lives in /etc/crontab [23:34:12] there is a crontab there, but it's not mine. [23:34:18] yup [23:34:21] trying to find it myself [23:57:57] kraigparkinson: running a few minutes late [23:58:14] np