[10:42:39] (PS1) Milimetric: Fix broken tests after hasty merge by me [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/121628 [10:42:56] (CR) Milimetric: [C: 2] Fix broken tests after hasty merge by me [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/121628 (owner: Milimetric) [10:47:31] (PS1) Milimetric: Improve survival metric query performance [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/121630 [12:17:21] hey qchris [12:17:28] Hi average! [12:18:09] so ottomata helped me yesterday, and I was able to have a mediawiki-vagrant install going [12:18:19] now I'm trying to re-set-up kraken on my machine [12:18:23] Awesome! [12:18:28] was there something in ~/.m2/ ? some xml that was required ? [12:18:44] There was. [12:18:54] But I this is gone since some time IIRC. [12:19:43] Just follow kraken's README.md. [12:19:49] That should get you going. [12:19:50] ok [12:25:35] Downloading: https://repository.cloudera.com/artifactory/cloudera-repos/org/wikimedia/analytics/kraken/kraken-generic/0.0.2-SNAPSHOT/maven-metadata.xml [12:25:48] this is the line it's trying to fetch from cloudera [12:25:56] that file doesn't reside ont hat url [12:28:11] hm, I hit mvn compile [12:28:16] and now it's trying to pull this file [12:28:18] Downloading: http://nexus.wmflabs.org/nexus/content/repositories/snapshots/org/wikimedia/analytics/kraken/kraken-generic/0.0.2-SNAPSHOT/maven-metadata.xml [12:29:15] qchris: I am following this now https://github.com/wikimedia/kraken/blob/master/build.md [12:30:28] average: Sorry. I missed your above messages. But I'll have leave for lunch. [12:30:35] We'll sort it out when I am back. [12:32:00] ok [12:32:11] However, not sure how relevant the build.md is. It got used to deploy stuff etc. A simple "mvn clean package" should be all that you need after README.md. [12:32:47] mvn package stalls when trying to download this [12:32:48] Downloading: http://nexus.wmflabs.org/nexus/content/repositories/snapshots/org/wikimedia/analytics/kraken/kraken-generic/0.0.2-SNAPSHOT/maven-metadata.xml [12:33:08] if it's possible to get over this hump, then I'm able to get jars [12:33:22] it can wait though, got a few more things to sort in the meantime too [13:02:29] (PS36) Nuria: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [13:09:56] average: Nexus is still listed in the repos, but the instance is down. [13:10:15] Dropping it from the root pom.xml should do the trick. [13:10:31] I am trying that and I'll prepare a patch afterwards. [13:12:35] (PS37) Nuria: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [13:19:28] (PS1) QChris: Drop nexus.wmflabs.org from list of repositories [analytics/kraken] - https://gerrit.wikimedia.org/r/121662 [13:19:59] average: kraken builds faster (and without problems) after ^ [13:20:07] qchris: will use that now [13:20:30] average: And if it works, please not only use, but also CR ;-) [13:24:06] qchris: CR=? [13:24:17] average: Sorry. Code-Review :-) [13:24:43] average: I meant to say that if it works, please also review the change and merge it, if it looks good to you. [13:25:09] qchris: agree, mvn packaging now, when I have jars I'll +2 (or +1 ?) [13:26:49] Click "Review", Select "+2 Looks good to me, approved" for "Code-Review". And then click "Publish and Submit". [13:27:09] (But of course only, if it really looks good to you and you approved) [13:27:45] ok, finished running mvn package [13:27:46] getting this [13:27:47] [ERROR] Failed to execute goal on project kraken-hive: Could not resolve dependencies for project org.wikimedia.analytics.kraken:kraken-hive:jar:0.0.2-SNAPSHOT: Could not find artifact org.wikimedia.analytics.kraken:kraken-generic:jar:0.0.2-SNAPSHOT in maven-twttr (http://maven.twttr.com) -> [Help 1] [13:28:03] also, I do mvn package in kraken/kraken-hive [13:29:05] it's weird that it looked for kraken-generic in the twitter package-repo [13:29:12] kraken-hive? ... which sha1 are you at? [13:29:32] (No kraken-hive for me on master) [13:29:48] qchris: a74ff2102a195f7d03d65b15a3cd545cc4c69827 [13:30:07] yes, it's part of https://gerrit.wikimedia.org/r/#/c/96738/ [13:30:28] average: but that's not yet merged into master. [13:30:47] Let's get master fixed, and afterwards rebase you patch on top of it. [13:30:56] ok [13:31:00] going back to master [13:34:22] ok so I'm on master [13:38:13] fetching change 121662 and trying it out [13:40:17] qchris: I still have some failing dclass tests around [13:40:20] but .. [13:40:35] apart from dclass, everything else builds succesfuly [13:40:42] I installed the packages from apt.wikimedia.org [13:40:48] average: do you have the correct dtree file? [13:41:55] qchris: the one that comes with this package http://apt.wikimedia.org/wikimedia/pool/main/d/dclass/libdclass-data_2.2.2-1_all.deb [13:42:19] Which one is that? [13:42:26] user@xw:/tmp$ sha1sum /usr/share/libdclass/openddr.dtree [13:42:26] cb5a5d6b2869b6cae72a1525185b1af7884084d1 /usr/share/libdclass/openddr.dtree [13:42:38] what sha1 do you get ? [13:42:45] e5ac59a59e5a0f2015f95139c3c855597d199c27 [13:42:52] hm, moment [13:42:55] I am on 1.13 [13:42:59] are you on 1.19? [13:43:24] wget http://garage-coding.com/releases/udfs-v1/dtrees.zip [13:43:25] 1.13 seems the correct one: README.md: Get a openddr.dtree file (version 1.13) [13:43:42] I should be able to find the right one in that zip [13:44:05] sha1sum /home/christian/sources/Wikimedia/analytics/dtree-files/* [13:44:05] e5ac59a59e5a0f2015f95139c3c855597d199c27 /home/christian/sources/Wikimedia/analytics/dtree-files/openddr.dtree-1.13 [13:44:05] cb5a5d6b2869b6cae72a1525185b1af7884084d1 /home/christian/sources/Wikimedia/analytics/dtree-files/openddr.dtree-1.19 [13:44:15] So you're indeed using 1.19. [13:44:26] Downgrade to 1.13 and the tests should pass. [13:44:40] ok, will do [13:47:27] 1355327886.dtree in dtrees.zip :) [13:47:30] that's 1.13 [13:47:33] cp-ed that, works now [13:47:40] mvn package again now [13:47:50] testDoGeoLookupIpv6US(org.wikimedia.analytics.kraken.geo.GeoIpLookupTest): /usr/share/GeoIP/GeoIPCity.dat (No such file or directory) [13:47:53] testDoGeoLookupIpv6Europe(org.wikimedia.analytics.kraken.geo.GeoIpLookupTest): /usr/share/GeoIP/GeoIPCity.dat (No such file or directory) [13:48:26] Ahhm... you did read the README.md. Right? [13:48:49] going back to the README [13:48:58] The section "## GeoIP files" seems relevant. [13:50:07] scp spetrea@stat1002.eqiad.wmnet:/home/spetrea/maxmind_archive.zip . [13:59:17] milimetric: can't hear you [14:00:29] it's loading.. [14:00:32] this is unusual [14:02:31] average: We switched to the google batcave. [14:02:43] ok [14:02:52] https://plus.google.com/hangouts/_/calendar/d2lraW1lZGlhLm9yZ19jYjM3bXU0OGNuaHRkN2hybmE4czI3b25hb0Bncm91cC5jYWxlbmRhci5nb29nbGUuY29t.c6j7qidqs491nhi7ovk9pi4h14?authuser=1 [14:04:23] Hey milimetric, I created an updated the survival metric SQL that's super performant and incorporates the archive table. https://trello.com/c/Sd60JBCf/189-optimize-survival-query [14:40:38] (PS14) Milimetric: Allow reports to be rerun [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111914 [14:40:44] (PS6) Milimetric: Add utility functions for scheduling [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119992 [14:44:05] halfak: trello is commercial/proprietary software that leads to vendor lock-in(just like Mingle, just like Google Chrome and Google Hangouts). please use bugzilla and mark your SQL query as enhancement [14:45:36] average, trello is our primary way of collaborating on projects cross-team. You'll notice that I posted the SQL on the wikipage that documents the metric and only noted this on the trello board. [14:46:12] So, average, if you want my work to fit in your process, I welcome your work to do so. I'll be happy to support it. [14:47:13] trello is a very bad choice for a project management tool from my point of view. at least use kanban(an opensource clone of trello), that way, you at least get to own your data https://github.com/somlor/kanban [14:47:53] average, I am not a project manager. I did not decide to use trello -- my team did. However, I do find it to be a satisfactory tool. [14:50:29] well, yeah, old discussion, no solution so far, maybe someone comes up with something(I know there are some discussions about this on the mailing list) [14:56:58] There's a project exploring a nice tool in labs. [14:57:04] * halfak is failing to remember the name.  [14:57:10] Looking through my email now. [14:59:16] http://fab.wmflabs.org/ [14:59:19] average, ^ [14:59:24] (CR) Nuria: [C: 2] Add utility functions for scheduling [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119992 (owner: Milimetric) [14:59:33] We're looking into switching to something like that now. [15:02:31] halfak: looks nice, why not use that instead of trello ? [15:03:06] It would mess up our process to just up and switch. It's not clear that it will be better yet, so we might have to switch back anyway. [15:03:13] So, we're reviewing it when we have time. [15:03:55] It's important to remember that the most important thing that we do is get things done. The tools that we use to get things done are of secondary concern. [15:10:56] (CR) Nuria: [C: 2] "Changes regarding testing on this patch are in a subsequent patch. Merging so as to keep connections among gerrit changes simple." [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111914 (owner: Milimetric) [15:15:40] halfak: what I'm trying to say is building opensource software while relying on proprietary software is weird.. but we can have different oppinions, I don't mind [15:16:25] Totally. I see this issue with editors like sublime text and iOS, but I wouldn't think of chasing those people down and asserting that they change. [15:34:04] ottomata: 1:1? [15:34:39] yupuppp [15:35:09] dans le hangout [15:35:15] found it [16:14:29] ... [16:21:50] (CR) Stefan.petrea: [C: 2] "This change looks good to me. Previously had a few problems building kraken, but with this change it works fine." [analytics/kraken] - https://gerrit.wikimedia.org/r/121662 (owner: QChris) [16:22:44] Thanks! [16:23:24] qchris, quick q re archiva repositories [16:23:35] Ha. Shoot. [16:26:00] so, ok, default archiva comes with two defined repositories [16:26:03] internal and snapshot [16:26:17] k [16:26:25] thus far I have been adding third party artifacts to internal, and deploying to snapshots [16:26:37] if we kept this way, we'd probably deploy releases to internal [16:26:43] but, i'm thikning we should have a 3rd repo [16:26:46] 'releases' [16:26:51] that is for wmf releases [16:26:53] Sounds good to me. [16:27:02] I'd also separate our releases from 3rd party software. [16:27:05] ja? i'm not sure it matters, its kinda nice to have them separatedx out [16:27:06] yeah [16:27:14] Totally. [16:27:15] i'm not sure where camus would go since we maintain a slight fork of it [16:27:27] i guess our releases [16:27:30] I am not listening to the camus question :-P [16:27:31] since we put our own version on it [16:27:33] haha ok [16:27:53] But yes. I also think it's probably on our plate. [16:27:55] ok carry on, thanks! i will make a releases [16:28:00] repo [16:29:28] Thinking about it again. Ja. Your camus build should totally not go into the 3rd party repository, but into the wmf (releases|snapshot) repository. [16:30:53] jaja k danke [16:38:52] (PS20) Stefan.petrea: [WIP] kraken-hive UDFs [analytics/kraken] - https://gerrit.wikimedia.org/r/96738 (owner: QChris) [16:49:04] ne more thing qchris, should I rename the 'internal' repository to 'external' then? [16:49:07] might make more sense? [16:49:11] external, releases, snapshots [16:49:12] ? [16:49:27] Sounds good to me. [16:49:36] ok! [16:49:40] Does manybubbles have an opinion about it? [16:49:57] Oh. Why is he not in the analytics channel? [16:50:02] E-D [16:50:09] s/E-D/:-D/ [16:50:16] good q, I will ask [16:57:03] (PS21) QChris: [WIP] kraken-hive UDFs [analytics/kraken] - https://gerrit.wikimedia.org/r/96738 [16:57:57] (CR) QChris: "I did not address any of the issues code-review brought up," [analytics/kraken] - https://gerrit.wikimedia.org/r/96738 (owner: QChris) [17:25:57] (PS7) Terrrydactyl: Removing cohorts from database. [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119343 [18:10:07] (Abandoned) Milimetric: Implement security for public reports [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/118068 (owner: Milimetric) [18:19:42] (PS38) Milimetric: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 [22:20:26] (CR) Milimetric: [WIP] Run recurring reports using the scheduler (5 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [22:25:50] (PS39) Milimetric: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 [22:26:30] (CR) Milimetric: [C: 1] "let's merge this sucker, it's good enough for now" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric)