[06:29:38] <grrrit-wm>	 (PS1) Rfaulk: mod. switch view logic to catch queued or running jobs before checking cache. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/79312
[06:30:12] <grrrit-wm>	 (CR) Rfaulk: [C: 2 V: 2] mod. switch view logic to catch queued or running jobs before checking cache. [analytics/user-metrics] - https://gerrit.wikimedia.org/r/79312 (owner: Rfaulk)
[10:52:12] <qchris>	 milimetric: Good morning!
[10:52:26] <qchris>	 milimetric: I got confused with timzones ... was it EST or EDT?
[10:52:46] <qchris>	 milimetric: Do we have the grooming session in ~10 minutes?
[10:53:03] <average>	 qchris: hi
[10:53:16] <qchris>	 average: Hi
[10:53:33] <average>	 qchris: thanks for the patchsets
[10:53:45] <average>	 for the new ticket, I am testing the old dtree file
[10:53:55] <average>	 I am getting some test failures in kraken-pig
[10:54:12] <qchris>	 average: Really? It passes for me with the old dtree files.
[10:54:17] <qchris>	 Which tests fail for you?
[10:55:54] <average>	 qchris: please see e-mail just sent
[10:56:03] * qchris  checks email

[10:56:24] <average>	 look at the bottom of the out.txt because it's quite big
[10:56:26] <average>	 I ran with -X
[10:58:33] <average>	 qchris: what openddr.dtree file are you using ?
[10:58:41] <average>	 qchris: what package did you get it from ?
[10:58:46] <qchris>	 Version 1.13.
[10:58:51] <qchris>	 drdee sent this file to me.
[10:59:10] <qchris>	 When I started in our team, he said that we should use this file
[11:00:29] <qchris>	 I'll send you the file.
[11:02:35] <average>	 ok
[11:02:50] <average>	 qchris: so I guess I should put that file inside the libdclass-data package right ?
[11:03:07] <average>	 and also document how we can produce dtree files
[11:03:13] <average>	 that's what #1017 says right ?
[11:03:32] <qchris>	 average: I was under the impression that we used the dtree file in version 1.13 already. At least that's what drdee told me.
[11:03:47] <qchris>	 average: If we do not, let's discuss with drdee what we are actually using.
[11:04:02] <average>	 that can be answered with a diff
[11:04:24] <qchris>	 average: #1017?
[11:04:30] <qchris>	 average: Oh you mean #1070. Yes.
[11:04:41] <average>	 #1070 yes , sorry
[11:06:02] <average>	 qchris: 1.13 or 1.1.3 ?
[11:06:18] <qchris>	 I thought 1.13. Let me check again.
[11:06:44] <qchris>	 It says 1.13 for me.
[11:07:00] <average>	 user@user-K56CM:/tmp/test1$ ls
[11:07:00] <average>	 2.2.2-1  libdclass-data_2.2.2-1_all.deb  openddr.dtree.v.1.13
[11:07:00] <average>	 user@user-K56CM:/tmp/test1$ diff -q 2.2.2-1/usr/share/libdclass/openddr.dtree openddr.dtree.v.1.13
[11:07:03] <average>	 Files 2.2.2-1/usr/share/libdclass/openddr.dtree and openddr.dtree.v.1.13 differ
[11:07:06] <average>	 qchris: ^^
[11:07:31] <average>	 openddr.dtree.v.1.13  is the file you sent me on the e-mail minutes ago
[11:07:42] <qchris>	 The deb holds the 1.19 version.
[11:07:57] <qchris>	 (At least it did for me when I tried the other day)
[11:08:10] <qchris>	 Yes, they differ.
[11:08:22] <qchris>	 Acutally, they differ a lot as hinted in #1070 :-(
[11:08:41] <average>	 if I include this v.1.13 in the libdclass-data would this take us forward in solving this ?
[11:09:02] <qchris>	 Yes. But we'll have towait for an answer from tomasz
[11:09:48] <qchris>	 Sorry. I got confused.
[11:10:14] <qchris>	 I am not sure if that takes us forward, as I do not know how we currently install the dtree file.
[11:10:31] <average>	 by deb packages
[11:10:39] <qchris>	 We could also simply mask the new dtree file, if the old dtree file is already provided by some other package.
[11:10:43] <average>	 the libdclass-data deb package provides the dtree files
[11:11:43] <qchris>	 From my point of view, and by what we discussed with drdee yesterday, we should keep the old dtree file for now. I have no preference on how we realize that.
[11:11:51] <qchris>	 We could either keep the old debs.
[11:11:58] <qchris>	 Create new debs with the old files
[11:12:03] <qchris>	 or whatever.
[11:12:21] <qchris>	 I suppose the one who does the installing has to speak up what's easiest for him.
[11:14:10] <average>	 well basically I foresee this going like this :
[11:14:19] <average>	 1) I can submit a new gerrit patchset to add v.1.13
[11:14:23] <average>	 2) Faidon reviews and approves
[11:14:39] <average>	 3) Faidon uploads new packages to apt.wikimedia.org
[11:14:56] <average>	 4) Andrew can install the new libdclass-data on the cluster nodes
[11:15:16] <average>	 and then we have the dtree v1.13 available
[11:15:24] <qchris>	 Is the new (1.19) version already installed on the cluster nodes?
[11:15:35] <qchris>	 If not, we could just keep what we have.
[11:15:51] <average>	 I'm not sure if all cluster nodes have 1.19 already. I believe Andrew is in the process of installing it to all nodes
[11:15:56] <average>	 He already installed it for some nodes
[11:16:15] <qchris>	 I guess it's best to discuss with him directly.
[11:16:29] <average>	 ok
[11:16:41] <qchris>	 average: Do the tests pass with the dtree file that I sent you?
[11:19:08] <average>	 running mvn test now
[11:19:36] <average>	 qchris: https://gist.github.com/wsdookadr/2125b55e6919d76f0dbe
[11:19:41] <average>	 this is the output of "mvn test"
[11:19:45] <average>	 it's failing on kraken-pig
[11:19:54] <average>	 I got your latest patchset btw
[11:20:16] <average>	 I also replaced openddr.dtree with the one you gave me
[11:20:36] <qchris>	 Can you show me the output of those two tests (testDesktopPageviewCount, testMobilePageviewCount) as well?
[11:20:46] <qchris>	 So we can see where the tests fail?
[11:21:54] <average>	 qchris: I'm having some trouble running those
[11:21:57] <average>	 I'm in the hangout
[11:22:09] <average>	 pls join
[11:22:11] <qchris>	 Ok. Let me start the google machine.
[11:22:59] <average>	 ok
[11:24:58] <grrrit-wm>	 (CR) QChris: [V: -1] "For Ştefan:" [analytics/kraken] - https://gerrit.wikimedia.org/r/75349 (owner: Stefan.petrea)
[11:26:01] <qchris>	 average: I am in the hangout.
[11:41:24] <grrrit-wm>	 (CR) QChris: "Removing block. Tests now pass for Ştefan. It was just some missing files that were necessary to run those tests." [analytics/kraken] - https://gerrit.wikimedia.org/r/75349 (owner: Stefan.petrea)
[14:19:46] <drdee>	 heay
[14:19:51] <drdee>	 i mean hiya
[14:31:46] <average>	 drdee: hi
[14:31:53] <drdee>	 hey average!
[14:32:01] <drdee>	 great work with qchris on dclass!
[14:42:29] <average>	 thanks, qchris helped a lot
[16:50:05] <qchris>	 Anyone around whom I could discuss the wp0 log decrease before scrum?
[16:56:17] <drdee>	 ask yurik
[16:56:32] <yurik>	 ?
[16:57:01] <qchris>	 Thanks, drdee.
[16:57:11] <yurik>	 qchris, are you talking about zero.log?
[16:57:29] <qchris>	 yurik: about mobile caches in general
[16:57:39] * average  is joining hangout

[16:58:09] <qchris>	 yurik: We're about to have a meeting right now, may I come back with the problem to you after the meeting?
[16:58:24] <yurik>	 qchris, sure, just ping me here
[18:08:40] <grrrit-wm>	 (CR) Diederik: [C: 2 V: 2] "Looking good! I saw some really minor issues but we clean that up next time (whitespace, years in license header) blablabla." [analytics/kraken] - https://gerrit.wikimedia.org/r/75349 (owner: Stefan.petrea)
[18:11:36] <drdee>	 qchris, average: it seems that on jenkins it cannot find the dclassjni files
[18:11:37] <drdee>	 https://integration.wikimedia.org/ci/view/Analytics/job/Kraken/11/org.wikimedia.analytics.kraken$kraken-dclass/testReport/junit/org.wikimedia.analytics.dclassjni/DclassWrapperTest/testDclassWrapper/
[18:11:51] <drdee>	 java.lang.UnsatisfiedLinkError: no dclassjni in java.library.path
[18:15:20] <average>	 jenkins probably doesn't have the new dclass instaled
[18:15:22] <average>	 *installed
[18:16:31] <average>	 puppet production branch needs to be forked and package added to the jenkins machine
[18:20:26] <average>	 puppet$ grep "dclass" modules/contint/manifests/packages.pp  'libdclass0', 'libdclass0-dev', 'libdclass-jni', 'libdclass-java', 'libdclass-data',
[18:21:01] <average>	 http://goo.gl/BbHkcC
[18:21:04] <average>	 this is what I wanted to write
[18:21:09] <average>	 basically it seems to be already in puppet
[18:21:46] <average>	 ottomata: if new packages pop up on apt.wikimedia.org , are they automatically upgraded on machines that depend on them ?
[18:22:19] <average>	 ottomata: for example, right now, jenkins appears to be missing dclass packages
[18:36:36] <drdee>	 average: i asked ottomata to install dclass on the jenkins machine and he did that
[18:36:47] <drdee>	 so that's why i am surprised that it's not working
[18:48:28] <average>	 I'd login the jenkins machine and do troubleshooting, but I don't have access(maybe hashar can have a look). For qchris and myself the tests ran locally and no problems with the library were found. It may be that jenkins has old versions of dclass installed.
[18:50:16] <drdee>	 nah, ottomata installed it this week
[19:28:49] <milimetric>	 i gotta run and get some lunch
[19:28:53] <milimetric>	 be back soon
[19:30:19] <drdee>	 k
[19:47:06] <drdee>	 average, on  your way with wikimetrics?
[19:56:05] <drdee>	 average: is card  https://mingle.corp.wikimedia.org/projects/analytics/cards/824 clear?
[20:47:38] <grrrit-wm>	 (PS1) Milimetric: re-organizing the return value of report nodes [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/79429
[20:48:12] <grrrit-wm>	 (CR) Milimetric: [C: 2 V: 2] re-organizing the return value of report nodes [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/79429 (owner: Milimetric)
[21:07:51] <qchris>	 milimetric: Thanks for the hint about trying to make numbers match although they did not.
[21:08:22] <qchris>	 milimetric: After some playing around, I round really filter out groups, as proportions did not make sense.
[21:08:43] <qchris>	 milimetric: And in the end found a (at least for me) convincing explanation.
[21:08:53] <qchris>	 milimetric: Thanks a lot!
[21:27:46] <qchris>	 drdee: Sorry. I missed the pings about Kraken in Jenkins.
[21:27:59] <qchris>	 drdee: It seems Jenkins is still using the github repo.
[21:28:42] <qchris>	 drdee: Github has an old version of loading the library (expects fixed path, ...).
[21:29:01] <qchris>	 drdee: I'll see if I can break Jenkin's kraken :-)
[21:40:53] <drdee>	 duuhhhhh
[21:41:15] <drdee>	 qchris: i can fix that as well
[21:42:02] <qchris>	 drdee: Ok. I just got distracted about the logging problems again.
[21:42:15] <drdee>	 ok, i will take care of jenkins
[21:42:18] <qchris>	 drdee: I thought there was a slight chance to get it fixed before the weekend
[21:42:38] <qchris>	 drdee: But it turns out people who could do it are away :-/
[21:42:47] <drdee>	 :(
[21:44:21] <qchris>	 However, I am not sure if switching to gerrit is enough. I am not sure if the shared object's directory is in the ld path by default.
[21:44:37] <qchris>	 But I guess we'll find out soonish :-)
[21:47:55] <drdee>	 we will indeed :)
[22:13:30] <milimetric>	 sorry, been busy looking after that silly runaway chicken :)  have a good weekend everyone
[22:13:38] <milimetric>	 email me if you need me