[08:08:32] <grrrit-wm>	 (PS9) Nuria: Adding coding guidelines to README.md file [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120542
[08:08:34] <grrrit-wm>	 (CR) jenkins-bot: [V: -1] Adding coding guidelines to README.md file [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120542 (owner: Nuria)
[08:52:05] <grrrit-wm>	 (PS10) Nuria: Adding coding guidelines to README.md file [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120542
[08:57:20] <grrrit-wm>	 (PS6) Nuria: Added tests for coalesced recurrent reports [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/122638 (owner: Csalvia)
[11:02:51] <grrrit-wm>	 (PS10) Nemo bis: Adapt dates used in test 10 to be relative to now [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[11:03:25] <grrrit-wm>	 (CR) Nemo bis: "I merged the two commit messages proposals. I'm looking forward to seeing the tests pass!" [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[11:18:46] <average>	 Nemo_bis++
[11:21:27] <average>	 qchris++
[11:29:04] <Nemo_bis>	 :)
[11:29:05] <average>	 Nemo_bis: hey, just notified Hashar that the tests didn't trigger. what's your e-mail so I can Cc you too ?
[11:31:50] <average>	 ah yeah, it was in gerrit all along
[11:31:51] <average>	 sorry
[11:35:05] <average>	 oh they did trigger
[11:35:16] <average>	 omg, I'm so silly
[12:31:18] <grrrit-wm>	 (CR) QChris: "> The analytics1012 issue is fixed, [...]" (1 comment) [analytics/kraken] - https://gerrit.wikimedia.org/r/121531 (owner: Ottomata)
[12:32:04] <Ironholds>	 yay!
[12:50:49] <grrrit-wm>	 (CR) QChris: [V: 1] "Yeah, sure. You like to see my name in commit messages :-(" [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[13:42:16] <grrrit-wm>	 (PS11) Nemo bis: Adapt dates used in test 10 to be relative to now [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[13:42:51] <grrrit-wm>	 (CR) Nemo bis: "Sorry for the name. If you actively dislike mentions, you should feel free to directly remove them." [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[13:54:27] <grrrit-wm>	 (CR) QChris: [C: 1] "> Sorry for the name." [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[14:01:09] <qchris_meeting>	 ottomata: Standup :-D
[14:01:22] <ottomata>	 ack
[14:19:03] <grrrit-wm>	 (CR) Stefan.petrea: "In a previous life, in a standup meeting, I mentioned that for change 96738 I would prefer to have only comments , and not patchsets from " [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[14:58:50] <grrrit-wm>	 (CR) Stefan.petrea: "Note for the brave and the corageous: Feel free to make any action on this gerrit change." [analytics/kraken] - https://gerrit.wikimedia.org/r/96738 (owner: QChris)
[15:01:44] <ottomata>	 hey, i have a few atom.io invites, if anyone wants
[15:01:56] <ottomata>	 it seems cool, but also kinda slow for me
[15:07:02] <chsalvia>	 freenode is being flaky today
[15:44:05] <j0kr>	 whois ottomata
[15:49:00] * average  thinks a leading / was missed

[15:50:59] <ottomata>	 whois j0kr?!?!?!
[15:51:01] <ottomata>	 EH?!
[15:51:01] <ottomata>	 EH!?
[15:51:06] <ottomata>	 :p
[15:51:25] <j0kr>	 I'm Kevin.
[15:51:39] <j0kr>	 I don't have a cloak set up yet.
[15:52:38] <ottomata>	 qchris_away: ah you are away
[15:53:54] <grrrit-wm>	 (CR) Ottomata: "Interesting." [analytics/kraken] - https://gerrit.wikimedia.org/r/121531 (owner: Ottomata)
[15:56:02] <grrrit-wm>	 (PS4) Ottomata: Adding camus wrapper script and camus properties file [analytics/kraken] - https://gerrit.wikimedia.org/r/121531
[17:04:03] <qchris>	 ottomata: re. What's up?
[17:05:20] <ottomata>	 yo, nuthin, just the patchset
[17:06:22] <qchris>	 Oh. Ok.
[17:06:30] <qchris>	 I just reviewed it and wondered ...
[17:06:44] <qchris>	 Do we actually rebase commits or just merge for kraken?
[17:06:51] <qchris>	 I guess just merge ... right?
[17:07:34] <grrrit-wm>	 (CR) QChris: [C: 2] Adding camus wrapper script and camus properties file [analytics/kraken] - https://gerrit.wikimedia.org/r/121531 (owner: Ottomata)
[17:07:43] <ottomata>	 i'd rebase, but whatever
[17:07:53] <grrrit-wm>	 (PS5) QChris: Adding camus wrapper script and camus properties file [analytics/kraken] - https://gerrit.wikimedia.org/r/121531 (owner: Ottomata)
[17:08:09] <grrrit-wm>	 (CR) QChris: [C: 2 V: 2] Adding camus wrapper script and camus properties file [analytics/kraken] - https://gerrit.wikimedia.org/r/121531 (owner: Ottomata)
[17:27:32] <grrrit-wm>	 (CR) Milimetric: [C: 2] Add ability to create recurrent reports via UI [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric)
[18:09:01] <grrrit-wm>	 (CR) QChris: [C: 2] Adapt dates used in test 10 to be relative to now [analytics/wikistats] - https://gerrit.wikimedia.org/r/123603 (owner: Stefan.petrea)
[18:14:05] <grrrit-wm>	 (Abandoned) QChris: DO NOT COMMIT. Dummy commit to have Jenkins rerun tests [analytics/wikistats] - https://gerrit.wikimedia.org/r/123595 (owner: QChris)
[18:14:20] <grrrit-wm>	 (PS3) QChris: Fix typo [analytics/wikistats] - https://gerrit.wikimedia.org/r/118257 (owner: Nemo bis)
[18:14:28] <grrrit-wm>	 (PS3) QChris: Comment some path tests which overrode standard ones [analytics/wikistats] - https://gerrit.wikimedia.org/r/118261 (owner: Nemo bis)
[18:14:34] <grrrit-wm>	 (PS3) QChris: Remove all trailing whitespace [analytics/wikistats] - https://gerrit.wikimedia.org/r/118266 (owner: Nemo bis)
[18:14:42] <grrrit-wm>	 (PS5) QChris: Typofix in comments [analytics/wikistats] - https://gerrit.wikimedia.org/r/118366 (owner: Nemo bis)
[18:14:50] <grrrit-wm>	 (PS5) QChris: [Full dump analysis] Reduce edits_only and reverts_only intricacy [analytics/wikistats] - https://gerrit.wikimedia.org/r/118436 (owner: Nemo bis)
[18:21:50] <grrrit-wm>	 (CR) QChris: [C: 2] Fix typo [analytics/wikistats] - https://gerrit.wikimedia.org/r/118257 (owner: Nemo bis)
[18:23:48] <grrrit-wm>	 (PS6) QChris: Typofix in comments [analytics/wikistats] - https://gerrit.wikimedia.org/r/118366 (owner: Nemo bis)
[18:24:37] <grrrit-wm>	 (CR) QChris: [C: 2] Typofix in comments [analytics/wikistats] - https://gerrit.wikimedia.org/r/118366 (owner: Nemo bis)
[18:26:09] <grrrit-wm>	 (PS4) QChris: Typofix in squids report [analytics/wikistats] - https://gerrit.wikimedia.org/r/100368 (owner: Nemo bis)
[18:27:08] <grrrit-wm>	 (CR) QChris: [C: 2] Typofix in squids report [analytics/wikistats] - https://gerrit.wikimedia.org/r/100368 (owner: Nemo bis)
[18:29:02] <grrrit-wm>	 (PS4) QChris: Move download limiter to proper place and comment it as it's only a test [analytics/wikistats] - https://gerrit.wikimedia.org/r/92056 (owner: Nemo bis)
[18:29:32] <grrrit-wm>	 (PS3) QChris: Add perl dependencies to a CollectMailArchives.pl comment [analytics/wikistats] - https://gerrit.wikimedia.org/r/91998 (owner: Nemo bis)
[18:29:48] <Nemo_bis>	 wohooooo success
[18:29:52] <grrrit-wm>	 (PS3) QChris: Archives are downloaded in .txt.gz format: fix matching and opening [analytics/wikistats] - https://gerrit.wikimedia.org/r/92066 (owner: Nemo bis)
[18:31:14] <qchris>	 Nemo_bis: I merged all the things that I could easily verify and were sufficiently small.
[18:31:20] <qchris>	 The dashboard at
[18:31:25] <qchris>	 https://gerrit.wikimedia.org/r/#/projects/analytics/wikistats,dashboards/default
[18:31:30] <qchris>	 is now much cleaner.
[18:31:49] <qchris>	 Is it ok if I leave the rest for ezachte, as I do not understand too much of wikistats?
[18:38:15] * average  thinks 63478 could be closed on bugzilla

[18:44:40] * qchris  thinks average is right.

[18:44:50] <qchris>	 I closed the bug.
[18:45:43] <average>	 nice :)
[19:21:20] <grrrit-wm>	 (CR) QChris: "> I suspect that these errors have something to do with the fact that" [analytics/kraken] - https://gerrit.wikimedia.org/r/121531 (owner: Ottomata)
[19:34:55] <qchris>	 Not sure about policies ...
[19:35:15] <qchris>	 Say $PERSON is a volunteer, and signs an NDA. Can she get access to stat1003?
[19:37:22] <halfak>	 Hey qchris.  That is a good question.
[19:37:38] <halfak>	 I'm not aware of a policy.  I suppose ottomata might be.
[19:38:18] <halfak>	 In the past, managers were given access to stat1/003 upon request.
[19:38:20] <qchris>	 Ok. I guess I'll have to wait until ottomata is back.
[19:38:27] <Ironholds>	 yeah, it's more of an Ops question.
[19:38:44] <Ironholds>	 I know we've had volunteers with /root/ in the past, but none are active any more.
[19:38:54] <qchris>	 There are some volunteer home directories on stat1/stat1003, so I thought ... maybe :-D
[19:39:26] <Ironholds>	 now-volunteers or always-volunteers?
[19:39:32] <halfak>	 ^
[19:39:56] <Ironholds>	 because we suck at cleaning up after ourselves
[19:40:03] <qchris>	 Some now-volunteers. Some where I am not sure.
[19:40:06] <Ironholds>	 Leslie still has her home directories. No way to get to them, because she left, but.. ;p
[19:40:10] <Ironholds>	 examples of the latter?
[19:40:29] <qchris>	 No.
[19:40:42] <qchris>	 I do not want to be responsible of people loosing their account.
[19:40:50] <qchris>	 (If they still had it as an oversight)
[19:41:14] <halfak>	 Seems like they should though.
[19:41:33] <qchris>	 Depends :-)
[19:41:36] <Ironholds>	 well, plus they presumably don't have valid SSH keys, rendering it irrelevant
[19:41:42] <Ironholds>	 in fact, have you looked at it that way?
[19:41:48] <Ironholds>	 compare the directory list to the puppet manifest.
[19:42:18] <Ironholds>	 If there are entries that are in one but not in the other, they should probably be discounted from any calculation of our current practices.
[19:42:35] <qchris>	 Puppet files need not reflect policies or possibilities.
[19:43:00] <halfak>	 Hey milimetric.  If I needed to file a request for some support from analytics devs, where's the right place to do that?
[19:43:05] <qchris>	 Ottomata will remove doubt when he comes back :-)
[19:43:24] * halfak  is not sure who to bother about these things until Kevin is on IRC

[19:43:53] <halfak>	 Actually, qchris might know.  See my message to milimetric above.
[19:45:18] <qchris>	 halfak: Sounds like you need help right away. Can do your support thing right away?
[19:45:23] <qchris>	 Or does it take more time?
[19:45:39] <halfak>	 I'm going to need some data purged from EventLogging after a 90 day window per our data retention guidelines.
[19:45:47] <halfak>	 I suspect the process to be easy.
[19:45:59] <halfak>	 But I need to know that someone will do it before we start collecting the data.
[19:46:18] <qchris>	 I guess that part has to go through our product owner.
[19:46:25] <yuvipanda>	 is it gonna be purged from the json deep storage files too?
[19:46:28] <yuvipanda>	 and any backups we have?
[19:46:28] <halfak>	 Is that Kevin?
[19:46:37] <halfak>	 Yes yuvipanda
[19:46:51] <halfak>	 "deep storage files"
[19:47:01] <yuvipanda>	 alright. do keep me in the loop when that happens, halfak
[19:47:17] <yuvipanda>	 *cold* storage files might be a better description but deep storage sounds cooler
[19:47:23] * yuvipanda  digs a 10km hole for storage

[19:47:38] <halfak>	 10km hole is pretty deep.
[19:48:22] <halfak>	 http://imgur.com/6PSKTU3
[19:48:32] <yuvipanda>	 haha
[19:48:39] <Ironholds>	 qchris, no, my point was that if volunteers have file storage but not access to that file storage, it probably indicates they're in the category of "people who left"
[19:49:42] <halfak>	 yuvipanda, re. loop keeping.  Affirmative. :)
[19:49:49] <yuvipanda>	 halfak: :)
[19:50:34] <milimetric>	 halfak: Kevin is j0kr (with a zero) in case you didn't know
[19:50:35] <qchris>	 Ironholds: Good point. However ... stat1003 should be fairly clean of "people who left"
[19:50:48] <Ironholds>	 s'true
[19:50:52] <halfak>	 Wut!
[19:51:02] * halfak  goes to add IRC handly to contact list. 

[19:51:05] <halfak>	 Thanks milimetric
[19:51:07] <milimetric>	 but yeah, I think he's the one to ping about dev work halfak
[19:51:26] <milimetric>	 but feel free to ask questions on here and ping any one of us, we're all friends :)
[19:51:44] <yuvipanda>	 qchris: btw, do you guys have a Nexus instance for maven or somesuch?
[19:51:48] <halfak>	 Totally.  I'm just trying to be respectful of your time/process. :)
[19:51:52] * yuvipanda  has http://tools.wmflabs.org/android-maven-repo setup

[19:52:13] <milimetric>	 btw halfak: I just got done bragging about your analysis that went into fine-tuning the EEVS metrics to Erik M.  He said he was going to look deeper into it
[19:52:22] <qchris>	 yuvipanda: nexus is dead. All hail arichva.
[19:52:28] <yuvipanda>	 arichva
[19:52:29] <yuvipanda>	 ?
[19:52:36] <yuvipanda>	 I am just scping things to an http server ;)
[19:52:48] <qchris>	 http://archiva.wikimedia.org/
[19:52:51] <milimetric>	 :) archiva is like apt.wikimedia.org but for java stuff
[19:53:09] <yuvipanda>	 DNS dead?
[19:53:15] <milimetric>	 it's our production sanctioned "install JVM stuff" solution
[19:53:31] <qchris>	 yuvipanda: for nexus: yes. for archiva: no.
[19:53:33] <yuvipanda>	 milimetric: yeah, I use maven heavily in the app too.
[19:53:43] <milimetric>	 yeah, the search team does too
[19:53:48] <yuvipanda>	 qchris: no, there seems to be unrelated DNS issues, plz2ignore :)
[19:53:49] <halfak>	 Woot milimetric.  :)
[19:54:06] <milimetric>	 i think spreading the word about archiva is one of the suggestions for the newly-proposed tech talks 2.0
[19:54:25] <qchris>	 yuvipanda: ottomata even did some writeup ... let me find that
[19:54:31] <yuvipanda>	 qchris: ah!
[19:54:33] <yuvipanda>	 yeah
[19:54:34] <yuvipanda>	 that'll be nice
[19:54:49] <qchris>	 yuvipanda: https://wikitech.wikimedia.org/wiki/Archiva
[19:54:51] <yuvipanda>	 although hopefully there's an easy-to-use publisher for it so I can move things off tools.wmflabs.org to that
[19:55:04] <yuvipanda>	 although it isn't that big a priority, since I am not deploying stuff to cluster
[19:55:27] <qchris>	 yuvipanda: But I hope I can get us to use a simpler maven setup. The current requirements are not really nice.
[19:55:56] <yuvipanda>	 qchris: oh? just a simple file / http based one?
[19:56:15] <qchris>	 "mvn deploy"?
[19:56:23] <qchris>	 "mvn release"?
[19:57:00] <qchris>	 I mean ... not sure what the use case is.
[19:57:36] <qchris>	 https://wikitech.wikimedia.org/wiki/Archiva#Deploying_to_Archiva
[19:58:24] <yuvipanda>	 qchris: hmm, for me it is just putting some 3rd party stuff that isn't possible on maven central (specifically the android libs)
[19:58:35] <yuvipanda>	 well, not 'not possible' just that harder to do
[19:58:40] <yuvipanda>	 I'll probably stick to tools for onw
[19:58:53] <qchris>	 WMF archiva already has jars that are not on central.
[19:59:04] <qchris>	 camus for example.
[19:59:19] <yuvipanda>	 right
[19:59:28] <qchris>	 So you could use archiva for that.
[20:00:10] * yuvipanda  files bug

[20:02:19] <yuvipanda>	 qchris: filed https://bugzilla.wikimedia.org/show_bug.cgi?id=63698
[20:02:46] <qchris>	 yuvipanda: Looks good :-D
[20:02:52] <yuvipanda>	 qchris: :)
[20:47:17] <grrrit-wm>	 (CR) QChris: [C: -1] Adding coding guidelines to README.md file (4 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120542 (owner: Nuria)
[21:01:03] <terrrydactyl>	 trying to run scripts/tests and getting this error: ImportError: No module named mock. i tried running scripts/install, but not sure if mock was installed properly because it didn't fix the error.
[21:01:16] <terrrydactyl>	 anyone have an idea of how to fix this?
[21:01:41] <average>	 maybe
[21:01:49] <average>	 https://pypi.python.org/pypi/mock/1.0.1
[21:01:56] <average>	 terrrydactyl: can you check if this is installed ?
[21:02:13] <average>	 maybe just   pip install --user mock
[21:03:32] <qchris>	 Today wikimetrics saw a merge that added "mock==1.0.0" to requirements.txt
[21:04:00] <qchris>	 terrrydactyl: Are you on vagrant?
[21:04:02] <terrrydactyl>	 average: oops, i just realized that i installed scripts/test to my local environment, but not my vagrant one.
[21:04:15] <terrrydactyl>	 qchris: might be why. haha
[21:04:36] <qchris>	 terrrydactyl: :-D
[21:04:39] <terrrydactyl>	 looks like i don't get that error anymore.
[21:04:45] <qchris>	 \o/
[21:06:13] <terrrydactyl>	 but i'm getting some test failures that seem to be unrelated to my changes
[21:06:36] <terrrydactyl>	 (at least i think they
[21:06:45] <terrrydactyl>	 're unrelated. i git stashed my changes)
[21:07:01] <qchris>	 terrrydactyl: Let me fire up my vagrant instance, so I can reproduce (will take some minutes)
[21:07:10] <terrrydactyl>	 okay, thanks qchris :)
[21:15:46] <qchris>	 terrrydactyl: Same here ... seeing many 'E's
[21:15:52] <qchris>	 (on vanilla master)
[21:16:13] <qchris>	 IIRC there was some talking about upgrading some parts and upgrade scripts.
[21:17:01] <qchris>	 milimetric: You around?
[21:17:13] <qchris>	 nuria: ^
[21:17:57] <milimetric>	 hi qchris
[21:18:13] <milimetric>	 oh I'm talking to terrrydactyl, I missed her ping earlier
[21:18:23] <qchris>	 Does the latest wikimetrics change require special upgrade?
[21:18:29] <qchris>	 Ah. Ok.
[21:18:34] <milimetric>	 ok, so it seems people are having problems in general
[21:18:41] <milimetric>	 let's talk in channel then
[21:18:49] <terrrydactyl>	 yep
[21:19:03] <milimetric>	 latest wikimetrics just has some new dependencies
[21:19:11] <milimetric>	 which should be installed via scripts/install
[21:19:21] <terrrydactyl>	 i ran scripts/instal and it fixed the mock problems
[21:19:28] <qchris>	 (I ran "alembic upgrade head" and "scripts/install")
[21:19:32] <milimetric>	 ah, ok
[21:19:56] <terrrydactyl>	 http://paste.ofcode.org/AsNkuKZ2SCzHD5QxnGMS6q
[21:19:59] <milimetric>	 terrrydactyl: did you run alembic upgrade head?
[21:20:03] <terrrydactyl>	 this is what my scripts/test looks like
[21:20:07] <terrrydactyl>	 no, i will run that now
[21:20:30] <milimetric>	 ok, that will get the database into the right state
[21:20:46] <milimetric>	 then, you'll have to rebase your change on master and fix any conflicts, but I can help with that
[21:21:07] <milimetric>	 qchris: are you having problems as well?
[21:21:27] <qchris>	 Yes.
[21:21:47] <qchris>	 ERROR: test_create_reports_for_missed_days_0 (tests.test_models.test_run_report.RunReportClassMethodsTest)
[21:22:04] <qchris>	 IntegrityError: (IntegrityError) (1062, "Duplicate entry '428-2014-03-23 00:00:00' for key 'uix_report'") 'INSERT INTO report ...
[21:22:06] <milimetric>	 well, so we don't have to work on this now, I can get up early tomorrow and help if you're going to bed qchris
[21:22:30] <qchris>	 milimetric: No need to help me fix this.
[21:22:36] <milimetric>	 qchris: yes, that is a known issue.
[21:22:45] <milimetric>	 a very few of the tests are non-deterministic
[21:23:01] <qchris>	 I was just trying to verify with terr rydactil that wikimetrics master is valid.
[21:23:05] <milimetric>	 the reason is mentioned in the last commit message under the "Missing" section
[21:23:21] <milimetric>	 thanks for helping with that
[21:23:28] <milimetric>	 and sorry I was MIA
[21:23:43] <terrrydactyl>	 so was qchris' problem different from mine?
[21:24:05] <milimetric>	 everyone has qchris' problem, it's a problem with integration testing
[21:24:15] <terrrydactyl>	 ah okay
[21:24:28] <qchris>	 I like the feeling the words "non-deterministic tests" cause in my head. ... It makes me dizzy a bit :-D
[21:24:31] <milimetric>	 there's a bug filed for it, so for now we should ignore errors with the RunReportClassMethodsTest tests
[21:24:46] <milimetric>	 hahaha qchris, yeah, it's not as crazy as it sounds
[21:25:04] <milimetric>	 basically I try to stop the scheduler so that I can run some tests as if the scheduler isn't started
[21:25:15] <milimetric>	 and it works most of the time
[21:25:17] <milimetric>	 but fails sometimes
[21:25:22] <qchris>	 :-)
[21:25:49] <milimetric>	 terrrydactyl: so how goes it with your setup?
[21:26:20] <terrrydactyl>	 well, alembic upgrade head didn't change anything.. was that supposed to be run in vagrant?
[21:26:54] <terrrydactyl>	 i got 2 fewer errors but 2 more failures, heh
[21:27:03] <terrrydactyl>	 after running the upgrade
[21:27:12] <milimetric>	 alembic upgrade head should be run in whatever environment you're developing under
[21:27:26] <milimetric>	 if you're developing in vagrant, you don't need to do anything outside of vagrant
[21:27:49] <terrrydactyl>	 vagrant is where all my databases are, so it should probably be running there.
[21:28:00] <terrrydactyl>	 but it didn't fix my issue. :(
[21:28:46] <milimetric>	 so you're having the same errors
[21:28:51] <milimetric>	 did you remember to stop the queue?
[21:28:56] <milimetric>	 and check for ghost queues?
[21:29:26] <terrrydactyl>	 oh. what. i totally checked earlier, but i guess i started it up again somewhere.
[21:29:28] <terrrydactyl>	 ugh
[21:29:58] <terrrydactyl>	 that might be the problem
[21:30:00] <milimetric>	 no worries - important thing is:
[21:30:03] <milimetric>	 1. i believe you
[21:30:31] <milimetric>	 2. the non-deterministic tests problem I was talking above is solved the same way as this silly ghost queue problem
[21:30:39] <terrrydactyl>	 do you know why sometimes pkill doesn't work and you have to individually delete the processes?
[21:30:49] <milimetric>	 so when we address that newly filed bug (which should be soon), all this will be much simpler
[21:30:53] <milimetric>	 yes
[21:31:12] <milimetric>	 kill -9 is required when the processes aren't responding to normal killing
[21:31:29] <milimetric>	 and one reason they can do that is if they're in some abnormal state
[21:32:33] <terrrydactyl>	 ah okay
[21:32:58] <qchris>	 milimetric: The you mentioned is
[21:33:00] <qchris>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=63680
[21:33:01] <qchris>	 ?
[21:33:03] <terrrydactyl>	 well, hopefully i won't have to worry about it later
[21:33:14] <milimetric>	 yes qchris
[21:33:18] <qchris>	 milimetric: Thanks.
[21:33:36] <milimetric>	 if we fix that bug we can basically ignore integration tests as part of our development flow
[21:33:36] <terrrydactyl>	 milimetric: okay, tests ran ok, thanks!
[21:33:52] <milimetric>	 which solve ALL of the problems mentioned above
[21:34:04] <milimetric>	 and is a good idea in general
[21:34:18] <milimetric>	 :) I have a lot to learn about how to grow projects organically
[21:34:32] <terrrydactyl>	 we're all learning
[21:34:38] <milimetric>	 I've never had so many people hop onto a project in the middle
[21:34:54] <milimetric>	 people usually hop on after a distinct phase
[21:34:59] <milimetric>	 anyway, good to hear terrrydactyl
[21:35:06] <milimetric>	 soldier on! :)
[21:36:30] <terrrydactyl>	 btw, i kind of built the delete_wiki_user branch on top of the delete_cohort branch since i wanted to use some of the code as reference. will this cause problems? i can always stash my changes and make a new branch, but wonder if that's necessary.
[21:39:05] <grrrit-wm>	 (PS1) QChris: Revert "Add ability to create recurrent reports via UI" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/124751
[21:39:56] <grrrit-wm>	 (PS2) QChris: Revert "Add ability to create recurrent reports via UI" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/124751
[21:41:45] <milimetric>	 terrrydactyl: not necessary to change your flow, I think the way you set it up seems sensible
[21:42:05] <terrrydactyl>	 okay cool, i will carry on then.
[21:42:08] <milimetric>	 when you rebase, rebase your delete_cohort branch first, solve conflicts there, then rebase your delete_wiki_user branch second
[21:42:43] <milimetric>	 qchris: are you saying you'd like to merge those reverts or are you just adding it for your own purposes?
[21:43:17] <milimetric>	 qchris: to be clear, non-deterministic tests existed before this, the issue is not this commit
[21:43:32] <qchris>	 They did?
[21:43:34] <qchris>	 :-(((
[21:43:46] <qchris>	 (They never hit me up to today)
[21:43:48] <milimetric>	 yes, integration tests are sort of that way by nature
[21:43:53] <qchris>	 ?
[21:43:59] <qchris>	 They shouldn't be :-D
[21:44:05] <milimetric>	 integration between multiple asynchronous systems anyway
[21:44:36] <milimetric>	 because by nature you have to start/stop different pieces of the system or time.sleep to let other parts catch up
[21:45:07] <qchris>	 Starting/Stopping pieces is part of the documentation ... I had the queue stopped.
[21:45:17] <milimetric>	 oh no, I mean dynamically as part of the test
[21:45:48] <milimetric>	 for example, if the scheduler mode is run by the test suite __init__
[21:45:56] <qchris>	 Those things are not a problem. When testing the xmldumps, I created and tore down mysql cluster on the fly ...
[21:45:57] <milimetric>	 so let's say I want to test days_missing
[21:46:52] <milimetric>	 well, they're certainly a problem.  maybe a solvable problem, but still a problem - clearly :)
[21:47:26] <milimetric>	 so anyway, to test wikimetrics/models/report_nodes/run_report.py:RunReport.days_missing
[21:47:32] <milimetric>	 you'd have to stop the scheduler
[21:47:44] <milimetric>	 otherwise, since it runs every 1 second, after you create test data
[21:47:49] <milimetric>	 it will alter the database
[21:48:49] <qchris>	 I am not sure I like the test setup.
[21:48:54] <qchris>	 So how to move forward?
[21:48:59] <qchris>	 Should I abandon the change?
[21:49:24] <qchris>	 I do not like that we have or add non-deterministic tests.
[21:49:25] <milimetric>	 well, I definitely don't like the test setup, but that bugzilla bug, once fixed, should make things a lot better, no?
[21:50:07] <milimetric>	 so I think we can move the integration tests to run maybe on the staging server, or to spin up a vagrant instance and run them there
[21:50:14] <qchris>	 But breaking something broken even more just to clean up afterwards is not a viable approach to me.
[21:50:20] <milimetric>	 either way, we would remove them from the main development flow
[21:50:57] <milimetric>	 I agree qchris, this is why I wanted desperately to talk about the scope of gerrit changes as part of our last retrospective
[21:51:44] <milimetric>	 I said I didn't want to merge this change as is, and the team pushed back on me very hard
[21:51:54] <milimetric>	 and said we would do it differently going forward
[21:52:03] <milimetric>	 so ok, let's do it differently going forward
[21:52:53] <qchris>	 But I cannot recall the team to aggree on adding non-deterministic tests.
[21:52:53] <milimetric>	 but that means that with this change I had two options:
[21:52:53] <milimetric>	 delay for another one or two weeks while we split up and clean up the change
[21:52:53] <milimetric>	 or merge
[21:52:54] <milimetric>	 people were really against the delay option
[21:53:54] <qchris>	 Meh. I'll abandon the change.
[21:54:05] <qchris>	 Not sure if this is right.
[21:54:18] <qchris>	 I was under the impression the huge change would not break anything.
[21:56:40] <milimetric>	 eh... huge change didn't technically break anything new, it just made an old problem more obvious
[21:57:09] <milimetric>	 I'll send an email to j0kr and tnegrin to try and prioritize https://bugzilla.wikimedia.org/show_bug.cgi?id=63680
[21:57:47] <qchris>	 Sounds good.
[21:57:48] <qchris>	 Thanks.
[21:59:58] <grrrit-wm>	 (Abandoned) QChris: Revert "Add ability to create recurrent reports via UI" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/124751 (owner: QChris)