[00:00:01] milimetric: i can do it now [00:00:06] nuria: I have time later if you want [00:00:13] milimetric: it's fine [00:00:20] should we do dashiki too? [00:00:22] I don’t know why it disapeared form the calendar [00:00:23] milimetric: will do [00:00:36] milimetric: give me 10 mins [00:00:41] dashiki has annotations too :-) [00:00:59] kevinator: you should edit the annotations actually, to make sure they say what you want them to say [00:02:05] kevinator: https://meta.wikimedia.org/wiki/Dashiki:CategorizedMetrics and follow the pageName "links" from the json (they're not links - but they could be in a fancy config viewer!) [00:02:24] yup, i was just reading that page. [00:02:47] Erik was just asking me about Vital Signs and I want to creat an annotation for Rolling Active Editor [00:03:51] I’d like to know when I can tell him annotations are live (once I edit them). [00:19:17] milimetric: thanks for taking care of the dashboard so quickly. :) [00:19:26] np [00:19:33] that's our job :) [00:19:41] doing it well [00:35:42] kevinator: wikimetrics is down after i deployed, trying to see what is going one [00:35:46] "going on" [00:35:59] ok [00:42:58] milimetric: hey [00:43:05] got a sec? [00:43:12] ori: sure [00:43:52] first, i owe you an explanation regarding the blog stats stuff -- i ended up being a little uncertain about what tilman required, which is why the abacist thing stalled, but i think i have a better idea now, so i'll be looking at your patch again shortly. [00:44:39] no prob about that, thanks for the explanation [00:44:39] second, kevinator just stopped by to ask me about the x-analytics header and why i reverted something-or-other. i think he was prompted to ask by something you had said -- can you explain? i don't remember the status of that myself, to be honest. [00:44:57] we have other stuff to put in a private dashboard and I got good use out of playing with highcharts either way [00:45:02] yes [00:45:30] ok, so we have an X-Analytics extension that's waiting to be deployed when we're "ready" [00:45:30] ohhhhh, were you asking about the mobilefrontend commit? [00:45:58] and I understood that one of the blockers was that the MobileFrontEnd extension changed the X-Analytics header in a way incompatible with the new extension [00:46:04] in other words, they should both append [00:46:12] and then I saw that you patched MobileFrontEnd to fix this [00:46:17] and then I saw that you reverted said patch [00:46:34] so I was confused and didn't know what the status was and whether or not we're clear to deploy X-Analytics [00:46:46] my patch had a stupid bug, not fundamentally related to the change [00:46:51] ah [00:46:56] I think I re-did it after correctly, but I am confused myself now, since it has been a while [00:47:01] give me a moment to take a look and re-assess [00:47:04] no prob [00:47:05] (thanks for clarifying) [00:47:09] thanks for looking into it [00:52:58] OK, now I see. There are two WebResponse objects in MediaWiki-Core: WebResponse and FauxWebResponse. The latter is used for unit-testing. I saw that the latter had a getHeader method and assumed the former did too, and my patch called that method. Because the test suite uses FauxWebResponse, the test passed, but it broke in production. [00:53:31] It was a little stupid to have that method only on FauxWebResponse, so I added it to WebResponse as well, and then reverted the revert on MobileFrontend, because the code was now correct (https://gerrit.wikimedia.org/r/#/c/171891/) [00:53:50] so *that* particular problem is all sorted out. But -- [00:54:20] aw, there's always a butt [00:54:55] there is presently no hook that is reliably run just before outputting the response headers [00:55:46] because the actual flushing of the output buffer is done in several different places in core, by calling native PHP ob_flush() & co. functions. [00:56:38] so it's easy to have the extension inject the X-Analytics header into the response for the majority of cases, but there are tricky edge cases (for things like 404s, IIRC) [00:56:53] what we could do about it is: [00:57:10] huh... interesting [00:58:00] (a) Move all the output buffer handling code in core into a single location, and make the various places that currently flush the output buffer delegate to that (and then we could add a hook) [00:58:38] (b) Make a list of the types of requests that wouldn't have the header and ask the analysts if they'd be OK with that [00:58:52] definitely not b :) [00:59:02] anytime I start saying would you be ok if you didn't have ... [00:59:04] they're like NO [00:59:11] or (c) Reconsider whether doing it in MediaWiki is really such a good idea, and instead apply the header at a different level (either Varnish, as before, or Apache). [00:59:53] ok, thank you very much ori. This makes perfect sense and we shall discuss tomorrow [01:00:02] np, thanks for the poke [01:00:07] have a good evening [01:00:25] you too :) [01:04:23] kevinator: I forgot this morning at tasking :( [01:04:32] but the bug with generate.py that baha reported yesterday [01:04:51] that's a fairly annoying thing for them - their graphs are broken until it's fixed [01:05:06] i gotta go collapse [01:05:07] <3 nite [01:05:13] ok, [01:05:14] nite [01:05:21] we’ll talk about it after standup [03:32:52] VisualEditor, Analytics-EventLogging: VisualEditor not loading unless edit page is refreshed with the cache cleared. - https://phabricator.wikimedia.org/T86112#965184 (Negative24) I have now tested it with different browsers and machines and can reproduce it on the Performance wiki but not any other. I looked... [04:21:38] milimetric: thank you :) [04:21:47] no problem [04:22:21] milimetric: code is managed via github. so, if needed I can send pull request? [04:22:31] s/I/We [04:22:43] (or it is in Gerrit too?) [04:22:59] kart_: that's just a gerrit mirror, I just find it easier to browse in githu [04:23:21] so yeah, you can manage that code yourselves in gerrit - feel free to change / review / merge amongst yourselves [04:23:33] but until that puppet change I linked you to is merged, nothing will run on stat1003 [04:23:56] also - be sure to sign up for our Event Logging and Dashboarding talk we're planning for the Developer Summit (in a couple of weeks) [04:24:06] VisualEditor, Analytics-EventLogging: VisualEditor not loading unless edit page is refreshed with the cache cleared. - https://phabricator.wikimedia.org/T86112#965205 (Negative24) Other people are having the same problem [[ http://www.mediawiki.org/wiki/Extension_talk:VisualEditor#Nothing_happens_when_I_click... [04:26:14] milimetric: Sure. I'll be there. [04:53:25] (PS1) Milimetric: Build for release [analytics/dashiki] - https://gerrit.wikimedia.org/r/183773 [04:53:37] (CR) Milimetric: [C: 2 V: 2] Build for release [analytics/dashiki] - https://gerrit.wikimedia.org/r/183773 (owner: Milimetric) [04:57:39] Analytics-Tech-community-metrics: Graph waiting time for shell access requests - https://phabricator.wikimedia.org/T66702#965235 (scfc) I don't quite understand why there needs to have been a problem in the past to ensure that there is not any in the present (or future). For shell access requests, we adverti... [07:39:07] It seems we need .gitreview [07:41:50] nope. git.wikimedia.org doesn't know about limn-language-data :( [07:42:00] YuviPanda: can you check? ^ [07:47:11] (PS1) KartikMistry: Add .gitreview file [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183787 [07:47:16] well, added. [07:53:45] (PS1) KartikMistry: Correct list of wikis [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183788 [08:04:42] kart_: https://github.com/wikimedia/analytics-limn-language-data exists :) [08:18:37] YuviPanda: yes. submitted stuffs :) [08:32:35] (CR) Amire80: [C: 1] Correct list of wikis [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183788 (owner: KartikMistry) [08:41:36] (PS2) Gilles: Add image right clicks to stats [analytics/multimedia] - https://gerrit.wikimedia.org/r/182404 (https://phabricator.wikimedia.org/T75962) (owner: Gergő Tisza) [08:41:54] (CR) Gilles: [C: 2] Add image right clicks to stats [analytics/multimedia] - https://gerrit.wikimedia.org/r/182404 (https://phabricator.wikimedia.org/T75962) (owner: Gergő Tisza) [08:41:59] (Merged) jenkins-bot: Add image right clicks to stats [analytics/multimedia] - https://gerrit.wikimedia.org/r/182404 (https://phabricator.wikimedia.org/T75962) (owner: Gergő Tisza) [08:47:50] (PS1) Gilles: Add image right-click stats to dashboard [analytics/multimedia/config] - https://gerrit.wikimedia.org/r/183793 (https://phabricator.wikimedia.org/T75962) [09:44:53] (PS2) KartikMistry: Correct list of wikis [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183788 [12:44:16] Analytics-Tech-community-metrics: Graph waiting time for shell access requests - https://phabricator.wikimedia.org/T66702#965573 (Qgil) p:Low>Volunteer? Well, it is a matter of priorities. I'd rather focus on areas that we know are problematic and benefit from metrics, like {T78768} or {T78639}. So no... [12:47:09] Analytics-Tech-community-metrics: div.navbar-inner obscures content on Internet Explorer 10 - https://phabricator.wikimedia.org/T55635#965582 (Qgil) http://korma.wmflabs.org/ has changed a lot since this bug was reported. @ashley, can you check whether this problem still occurs in IE, please? [12:53:02] Analytics-Tech-community-metrics: Allow contributors to update their own details in tech metrics directly - https://phabricator.wikimedia.org/T60585#965594 (Qgil) Now imagine for a minute that their Phabricator profile would contain such details. [13:00:34] (PS1) QChris: Document the need to update the Oozie overiew diagram [analytics/refinery] - https://gerrit.wikimedia.org/r/183841 [13:00:36] (PS1) QChris: Update Oozie overview diagram for webrequsest/refinery [analytics/refinery] - https://gerrit.wikimedia.org/r/183842 [13:59:02] Analytics-Tech-community-metrics: div.navbar-inner obscures content on Internet Explorer 10 - https://phabricator.wikimedia.org/T55635#965645 (ashley) Open>Resolved @Qgil: http://korma.wmflabs.org/ looks good on Internet Explorer 11 under Windows 7. I don't have access to IE10 myself, so I used browsers... [14:15:04] Analytics-Tech-community-metrics: Add i18n/l10n support to korma.wmflabs.org - https://phabricator.wikimedia.org/T62073#965663 (Nemo_bis) [14:22:12] (PS1) QChris: For Oozie's legacy_tsv stop depending on all webrequest_sources [analytics/refinery] - https://gerrit.wikimedia.org/r/183849 [14:44:00] ottomata: Thanks for fixing the maven change yesterday. I only today saw in the logs how broken the change was :-/ [14:44:20] oh, that reminds me [14:44:41] the transition to 0.4, did that do...something, to the ability of maven to compile and clean up the refinery-core project, for anyone who isn't me? [14:44:50] or am I just crazy? [14:45:34] Ironholds: When did you last pull? [14:45:41] * Ironholds thinks [14:45:43] 4, 5 days ago? [14:46:36] Either pull, or cherry-pick [14:46:39] https://gerrit.wikimedia.org/r/#/c/183509/ [14:46:49] Ironholds: ^ [14:47:03] HAH [14:47:04] ja [14:47:07] qchris, wanna see something funny? [14:47:16] Cat pictures? [14:47:18] Sure! [14:47:31] https://gerrit.wikimedia.org/r/#/c/181939/8/refinery-tools/pom.xml was my fix for the problem [14:47:50] * Ironholds will rebase his change next patch [14:47:58] :-) [14:48:07] yay, I can do basic pattern recognition! [14:52:59] hmm, qchris, i'm looking at this needs_webrequest_* change [14:53:16] Trying to join the hangout and badly failing again :-( [14:53:45] i unnnooooo [14:53:47] ottomata: The issue is that the zero jobs got blocked on faulty bits and uploads partitions. [14:53:53] yeah [14:54:23] um, maybe we sould have separate coordinators for each tsv? since they each use different datasets [14:54:53] We're having like a dozen tsvs. [14:55:00] yeahHHhhHhHh [14:55:01] hm [14:55:02] Would you want to maintain a dozen coordinators? [14:55:08] Instead of just one? [14:55:18] noooOOoooo [14:55:20] i would not . [14:55:23] hm [14:55:28] Me neither :-) [14:56:48] I jnow the _unchecked trick is not super clean. [14:56:57] s/jnow/know/ [14:57:06] but it's actually not bad. Is it? [14:57:20] milimetric: Can you do you hangout magic once more please? [14:57:47] I am hanging in "Trying to join ..." -> "Party is over" mode again. [14:57:59] Since I-do-not-know-how-many-tries. [14:58:01] qchris: let's be scientific :) [14:58:07] i just removed you - try now [14:58:18] * qchris tries [14:58:41] Still saying "Trying to join the call. Psease wait..." [14:59:35] qchris: ok now I added you, try now :) [14:59:36] Now it switched to "This party is over" [14:59:40] Ok. Retrying. [15:00:04] It is again saying "Trying to join the call. Please wait..." [15:00:16] Hey :-D [15:05:56] Analytics-Tech-community-metrics: Add i18n/l10n support to korma.wmflabs.org - https://phabricator.wikimedia.org/T62073#965731 (Qgil) p:Low>Volunteer? In fact, even if the principle of i18n/L10n is always good (says the non-native English speaker), considering that korma is about tech metrics and most o... [15:23:03] ottomata1: So about the need_webrequest_* stuff. [15:23:16] How to do this in a nicer way? [15:24:17] §Analytics-sprint-2015-01-08, Analytics-EventLogging: Client Engineer passes null to EventLogging::logEvent for optional boolean fields [8 pts] - https://phabricator.wikimedia.org/T78325#965767 (Milimetric) a:Milimetric [15:24:49] qchris: , yeah, thinking about that, and checking emails [15:24:52] also was just thinking [15:25:00] if we turn of bits, is this going to be a problem? [15:25:23] If then upload recovers fully, the problem is solved. [15:25:54] But I have to admit that it always felt strange to depend [15:26:04] oh, no [15:26:08] on all four webrequest_sources when some tsvs really only needed one of them. [15:26:12] i mean, won' t turning of bits cause more problems? [15:26:19] camus won't be creating directories for bits. [15:26:50] Now I get your point. [15:26:51] so, using the unchecked source won't help [15:27:12] Then I'll remove the bits altogether until the [15:27:20] first TSV comes around that needs it. [15:27:34] VisualEditor, Analytics-EventLogging: VisualEditor not loading unless edit page is refreshed with the cache cleared. - https://phabricator.wikimedia.org/T86112#965770 (Jdforrester-WMF) >>! In T86112#964172, @Negative24 wrote: > I'll try with a different browser when I get home. Jdforrester, you tested it on t... [15:28:27] hmm, ok [15:35:01] (PS1) QChris: Remove dependence on bits for Oozie's legacy TSVs jobs [analytics/refinery] - https://gerrit.wikimedia.org/r/183859 [15:35:57] (CR) QChris: "Dropping of depencence on bits is in" [analytics/refinery] - https://gerrit.wikimedia.org/r/183849 (owner: QChris) [15:44:30] moving to cafe, back shotly [15:53:30] qchris: me no comprendo how... a job can succeed in some nodes but fail in others..for example: oozie job -info 0019273-141210154539499-oozie-oozi-C [15:54:40] * qchris looks [15:55:59] nuria: Can it be that someone modified hdfs://analytics-hadoop/tmp/mobile-apps/coordinator.xml ? [15:56:33] (or one of the dependent things) [15:56:57] It does not look like it's failing/not failing on some nodes, but mostly on the time it is run. [15:57:17] "Error Message : E0738: The following 1 parameters are required but were not defined and no default values are available: archive_job_output_workflow_file" [15:58:52] nuria: See the mtime of /tmp/mobile-apps/workflow.xml in hdfs. [16:00:46] So I think the files that your coordinator uses changed while [16:00:58] the coordinator started new workflows. [16:01:15] And the changed files did not meet the requirements of the original coordinator. [16:01:37] Anyways ... since there is no response. I am heading off to lunch. [16:05:16] baack [16:06:24] (CR) Ottomata: [C: 2] Document the need to update the Oozie overiew diagram [analytics/refinery] - https://gerrit.wikimedia.org/r/183841 (owner: QChris) [16:06:30] (CR) Ottomata: [V: 2] Document the need to update the Oozie overiew diagram [analytics/refinery] - https://gerrit.wikimedia.org/r/183841 (owner: QChris) [16:06:40] (CR) Ottomata: [C: 2 V: 2] Update Oozie overview diagram for webrequsest/refinery [analytics/refinery] - https://gerrit.wikimedia.org/r/183842 (owner: QChris) [16:07:34] VisualEditor, Analytics-EventLogging: VisualEditor not loading unless edit page is refreshed with the cache cleared. - https://phabricator.wikimedia.org/T86112#965870 (Negative24) >>! In T86112#965770, @Jdforrester-WMF wrote: >>>! In T86112#964172, @Negative24 wrote: >> I'll try with a different browser when... [16:07:48] (CR) Ottomata: [C: 2 V: 2] For Oozie's legacy_tsv stop depending on all webrequest_sources [analytics/refinery] - https://gerrit.wikimedia.org/r/183849 (owner: QChris) [16:08:02] (CR) Ottomata: [C: 2 V: 2] Remove dependence on bits for Oozie's legacy TSVs jobs [analytics/refinery] - https://gerrit.wikimedia.org/r/183859 (owner: QChris) [16:10:05] hey nuria, you there? [16:10:11] ottomata: yesss [16:10:38] qchris: sorry, in meeting, thank you [16:11:02] qchris_away: thank you, meeting just started so i did not see your response [16:11:36] re your mobile apps change. so, qchis and I argued a bunch about how to name directories, and i think we agreed not to put verbs in the directoy names unless it really made sense [16:11:40] see discussion here [16:11:41] https://gerrit.wikimedia.org/r/#/c/162589/4 [16:12:36] so, q: is there a particular reason to use 'hyphen' for 'mobile-apps'? [16:12:38] and [16:12:47] can we drop the 'generate_' from the directoy name? [16:12:49] maybe just: [16:12:58] hm [16:13:17] also, i think i'd prefer to make uniques the first word, if you don't mind [16:13:25] mobile_apps/uniques_daily [16:13:35] or, hm, i wonder if 'daily' should be in the name at all. not sure. [16:17:08] ottomata: just let me know how to name it and will change them [16:17:51] to the convention we agreed [16:19:25] VisualEditor, Analytics-EventLogging: VisualEditor not loading from edit button unless page is refreshed - https://phabricator.wikimedia.org/T86112#965888 (Negative24) [16:20:45] well, there is a loose agreement, so it will likely always take a little bit of back and forth [16:20:51] depending on the situation [16:20:57] but, i would go with either: [16:21:13] mobile_apps/uniques_daily [16:21:13] or [16:21:13] mobile_apps/uniques [16:21:32] maybe since your coordinators will specificly work with daily , having _daily in the name is fine [16:21:34] nuria ^ [16:22:00] ottomata: ok, mobile_apps/uniques_daily will be [16:22:33] k [16:22:46] also, nuria, since you want me to review, maybe now is a good time to remove your testing stuff [16:22:56] i guess you can leave the hour stuff in for now [16:22:57] if you want [16:23:10] but at least you can set the properties file to proper deafults [16:23:15] and override using -D on the CLI when you are testing. [16:23:16] hm [16:23:25] nuria, do you want to run this on the refined webrequest table? [16:23:26] ottomata: yes, my next patch will not have testing stuff, so it is "mergeable" right away [16:23:56] ottomata: yes, i am doing so already on my tests [16:24:22] ok cool, hm. i suppose we need a refined dataset in the datasets.xml file then, eh? [16:30:03] ottomata: ahhhi thought it was just the hive table [16:30:23] (PS1) Ottomata: Add webrequest_*_refined oozie datasets [analytics/refinery] - https://gerrit.wikimedia.org/r/183868 [16:30:27] i passed "webrequest" but i guess you are right, that is wrong! [16:30:43] well, i mean, the data has to exist in that table, so you should probably depend on that :) [16:30:58] ottomata: job needs to trigger when data exists on table [16:31:00] it'd be nice if there was a way to make a 'hive partition' dataset [16:31:03] but meh, this will work :) [16:31:04] yes [16:31:12] nuria https://gerrit.wikimedia.org/r/#/c/183868/ [16:31:15] that make sense? [16:31:21] then you just have to use the *_refined dataset name [16:31:32] and also set webrequest_refined_data_directory in your properties file [16:31:40] ottomata: YES! [16:31:52] ottomata: want me to +1 on that or adding it to my changeset [16:32:06] +1 we merge, you rebase :) [16:32:42] (CR) Nuria: [C: 1] Add webrequest_*_refined oozie datasets [analytics/refinery] - https://gerrit.wikimedia.org/r/183868 (owner: Ottomata) [16:34:26] ottomata: k, will incorporate these changes too, in meeting now, but getting back to this right after [16:35:16] k cool [16:35:41] nuria, when you submit the next patchset... get rid of trailing spaces! :) [16:36:19] ottomata: yes, yes [16:54:48] (CR) Ottomata: [C: 2 V: 2] Add webrequest_*_refined oozie datasets [analytics/refinery] - https://gerrit.wikimedia.org/r/183868 (owner: Ottomata) [17:21:59] (CR) Ottomata: "Also, spaces instead of tabs please!" (8 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/181939 (owner: OliverKeyes) [17:23:50] Ironholds: you have been reviewed! [17:24:09] milimetric: tell me about this [17:24:09] https://gerrit.wikimedia.org/r/#/c/183734/1/manifests/misc/statistics.pp [17:24:11] you ready? [17:24:13] they ready? [17:25:36] Analytics-EventLogging: "[undefined] Missing property "action"" ext.eventLogging.core.js:302 - https://phabricator.wikimedia.org/T86319#966053 (Quiddity) NEW [17:28:51] ottomata, yay! [17:28:55] in a meeting but will get to it :) [17:37:10] lunchtime! [17:57:52] milimetric: few patches, if you have free time :) [17:58:40] nuria: can we turn off the sendBeacon experiment today? [17:58:44] or wait until next week? [17:59:11] ori: our 30 days caching expiration will start tomorrow so better next week. [17:59:19] as soon as monday really [17:59:50] :/ [18:00:08] i still don't really understand why anything except a sufficient volume of data matters here [18:00:16] it doesn't have to cover all articles [18:01:59] is it that we don't yet have enough information about the people sendBeacon doesn't work for? [18:02:55] kart_: feel free to merge the patches yourself and see how they affect the graph [18:03:25] ottomata: got a chance to merge this: https://gerrit.wikimedia.org/r/#/c/183734/ [18:03:25] ? [18:03:31] kart_: ^ that has to happen first [18:08:53] milimetric: do I have access? :D [18:09:16] to merge that? [18:09:27] yes [18:09:34] kart_: you can't merge the puppet change, but you should be able to merge the limn-language-data change [18:10:00] checked: I don't have limn-lang-data change access. [18:10:18] oh that sux, uh... let me see how we do that [18:10:35] who on the language team would you like to give +2 to? [18:10:57] milimetric: me and Amir should be ok. [18:11:18] Amire80 [18:11:21] what about Joel? This repository is more multi-purpose than just these queries [18:11:32] milimetric: sure. Joel too. [18:11:39] ah, ja, so that change is ready to go? [18:11:42] repo exists, etc? [18:12:58] ottomata: yep [18:13:54] ottomata: also, how do we give people rights to gerrit repos without asking Alexandru (that's my code name for our favorite beardy guy so he doesn't find this in the logs) [18:14:38] uhhh [18:14:48] iunno? give you the rights to give rights? [18:14:59] sounds good [18:15:54] ottomata, could you check out the legacy UDF when you get a sec, too? [18:16:00] milimetric: good night. Very very sleepy :) [18:16:31] k, have a good weekend kart_ [18:16:42] ottomata: i must be doing something wrong when passing parameters with -D as i always run intopermits issues [18:16:44] ja, Ironholds, that is next after ananths stuff :) [18:16:51] yay! thankee :) [18:17:07] Ironholds: i will try to take a look today too [18:17:15] ta! [18:17:31] ottomata: my command: "oozie job -run -Duser=nuria -Darchive_directory=hdfs://analytics-hadoop/tmp/mobile-apps/archive -config coordinator.properties" [18:25:57] ja? [18:26:32] ottomata: that looks good right? [18:27:15] ja thikn so [18:27:34] hmm, you prob need to set the oozie_directory [18:27:56] manually, since your coodinator.xml will be somewhere different than /wmf/refinery... for testing [18:27:56] ja? [18:33:09] nuria: https://graphite.wikimedia.org/render/?width=1613&height=780&_salt=1420799381.949&target=eventlogging.overall.raw.rate&target=eventlogging.overall.valid.rate&from=01%3A00_20150108&until=08%3A00_20150109 [18:33:55] milimetric: the mobile schema issue right? I think they probably need to know at what they are sending [18:34:15] milimetric: i forwarded the thread to mobile-tech [18:34:21] yeah, i mean, the problem's persisted since Christian saw it this morning [18:34:27] ok, good [18:34:35] milimetric: right, cause it comes from mw deployment [18:34:53] nuria: want to join the oozie Q&A session? [18:34:57] oh ok [18:34:59] milimetric: nothing has changed so until they do chnages on their end it will not validate [18:35:01] kevin says NO [18:35:05] yep [18:35:11] milimetric: makes sense? [18:35:57] ottomata: i *think* i have graduated from the intro session but plis let me know otherwise [18:36:09] np [18:37:06] milimetric: if most schemas have issues it's likely a problem on our end, 1 schema not validating after a code deploy it's likely and error on the code sending events, let me make that more clear on the thread [18:40:58] milimetric: let me look at logs and such [18:53:59] (CR) Amire80: [C: 1] "I'm not that much of a Gerrit expert." [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/183787 (owner: KartikMistry) [18:57:19] mforns: how are you doing with the EL server? [18:57:54] Analytics-Wikimetrics: Story: Wikimetrics User uses tooltip to view all projects for an expanded cohort member - https://phabricator.wikimedia.org/T86332#966345 (Capt_Swing) NEW [18:58:01] nuria, I opened the port, but it continued failing, and did not find the port whitelist you said... [18:58:03] :[ [18:58:16] I'm in the oozie Q&A [18:58:19] mforns: k [18:58:30] you saw that Ori has merged the change? [18:59:14] nuria, [18:59:21] ^ [18:59:57] https://www.irccloud.com/pastebin/vtiKFiOx [19:00:55] thanks nuria [19:09:33] Analytics-Wikimetrics: Story: Wikimetrics user views full list of cohort members - https://phabricator.wikimedia.org/T86335#966385 (Capt_Swing) NEW [19:35:39] nuria or mforns: where are the "set up dev server" instructions: http://www.mediawiki.org/wiki/Extension:EventLogging/Guide#Installing_the_EventLogging_devserver [19:35:48] that link leads to itself and is an empty section [19:36:03] milimetric, yes also got that [19:36:07] wait [19:36:25] here's something: http://www.mediawiki.org/wiki/Extension:EventLogging [19:36:51] And also, in the link you passed, grep for: Installing the EventLogging devserver [19:39:18] what I did to install it was: 1) pull latest changes on EventLogging 2) go to: /vagrant/mediawiki/extensions/EventLogging/server 3) sudo python setup.py install 4) eventlogging-devserver --port 8100 [19:40:01] ok, so you just run it manually, got it [19:40:08] yes [19:40:33] oh, I also added eventlogging role to vagrant [19:45:56] milimetric, however I continue having problems with the server (ERR_CONNECTION_RESET) [19:46:21] hm [19:46:23] nuria has given me some hints, I followed them, but nothing [19:46:33] i just got done fixing the documentation there, i'm installing it now [19:47:33] mforns: here, you can reword/moveL: https://www.mediawiki.org/wiki/Extension:EventLogging#installing_dev_server [19:48:29] thanks! [20:16:35] milimetric, nuria: I managed to execute the eventlogging-devserver in my machine (not vagrant) and access it. It seems there's another error that fires when a n event is received. I'll have a look at that, and see if that's related to the change [20:17:33] mforns: right, i saw that ori merged your change but we should make sure that the dev server works and that there are no errors besides the formatting [20:17:50] of cours [20:17:54] *course [20:31:06] nuria: I give up, how do you run the ext.eventLogging.js tests? [20:31:34] the python and php ones I get, but the JS ones are not documented somewhere I can find [20:31:40] milimetric: https://www.mediawiki.org/wiki/Manual:JavaScript_unit_testing [20:32:09] k, i'll read and add link from EL testing [20:32:11] milimetric: let me know [20:32:18] if it works [20:34:24] yes nuria, it worked without any setup actually. You just visit "Special:JavaScriptTests" and select QUnit. Magic! [20:34:36] Analytics-Tech-community-metrics: Add i18n/l10n support to korma.wmflabs.org - https://phabricator.wikimedia.org/T62073#966695 (Nemo_bis) I'm not aware of any production Wikimedia statistics site, or any site *about* development, which is *not* localised. Do you have examples? [20:35:53] (CR) Ottomata: [WIP] UDF to get country code from IP address UDF to determine client IP address given values from remote_addr and XFF headers Change-Id: I6 (4 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/183551 (owner: Ananthrk) [20:52:05] ottomata, do you think it's okay if I do some CR of the GeoUDF above? [20:53:29] please [20:54:16] Ironholds: I think I am not a reviewer on the Legacy pageview udf thing? [20:54:18] i don't see it anyway... [20:54:26] huhw [20:54:28] will add you [20:55:58] (CR) OliverKeyes: [WIP] UDF to get country code from IP address UDF to determine client IP address given values from remote_addr and XFF headers Change-Id: I6 (8 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/183551 (owner: Ananthrk) [20:56:48] ottomata, done! both the comments and the adding-as-reviewer [20:57:06] ananthrk, awesome patch :). It took me so goddamn long to get to grips with MaxMind's, ah, "style", of writing APIs. [20:57:20] you seem to have done so really darn fast! [21:00:52] Analytics, Wikipedia-App-iOS-App, MediaWiki-extensions-UniversalLanguageSelector, Mobile-Apps, Language-Engineering, Mobile-Web, Wikipedia-App-Android-App: there should be a comparison of clicks count on interlanguage on different platforms - https://phabricator.wikimedia.org/T78351#966795 (Deskana) p:Triag... [21:02:48] (CR) Ottomata: "Ok, I was just talking to Nuria a bit about this." [analytics/refinery/source] - https://gerrit.wikimedia.org/r/183551 (owner: Ananthrk) [21:06:10] (CR) Nuria: "Please have in mind that mappers can also be multithreaded: https://hadoop.apache.org/docs/r2.3.0/api/org/apache/hadoop/mapreduce/lib/map/" [analytics/refinery/source] - https://gerrit.wikimedia.org/r/183551 (owner: Ananthrk) [21:07:36] Ironholds: your editor needs tabs vs spaces rehab! :p [21:12:27] Analytics-EventLogging: "[undefined] Missing property "action"" ext.eventLogging.core.js:302 - https://phabricator.wikimedia.org/T86319#966929 (Quiddity) [21:21:02] (CR) Ottomata: "Ja, but I really doubt Hive uses this. We should check." [analytics/refinery/source] - https://gerrit.wikimedia.org/r/183551 (owner: Ananthrk) [21:28:34] (CR) Ottomata: "Fix yo tabs! :)" (14 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/182971 (owner: OliverKeyes) [21:32:53] milimetric or mforns : can one of you help me for a second [21:32:58] I need something proofread [21:32:59] sure [21:33:08] I’ll share the doc with you [21:33:12] ok [21:37:12] ottomata: can we deploy the "webrequest_mobile_refined" dataset definition to 1002? [21:37:40] ottomata: that way i can test with the true dataset job is going to run with, rest of changes are done [21:40:06] kevinator, do you want me to comment on the right or just modify if I see something [21:40:11] ottomata: you around? [21:40:32] mforns: depends. if it’s a typo, go ahead. otherwise post a comment [21:40:43] kevinator, ok [21:41:09] ottomata: holaaa? [21:42:19] yessh [21:42:21] nuria: ? [21:42:49] ottomata: can we deploy the newjar with the never datasets [21:43:03] ottomata: i.e. with the change we merged this morning [21:43:44] Analytics-Tech-community-metrics, Wikimedia-Git-or-Gerrit: Active code review users on a monthly basis - https://phabricator.wikimedia.org/T86152#967050 (Nemo_bis) >>! In T86152#963373, @Umherirrender wrote: > having that under control of wmf is better for updates and traffic. {{citation needed}} [21:44:28] ? [21:44:45] oo, actually, that new dataset is kinda broken, i mean, it breaks other things, because of hte addition of that refined webrequest directory variable [21:44:52]