[05:06:52] <fdans>	 morning!
[06:04:15] <elukey>	 early Francisco is early :)
[06:04:18] <elukey>	 morning fdans 
[06:11:40] <wikibugs>	 10Analytics, 10Analytics-Kanban: Move the Analytics infrastructure to Debian Buster - https://phabricator.wikimedia.org/T234629 (10elukey)
[07:10:04] <joal>	 hi team
[07:11:45] <wikibugs>	 (03PS21) 10Fdans: Add Pageviews Complete dumps backfilling job [analytics/refinery] - 10https://gerrit.wikimedia.org/r/597541 (https://phabricator.wikimedia.org/T252857)
[07:16:35] <elukey>	 bonjour!
[07:54:47] <wikibugs>	 (03PS2) 10Fdans: Add UDF that transforms Pagecounts-EZ projects into standard [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/597740 (https://phabricator.wikimedia.org/T252857)
[07:54:54] <wikibugs>	 10Analytics: Revision_text field of mediawiki_wikitext_current is Not properly mapped - https://phabricator.wikimedia.org/T253484 (10JAllemandou) Hi @DED - The problem is with showing the data only, not decoding rows. For instance: ` select   page_id,   page_namespace,   page_title,   page_redirect_title,   page...
[07:55:30] <wikibugs>	 10Analytics: Revision_text field of mediawiki_wikitext_current is Not properly mapped - https://phabricator.wikimedia.org/T253484 (10JAllemandou) 05Open→03Invalid
[08:00:30] <elukey>	 bbiab
[08:02:30] <djellel>	 Morning!
[08:03:24] <djellel>	 joal: I saw your reply to T253484. Is this a limitation in Hive ?
[08:03:24] <stashbot>	 T253484: Revision_text field of mediawiki_wikitext_current is Not properly mapped - https://phabricator.wikimedia.org/T253484
[08:03:45] <joal>	 djellel: I don't know :) I think it;s a bug in Hive UI
[08:03:58] <joal>	 djellel: data is fine however
[08:04:05] <joal>	 djellel: it's only the printing that failds
[08:04:17] <djellel>	 I figured. though, same goes for the command line.
[08:04:34] <joal>	 what do you mean command-line?
[08:05:03] <wikibugs>	 (03CR) 10Joal: Add special explode UDTF that turns EZ-style hourly strings into rows (032 comments) [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/596605 (https://phabricator.wikimedia.org/T252857) (owner: 10Fdans)
[08:05:13] <djellel>	 hive CLI
[08:05:41] <djellel>	 on a stat machine, I was trying to run my sql query from file and output to csv. 
[08:05:55] <joal>	 yes djellel - Hive has a bug printing, whether in CLI or in HUE (same engine)
[08:06:54] <djellel>	 not sure it's UI
[08:07:31] <djellel>	 It is reading a newline as a new record.
[08:08:26] <djellel>	 I guess I have to work with Spark
[08:08:45] <joal>	 djellel: from the 2 other queries I ran, data seems corretly extracted but printed incorrectly
[08:09:13] <djellel>	 yes, I confirm
[08:09:56] <djellel>	 using spark you mean ?
[08:10:08] <joal>	 ?
[08:10:28] <djellel>	 which queries did you run and where?
[08:10:29] <joal>	 Ah - no printing bug in spark nope
[08:10:56] <djellel>	 ok.
[08:11:29] <fdans>	 thank you for the review joal :)
[08:12:22] <djellel>	 I flagged an issue with Abstracts dumps. Who is responsible for that? I believe the bugreport should be picked up from the backlog.
[08:12:49] <joal>	 np fdans :) Let's wait for others input on whether to extract business logic or not :)
[08:15:19] <joal>	 no idea djellel - I don't even know what `Abstracts dumps` are - If they are XML related, it's probably in Ariel Glen realm (apergos on IRC)
[08:16:15] <djellel>	 T111775
[08:16:16] <stashbot>	 T111775: Infoboxes are mistaken for abstracts in page abstract dumps. - https://phabricator.wikimedia.org/T111775
[08:16:43] <joal>	 ack djellel 
[08:20:34] <djellel>	 ok, I'll ping @apergos on the task
[08:25:21] <djellel>	 joal: may I suggest that mediawiki_wikitext_current contains only the last snapshot, and mediawiki_wikitext_history all snapshots?
[08:25:30] <joal>	 nope djellel 
[08:25:45] <joal>	 djellel: we have snapshots for both, history contains full historical dumps
[08:26:32] <joal>	 djellel: we will at some point add a 'latest' partition (see https://phabricator.wikimedia.org/T252148)
[08:29:05] <djellel>	 it's more from a speed perspective. and +1 for latest! 
[08:29:26] <djellel>	 though, if `latest` is standalone table, you can partition it further.
[08:29:52] <joal>	 djellel: the snapshot field is a partition, therefore only data from the selected snapshot is read, not all
[08:30:16] <joal>	 and actually, wiki_db is a partition field as well
[08:30:45] <djellel>	 good point
[08:39:00] <djellel>	 joal: I found a solution to run my query on Hive :)
[08:39:04] <djellel>	 set hive.query.result.fileformat = SequenceFile;
[08:40:48] <wikibugs>	 10Analytics: Revision_text field of mediawiki_wikitext_current is Not properly mapped - https://phabricator.wikimedia.org/T253484 (10DED) 05Invalid→03Resolved To run the query on Hive if some fields contain a newline char:   > set hive.query.result.fileformat = SequenceFile;   reference: https://www.phdata.i...
[08:48:18] <joal>	 nice djellel 
[08:51:27] <elukey>	 I checked a bit gunicorn configs for superset, and atm we are using a prefork-sync scheme with 8 workers
[08:51:42] <elukey>	 what I want to try after 0.36 is using gevent
[08:51:56] <elukey>	 so allow each process to go async when doing I/O
[08:52:06] <joal>	 hm
[08:52:11] <joal>	 elukey: --verbose?
[08:53:12] <elukey>	 joal: gevent is a python lib to manage greenlet based on libevent. We use gunicorn as WSGI backend for Superset, currently set to just create 8 processes that each one can handle one request at the time
[08:53:58] <elukey>	 gunicorn can be set to use greenlets instead, more or less like how node works
[08:54:42] <joal>	 I think I get it - This for HTTP-level threads
[08:54:42] <elukey>	 I think that this is what Nuria experienced when she talked about Superset slowness
[08:54:52] <joal>	 right
[08:55:27] <elukey>	 she usually works when more people from SF use superset, so I think that 8 workers are not enough
[08:55:32] <elukey>	 in sync mode I mean
[08:55:35] <joal>	 makes sense
[08:56:03] <joal>	 elukey: I was also thinking - How complicated is it for us to start moving presto to collocated?
[08:56:54] <elukey>	 joal: it shouldn't be super hard but requires time to plan everything correctly, I guess that we could do it next Q
[08:57:14] <elukey>	 but we have to drop something probably if we want to prioritize it
[08:57:15] <joal>	 ok elukey - I think there would be tremendous value :)
[08:59:51] <wikibugs>	 (03PS1) 10Elukey: Release upstream version 0.36.0 [analytics/superset/deploy] - 10https://gerrit.wikimedia.org/r/598423 (https://phabricator.wikimedia.org/T249495)
[09:01:11] <djellel>	 Just in case, in enwiki:   ns = 0 is 15Mio, ns=3 is 15Mio, other namespaces have 19Mio records. 
[09:50:01] * elukey errand + lunch!
[10:21:14] <wikibugs>	 (03PS22) 10Fdans: Add Pageviews Complete dumps backfilling job [analytics/refinery] - 10https://gerrit.wikimedia.org/r/597541 (https://phabricator.wikimedia.org/T252857)
[11:16:40] <wikibugs>	 10Analytics-Kanban, 10Better Use Of Data, 10Product-Analytics, 10Patch-For-Review: Upgrade to Superset 0.36.0 - https://phabricator.wikimedia.org/T249495 (10elukey) New version deployed in staging, waiting for another round of tests before releasing to production.
[11:16:51] <wikibugs>	 10Analytics-Kanban, 10Better Use Of Data, 10Product-Analytics, 10Patch-For-Review: Upgrade to Superset 0.36.0 - https://phabricator.wikimedia.org/T249495 (10elukey) p:05Triage→03Medium
[11:17:33] <wikibugs>	 10Analytics-Kanban, 10Better Use Of Data, 10Product-Analytics: Superset Updates - https://phabricator.wikimedia.org/T211706 (10elukey)
[11:23:26] <wikibugs>	 10Analytics: Test superset running on gunicorn + gevent - https://phabricator.wikimedia.org/T253545 (10elukey)
[11:43:49] <wikibugs>	 10Analytics, 10Analytics-Kanban, 10Operations: Create a profile to standardize the deployment of JVM packages and configurations - https://phabricator.wikimedia.org/T253553 (10elukey) p:05Triage→03Medium
[11:48:42] <wikibugs>	 10Analytics, 10Analytics-Cluster, 10Analytics-Kanban, 10User-Elukey: Upgrade the Hadoop test cluster to BigTop - https://phabricator.wikimedia.org/T244499 (10elukey)
[11:52:36] <wikibugs>	 10Analytics, 10Operations, 10Traffic: Remove ganglia leftovers from ops/puppet - https://phabricator.wikimedia.org/T253555 (10ema)
[11:53:58] <wikibugs>	 10Analytics, 10Operations, 10Traffic: Remove ganglia leftovers from ops/puppet - https://phabricator.wikimedia.org/T253555 (10ema) p:05Triage→03Low
[11:54:51] <wikibugs>	 10Analytics, 10Operations, 10Traffic: Remove ganglia leftovers from ops/puppet - https://phabricator.wikimedia.org/T253555 (10ema)
[12:05:54] <wikibugs>	 10Analytics, 10Operations, 10Traffic, 10Patch-For-Review: Remove ganglia leftovers from ops/puppet - https://phabricator.wikimedia.org/T253555 (10ema)
[12:51:21] <wikibugs>	 (03PS23) 10Fdans: Add Pageviews Complete dumps backfilling job [analytics/refinery] - 10https://gerrit.wikimedia.org/r/597541 (https://phabricator.wikimedia.org/T252857)
[13:02:30] <gehel>	 joal: wdqs meeting?
[13:02:46] <joal>	 gehel: YES
[14:01:32] <mforns>	 hellooo team :]
[14:02:55] <joal>	 Hi mforns 
[14:03:06] <mforns>	 :]
[14:17:34] <wikibugs>	 (03CR) 10Mforns: [C: 03+1] "LGTM!" [analytics/reportupdater-queries] - 10https://gerrit.wikimedia.org/r/593092 (https://phabricator.wikimedia.org/T247099) (owner: 10Nuria)
[14:44:05] <nuria>	 a-team: will be working today on and off as schedule allows but not on meetings, it is a US holiday
[14:44:23] <joal>	 ack nuria 
[14:55:37] <wikibugs>	 (03CR) 10Nuria: [C: 03+2] "Merging as results coincide with the ones jennifer had earlier" [analytics/reportupdater-queries] - 10https://gerrit.wikimedia.org/r/593092 (https://phabricator.wikimedia.org/T247099) (owner: 10Nuria)
[14:55:39] <wikibugs>	 (03CR) 10Nuria: [V: 03+2 C: 03+2] Automate calculations for number of pages using wikidata items [analytics/reportupdater-queries] - 10https://gerrit.wikimedia.org/r/593092 (https://phabricator.wikimedia.org/T247099) (owner: 10Nuria)
[15:28:59] <mforns>	 heya elukey :] I'd like to install mysql under my user in an-launcher to be able to test Airflow with the ParallelExecutor. Any things I should know, be careful with or any heads up? I wouldn't like to break anything there. Also, I'm not sure I can apt install there...
[15:30:49] <elukey>	 mforns: if you need a temporary database, I can create one for you on analytics1030 (test host) so you can skip that
[15:31:15] <elukey>	 something like 'airflow_mforn_test'
[15:31:17] <mforns>	 elukey: oh, I think that would work, would be great!
[15:31:44] <mforns>	 sure, airflow_mforns_test is awesome
[15:37:32] <elukey>	 mforns: on an-launcher1001 you have a file called airflow_db_creds (in your home)
[15:37:39] <elukey>	 the db is created
[15:37:51] <elukey>	 analytics1030.eqiad.wment:3306 to access it
[15:38:00] <elukey>	 test it and tell me if it works :)
[15:44:07] <wikibugs>	 10Analytics: Grant not able to access superset - https://phabricator.wikimedia.org/T253281 (10elukey) What error is returned? Is the LDAP access failing? The uid is `gsingers` but I can't find it in httpd's logs..
[15:45:27] <mforns>	 elukey: super thanks! :D
[15:45:44] <mforns>	 configuring airflow now
[15:55:33] <mforns>	 elukey: Airflow docs ask me to: We rely on more strict ANSI SQL settings for MySQL in order to have sane defaults. Make sure to have specified explicit_defaults_for_timestamp=1 in your my.cnf under [mysqld]
[15:56:41] <mforns>	 Where in analytics1030 can I find that?
[15:59:22] <elukey>	 it is under /etc/my.cnf
[15:59:36] <elukey>	 IIRC we had to add it with Erik at the time
[15:59:38] <mforns>	 I found /etc/mysql/conf.d/mysql.cnf
[15:59:52] <mforns>	 oh
[16:00:22] <elukey>	 ah yes
[16:00:22] <elukey>	 # Required for search/airflow installation. This will also
[16:00:22] <elukey>	 # be the mariadb default some day, as the old behaviour
[16:00:22] <elukey>	 # is deprecated.
[16:00:23] <elukey>	 explicit_defaults_for_timestamp = on
[16:00:28] <elukey>	 so all good
[16:01:20] <mforns>	 ok! thx
[16:01:47] <elukey>	 going to step away for a bit!
[16:01:56] <mforns>	 k!
[16:30:14] <elukey>	 mforns: all good with the db?
[16:44:49] <mforns>	 elukey: one question
[16:45:07] <mforns>	 the password you created is for my user
[16:45:28] <mforns>	 now, I need to execute airflow as analytics, because the job is spark and needs access to the metastore
[16:46:10] <mforns>	 so, is test mysql also kerberized? will it have problems when the analytics user tries to connect to mysql using mforns credentials?
[16:46:17] <elukey>	 mforns: yes I used your username but in theory I could have used also "batman", the important bit is that you configure it in airflow's config
[16:46:32] <mforns>	 ok, cool :]
[16:46:36] <elukey>	 yes no krb :)
[16:46:46] <mforns>	 shame you didn't use batman though :[
[16:46:55] <mforns>	 xD just kidding
[16:50:58] <elukey>	 I wish I had!
[16:54:56] <mforns>	 elukey: no working... :[ it can not connect to mysql, maybe I built the connection URI wrong.. but I think it's OK, hmm
[16:55:27] <mforns>	 says access denied to mforns, using password yes
[16:57:34] <mforns>	 I used  mysql://mforns:<password>@analytics1030.eqiad.wmnet:3306/airflow_mforns_test
[17:00:04] <mforns>	 should I use analytics1030's IP directly? probably
[17:01:31] <elukey>	 in theory no
[17:01:35] <elukey>	 maybe I missed some grants
[17:03:48] <mforns>	 no, same problem
[17:04:41] <elukey>	 yes I can repro via mysql cli, lemme check what I missed
[17:04:48] <mforns>	 ok, thanks :]
[17:05:41] <mforns>	 I noticed you wrote airflow_mforn_test before (without the 's'), maybe that's it?
[17:08:56] <elukey>	 mforns: can you retry now?
[17:09:01] <mforns>	 sure
[17:09:42] <mforns>	 elukey: working!
[17:09:49] <elukey>	 mforns: sorry my bad, PEBCAK
[17:10:02] <mforns>	 no no, the opposite, thanks! :D
[17:10:06] <wikibugs>	 10Analytics, 10Analytics-Cluster, 10Analytics-Kanban, 10User-Elukey: Upgrade the Hadoop test cluster to BigTop - https://phabricator.wikimedia.org/T244499 (10elukey) Upgrade a second time and failed with a different issue. This time, I ended up with a lot of missing/under-replicated blocks and also ~7% of...
[17:24:20] <elukey>	 mforns: all good? If so I'll log off, otherwise I can help :)
[17:27:43] <mforns>	 elukey: all seems good! thanks a lot
[17:31:23] <elukey>	 super
[17:31:26] * elukey off!
[17:59:54] <wikibugs>	 10Analytics: Measure DNT usage across geographies and wikis - https://phabricator.wikimedia.org/T187376 (10Nuria) 05Open→03Declined
[18:00:27] <wikibugs>	 10Analytics: Measure DNT usage across geographies and wikis - https://phabricator.wikimedia.org/T187376 (10Nuria) declined per DNT being (mostly) a failed experiment not supported going forward by a number of browsers
[18:05:24] <wikibugs>	 10Analytics, 10Product-Analytics: [Spike] Should EventLogging support DNT? - https://phabricator.wikimedia.org/T252438 (10Nuria) >We have a whole group of smart, technically educated, privacy-minded experts weighing in and even we do not entirely agree what DNT means.  could not agree more, this is been the ca...
[22:56:26] <wikibugs>	 10Analytics, 10Event-Platform, 10Growth-Team, 10MediaWiki-Recent-changes, and 2 others: Remove deprecated RCFeedEngine support - https://phabricator.wikimedia.org/T250628 (10Krinkle) @Milimetric The above finishes the deprecation we started as part of setting up the EventBus/EventStreams consumer of RCFeed...