[00:58:12] Analytics-Tech-community-metrics, Possible-Tech-Projects: Improving MediaWikiAnalysis - https://phabricator.wikimedia.org/T89135#1714299 (Fhocutt) This looks interesting, and I can help with exploring and working with the MediaWiki API. [00:59:30] Analytics-Tech-community-metrics: Implement some missing information from the MediaWiki API - https://phabricator.wikimedia.org/T114440#1714305 (Fhocutt) I can be a resource for this. [02:35:42] PROBLEM - Analytics Cassanda CQL query interface on aqs1002 is CRITICAL: Connection timed out [02:37:22] RECOVERY - Analytics Cassanda CQL query interface on aqs1002 is OK: TCP OK - 3.004 second response time on port 9042 [02:42:32] PROBLEM - Analytics Cassanda CQL query interface on aqs1002 is CRITICAL: Connection timed out [02:49:22] RECOVERY - Analytics Cassanda CQL query interface on aqs1002 is OK: TCP OK - 3.001 second response time on port 9042 [02:54:31] PROBLEM - Analytics Cassanda CQL query interface on aqs1002 is CRITICAL: Connection timed out [03:04:42] RECOVERY - Analytics Cassanda CQL query interface on aqs1002 is OK: TCP OK - 0.004 second response time on port 9042 [03:09:53] PROBLEM - Analytics Cassanda CQL query interface on aqs1002 is CRITICAL: Connection timed out [03:16:32] RECOVERY - Analytics Cassanda CQL query interface on aqs1002 is OK: TCP OK - 0.011 second response time on port 9042 [03:38:33] PROBLEM - Analytics Cassanda CQL query interface on aqs1002 is CRITICAL: Connection timed out [03:46:53] RECOVERY - Analytics Cassanda CQL query interface on aqs1002 is OK: TCP OK - 3.003 second response time on port 9042 [04:30:23] PROBLEM - Analytics Cassanda CQL query interface on aqs1003 is CRITICAL: Connection refused [04:47:02] RECOVERY - Analytics Cassanda CQL query interface on aqs1003 is OK: TCP OK - 0.006 second response time on port 9042 [04:55:41] PROBLEM - Analytics Cassanda CQL query interface on aqs1002 is CRITICAL: Connection timed out [04:57:13] RECOVERY - Analytics Cassanda CQL query interface on aqs1002 is OK: TCP OK - 0.998 second response time on port 9042 [09:15:51] !log Restart cassandra on aqs1002 [09:15:54] Logged the message at https://www.mediawiki.org/wiki/Analytics/Server_Admin_Log, Master [09:45:40] (CR) Joal: [C: -1] Fix inconsistent mobile uniques reports due to partial job runs (1 comment) [analytics/refinery] - https://gerrit.wikimedia.org/r/244604 (https://phabricator.wikimedia.org/T114406) (owner: Madhuvishy) [11:10:32] (CR) Joal: "Comments inline." (7 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/243990 (https://phabricator.wikimedia.org/T113521) (owner: Madhuvishy) [11:13:25] hi a-team1 [11:13:30] mforns: ! [11:13:37] xD [11:13:44] Are you good enough to really be here ? [11:13:52] yes, a lot better [11:13:54] by good, I mean well, sorry [11:13:56] :) [11:13:58] Cool :) [11:14:13] xD I understood [11:14:18] Thanks for yesterday presentation, i was really great :) [11:14:21] would have written it the same way [11:14:27] oh, cool [11:14:30] IT was ... pfff ... big fingers [11:14:47] big fingers? [11:14:57] I say that when I make typos :) [11:15:05] ok ok [11:15:08] :] [11:15:42] Like not being able to properly type is a real handicap for our kind of job ! [11:15:47] :D [11:16:17] hehehe [11:16:21] I have broken cassandra a bit this night :) [11:16:32] Hammering it with hadoop is kinda tough [11:16:49] oh, but np, pageview api is still a tier...10 system [11:16:59] :) [11:17:12] what's the problem? [11:17:34] I think there was too much pressure from hadooop (load too high) [11:17:39] aha [11:17:54] how many machines does our cassandra have? [11:18:00] I will start investigating using another insert method (more bulk style) [11:18:05] aha [11:18:07] hopefully it could reduce load [11:18:15] cassandra have 3 machine [11:18:32] mmm, it seems enough for aggregated data... [11:18:37] no? [11:19:00] And even if I tell hadoop no to use too many writers, I am still writing with 6 writers x 4 [11:19:08] :) [11:19:16] aha [11:19:16] http://ganglia.wikimedia.org/latest/?r=day&cs=&ce=&c=Analytics+Query+Service+eqiad&h=&tab=m&vn=&hide-hf=false&m=bytes_in&sh=1&z=small&hc=4&host_regex=&max_graphs=0&s=by+name [11:19:25] * mforns looks [11:19:48] There is one thing I need to ask the services team: why is there so many bytes out from cassandra [11:19:55] bytes in , ok I get it, but out ? [11:20:02] Need to investigate [11:20:05] aha [11:20:20] aren't we using our own analytics cassandra cluster? [11:20:34] We are [11:20:45] It's just that the services team knows more about cassandra than I do [11:22:48] maybe when you insert data into cassandra, cassandra returns the inserted data [11:23:02] backfilling-wise: per-project daily/hourly is done, top daily is done, and per-article daily and hourly are still ongoing [11:23:15] aha [11:23:19] The per-article ones are the big ones in term of data size [11:23:26] sure [11:24:04] But still, it shouldn't be that long: ~3G gzipped compress daily to upload to cassandra [11:24:13] Should be faster [11:24:24] * joal gets back to investigate better loading ! [11:24:30] aha [11:24:30] ok [11:54:26] a-team, I'm away for 1 hour or so [11:54:31] later ! [11:54:35] ok, later! [12:57:50] * joal is back ! [13:33:40] (CR) Ottomata: "Probably the stuff in refinery-job can stay there, just the stuff that is now in refinery-core should move to refinery-camus." [analytics/refinery/source] - https://gerrit.wikimedia.org/r/240868 (https://phabricator.wikimedia.org/T113251) (owner: Joal) [13:35:42] (CR) Ottomata: Add libjars optional arg to Camus python wrapper script (1 comment) [analytics/refinery] - https://gerrit.wikimedia.org/r/244599 (owner: Madhuvishy) [13:42:25] (CR) Ottomata: "I would structure the src/main/avro schemas in the same way the rest of the class hierarchy is structured, e.g src/main/avro/org/wikimedia" (1 comment) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/243990 (https://phabricator.wikimedia.org/T113521) (owner: Madhuvishy) [14:06:44] hey ottomata I need to borrow your permissions for a bit [14:06:57] trying to figure out what's wrong with aqs [14:10:53] k [14:10:55] wassup? [14:11:03] i saw those alerts form aqs1002 lastnihg [14:11:04] tnight [14:11:06] night8 [14:11:07] AH [14:11:11] last night* [14:13:48] ottomata: I think it has suffered from overload :( [14:14:05] ottomata: I have restarted cassandra this morning, seems to be bakc in the game [14:15:08] Analytics-EventLogging, Database: db1046 innodb signal 6 abort and restart - https://phabricator.wikimedia.org/T104748#1715187 (Milimetric) Sorry, Jaime, I missed this problem when it happened. The project we all monitor is Analytics-Backlog, but we've been meaning to clean that up, there are too many co... [14:15:48] joal: it's still returning empty responses [14:15:59] so I wanted to change the logging to get local logs so we can see what the heck is going on [14:17:47] ottomata: so I just need to be able to fiddle with /etc/restbase/config.yaml (but puppet generates that... hm...) [14:20:07] milimetric: normally you have aqs-admin rights [14:20:15] milimetric: maybe it's not enough ? [14:20:33] joal: I'm lost, as always [14:20:39] milimetric, ottomata : The problem milimetric is describing is not related to this night issue :) [14:20:41] I can't seem to do "service restbase restart" [14:20:54] nor edit /etc/restbase/config.yaml [14:21:04] sudo ? [14:21:06] milimetric: sudo [14:21:07] ? [14:21:12] :) [14:21:12] no, asks for pw [14:21:14] you can't edit the config, since that is managed by puppet [14:21:15] hm [14:21:45] well, somehow I need to be able to debug this thing, and it's getting a bit ridiculous. workers are dying all the time and I have no idea why [14:22:00] nor any way to see the logs, because logstash doesn't seem to tell me anything [14:22:01] hm, i tlooks like you should be able to sudo service restbase restart [14:22:03] milimetric: sudo service restbase restart works for me on aqs1001 [14:22:10] the last hundred or so errors have the helpful message "HOST" [14:22:13] milimetric: i can temporarily let you edit the config file! [14:22:14] :) [14:22:14] it also worked with cassandra service [14:22:17] k [14:22:17] you want aqs1002? [14:22:21] sure [14:22:24] k [14:22:50] milimetric: you'll temporary be an ops guys !!! How cool :) [14:22:58] and then i need to be able to restart restbase - joal how did you restart cass? [14:23:00] there you go [14:23:01] try now [14:23:18] milimetric: sudo service restbase restart has worked for me on aqs1001 [14:24:57] ok joal sweet, that for some insane reason worked [14:25:04] :D [14:25:17] when I go "service restbase status" it says the command "service" doesn't exist [14:25:19] :P [14:25:29] mwarf :) [14:25:33] ok, if you tail /tmp/debug.log I'm about to try and figure out what's up with these empty messages [14:25:35] on 1002 [14:26:41] milimetric: are you sudoing? [14:26:56] /usr/sbin is not in your user's path [14:27:04] I was able to "sudo service restbase restart", yes [14:27:04] you have to sudo for it to even know where the 'service' command is [14:27:08] but I can't sudo other stuff [14:27:11] right [14:27:17] you can do: [14:27:19] %aqs-admins ALL = NOPASSWD: /usr/sbin/service cassandra * [14:27:19] %aqs-admins ALL = (cassandra) NOPASSWD: ALL [14:27:19] %aqs-admins ALL = NOPASSWD: /usr/sbin/service restbase * [14:27:19] %aqs-admins ALL = (restbase) NOPASSWD: ALL [14:27:19] %aqs-admins ALL = NOPASSWD: /bin/journalctl * [14:27:29] you should be able to sudo -u restbase [14:27:30] milimetric: tailing ! [14:27:31] and do anytihng [14:27:32] or maybe [14:27:35] sudo -u cassandra [14:27:42] so, anything restbase or cassandra user can do, I think you can do. [14:28:12] joal: I think I got it, formatting so I can paste properly [14:28:32] Yeah, I have seen it as well : timeuuid, right ? [14:30:14] https://www.irccloud.com/pastebin/IPaTzUv3/ [14:30:32] milimetric: yes [14:30:36] Rahhhhh [14:30:47] Looking into it now [14:31:06] I insert an empty string as a timeuuid --> cassandra doesn't complain [14:31:49] but the driver that restbase uses to hit cassandra has a problem with it? [14:32:05] seems so :( [14:32:11] It's an unused field for us [14:32:22] created by default by restbase [14:32:32] Since it's part of primary key, can't be set to null [14:34:47] Analytics-Cluster, Analytics-Kanban: Move camus properties out of refinery and into puppet - https://phabricator.wikimedia.org/T115114#1715208 (Ottomata) NEW a:Ottomata [14:35:19] joal: ok, and I'm assuming it's not as simple as update table set uuid='not empty string'? [14:35:28] or uuid=newid() or something? [14:35:54] milimetric: it's a pain to generate in java :( [14:36:06] milimetric: I'll double check on how we can do that [14:36:08] i mean directly in cassandra [14:36:13] Yup got it [14:36:24] (reading up too) [14:36:44] meanwhile i'm going to remove this logging and let it go to logstash again [14:37:56] milimetric: ok thanks [14:39:16] it'd be nice to have the logs on disk too by default, wouldn't it? [14:39:52] yeah, especially since the errors seem to not be going to logstash [14:40:25] i'll try and see if you can have both in that streams section (https://wikitech.wikimedia.org/wiki/RESTBase#Debugging) [14:40:30] Analytics-Kanban, Database: Delete obsolete schemas {tick} - https://phabricator.wikimedia.org/T108857#1715233 (mforns) [14:41:18] joal: so when I do select * from data limit 1; I don't see uuid, is it the _tid column or something? [14:41:31] also milimetric we should have a talk about insertion rate --> not really happy of current [14:41:42] milimetric: it is [14:42:35] milimetric: providing a playground for testing [14:43:03] hm? [14:43:19] "Test_Project" keyspace [14:43:26] milimetric: --^ [14:43:37] no, I'm here :) I'm just not understanding what you mean [14:43:51] ah. ok [14:43:54] say a full thought, my interpolation is bad in the morning :) [14:44:06] and the afternoon most of the time [14:44:42] I kind of know you just enough now to know you do as I do: self depreciation as good humour :-P [14:45:08] Currently feeding a new keyspace with some data to see if we can modify [14:45:35] ah ok [14:45:47] I'm trying to figure out if I can update the _tid on a specific record and get it to work [14:45:56] any idea what default value they expect there? [14:46:02] now() [14:46:06] sweet :) [14:46:12] is the easiest way to create :) [14:49:16] I have killed aqs1003 now :( [14:51:23] joal: killed? [14:51:44] not killed, but need restart (timeout, as for 1002 before [14:51:55] 1002 was timing out a lot more this morning [14:52:00] wasn't yesterday [14:52:35] so it looks like Cassandra doesn't let you call UPDATE on part of the PRIMARY KEY [14:52:37] makes sense [14:52:39] right, before me taking care of it :) [14:52:54] milimetric: makes sense indeed :( [14:53:00] mwarf [14:53:04] looks like there's a fast "COPY" method [14:53:14] Means full reimport with a fake timeuuid right ? [14:53:41] milimetric: can be tried [14:53:43] I'm trying to find an alternative, I'm going to try deleting and re-inserting just a row just to make sure this solves the problem [14:53:50] but also looking into COPY [14:54:02] PROBLEM - Analytics Cassanda CQL query interface on aqs1003 is CRITICAL: Connection refused [14:54:25] !log Cassandra restarted on aqs1003 [14:54:27] Logged the message at https://www.mediawiki.org/wiki/Analytics/Server_Admin_Log, Master [14:55:04] milimetric: you can use the test keyspace if you want --> ready to play [14:55:52] RECOVERY - Analytics Cassanda CQL query interface on aqs1003 is OK: TCP OK - 0.005 second response time on port 9042 [14:56:29] lol, I love how "now()" returns a guid [14:56:32] how completely stupid [14:56:44] joal: well, that won't be hooked up through restbase though [14:57:00] ok, so now if you do select * from data where "_domain" = 'analytics.wikimedia.org' and project = 'en.wikipedia' and access = 'all-access' and agent = 'all-agents' and granularity = 'daily' and timestamp = '2015100100'; [14:57:06] you get two records instead of one [14:57:36] which keyspace? [14:58:33] ok found it [14:58:51] sorry, per-project [14:58:58] haha, now i don't know how to delete the old one [14:59:17] 'cause i have to specify a timeuuid and i don't know how [14:59:38] empty string [15:00:21] ottomata, hi! qq: are eventlogging validation errors written to a log file also or just to kafka? [15:00:21] doesn't work milimetric ? [15:00:49] this doesn't work: select * from data where "_domain" = 'analytics.wikimedia.org' and project = 'en.wikipedia' and access = 'all-access' and agent = 'all-agents' and granularity = 'daily' and timestamp = '2015100100' and "_tid" = ''; [15:01:07] Invalid STRING constant () for "_tid" of type timeuuid [15:01:15] mwarf [15:01:51] aha! [15:01:51] select * from data where "_domain" = 'analytics.wikimedia.org' and project = 'en.wikipedia' and access = 'all-access' and agent = 'all-agents' and granularity = 'daily' and timestamp = '2015100100' and "_tid" > dfcec280-6e95-11e5-89ab-55ce467c43aa; [15:01:57] <> doesn't work, but > does [15:02:06] wow, well done :) [15:02:59] also, delete doesn't work with that where clause :/ [15:03:22] milimetric: MANNN ! [15:03:27] What have Ivdone :( [15:03:49] it's ok, we'll figure this out. Also, cassandra is supposed to be simple, wtf [15:04:01] milimetric: timeuuid is not simple . [15:04:08] We hsould have avoided that [15:04:20] can we help it? It's part of restbase, right? [15:04:43] It's part of restbase cassandra module I think [15:05:22] that _tid column should just have a default [15:05:31] right [15:05:32] if you don't specify a value on insert, does it fill it in? [15:06:08] nope, says _tid is missing [15:06:46] zactly [15:07:42] tried that before .... [15:08:57] figured it out, deleted the two others and just inserted a new record [15:09:00] duh [15:09:01] :) [15:09:05] ok, so now to query! [15:09:28] How have you managed that ? [15:09:46] worked milimetric !!! [15:09:46] oh snap!!! [15:09:55] :) k, so that's the only problem [15:09:56] phew [15:10:00] right [15:10:12] I was dug in, expecting to find like 30 other nested ones [15:10:22] :D [15:10:33] milimetric: how have you manged to delete the two rows ? [15:10:35] sweet, so.... hm.... now lemme see, it says you can COPY the data out to a CSV file [15:10:42] oh, just don't specify the _tid [15:10:44] milimetric: correct [15:10:53] just delete from ... where [15:11:18] ok, that worked because _tid is not part of the partition key [15:11:20] makes sense [15:11:56] milimetric: retrying again to load data without setting _tid [15:12:14] joal: well, won't that take a lot longer than figuring out how to set the _tid to now() everywhere? [15:12:41] milimetric: both are needed actually [15:12:56] ...? [15:13:10] Well, I don't want to coninue pushing wrong data ! [15:13:16] oh! [15:13:32] you said "retrying" I thought you meant you were deleteing everything and starting over [15:14:04] why don't we hang out in the batcave :) [15:14:35] OMW [15:17:23] (Abandoned) Ottomata: [WIP] Add properties file for importing mediawiki data [analytics/refinery] - https://gerrit.wikimedia.org/r/244594 (https://phabricator.wikimedia.org/T113521) (owner: Madhuvishy) [15:18:54] (PS1) Ottomata: Removing camus/ properties files. This has been moved to puppet [analytics/refinery] - https://gerrit.wikimedia.org/r/244694 (https://phabricator.wikimedia.org/T115114) [15:19:37] !log moved camus property files out of refinery repository and into puppet. Camus properties now live on an27 at /etc/camus.d, and camus log files are in /var/log/camus [15:19:39] Logged the message at https://www.mediawiki.org/wiki/Analytics/Server_Admin_Log, Master [15:19:40] joal: ^ :) [15:19:50] awesome ottomata [15:19:52] Thanks [15:20:07] (CR) Ottomata: [C: 2 V: 2] Removing camus/ properties files. This has been moved to puppet [analytics/refinery] - https://gerrit.wikimedia.org/r/244694 (https://phabricator.wikimedia.org/T115114) (owner: Ottomata) [15:20:32] adding new camus jobs is much nicer now [15:20:43] https://gerrit.wikimedia.org/r/#/c/244601/2/manifests/role/analytics/refinery.pp [15:26:32] Analytics-Cluster, Analytics-Kanban, Patch-For-Review: Move camus properties out of refinery and into puppet [5 pts] - https://phabricator.wikimedia.org/T115114#1715380 (Ottomata) [15:26:57] Analytics-Backlog, Analytics-Cluster, Analytics-Kanban: logrotate camus logs on analytics1027 [3 pts] - https://phabricator.wikimedia.org/T110598#1715384 (Ottomata) [15:27:01] Analytics-Backlog, The-Wikipedia-Library: Implement Schema:ExternalLinkChange - https://phabricator.wikimedia.org/T115119#1715388 (Halfak) NEW [15:28:11] Analytics-Backlog, The-Wikipedia-Library: Implement Schema:ExternalLinkChange - https://phabricator.wikimedia.org/T115119#1715396 (Halfak) I **boldly** added this to the #Analytics-Backlog in hope that they might be able to pick up requests like this since the #The-Wikipedia-Library doesn't have the engin... [15:32:28] Analytics-Backlog: Add the schema name to the EL EventError topic - https://phabricator.wikimedia.org/T115121#1715415 (mforns) NEW [15:34:28] milimetric: http://stackoverflow.com/questions/23191933/cassandra-inserting-timeuuid-error [15:34:37] :( [15:37:08] Analytics-Kanban, Wikimedia-Logstash, Patch-For-Review: Make Logstash consume from Kafka:eventlogging_EventError {oryx} [8 pts] - https://phabricator.wikimedia.org/T113627#1715442 (mforns) [15:37:12] Analytics-Kanban, Wikimedia-Logstash, Patch-For-Review: Make Logstash consume from Kafka:eventlogging_EventError {Oryx} [8 pts] - https://phabricator.wikimedia.org/T113627#1715443 (ggellerman) [15:37:43] joal: and we're not allowed the datastax driver? [15:38:12] milimetric: we are, just realised that I have a function to generate an uuid [15:38:34] you read the question and got sad, but didn't read the answer :) [15:38:55] I did, but thought that it was only a python thing :) [15:39:02] But found the thing in java ! [15:39:05] milimetric: --^ [15:39:07] :) [15:39:51] that question you linked was java! :) [15:40:00] http://stackoverflow.com/a/23198388/180664 [15:41:26] Analytics-Cluster, Analytics-Kanban, operations, Monitoring, Patch-For-Review: Replace uses of monitoring::ganglia with monitoring::graphite_* [5 pts] - https://phabricator.wikimedia.org/T90642#1715469 (Ottomata) [15:44:49] milimetric: I feel silly now :( [15:45:08] I'll ensure stuff get's back in track before weekend :) [15:45:25] psh, whatsamatter, you can't have a meeting debug code, run tests, read SO answers, and talk to me at the same time? [15:45:31] weaaak! [15:45:31] :P [15:45:42] :D [15:48:48] milimetric: select * from "Test_Project"."data"; [16:06:06] Analytics-Kanban, Analytics-Wikimetrics: Wikimetrics' cohort page is returning 500 in production {dove} [2 pts] - https://phabricator.wikimedia.org/T114881#1715497 (kevinator) Open>Resolved [16:06:28] Analytics-Backlog, Privacy, Varnish: Connect Hadoop records of the same request coming via different channels - https://phabricator.wikimedia.org/T113817#1715499 (madhuvishy) [16:06:39] Analytics-Kanban, netops, operations, Patch-For-Review: Puppetize a server with a role that sets up Cassandra on Analytics machines [13 pts] {slug} - https://phabricator.wikimedia.org/T107056#1715503 (kevinator) Open>Resolved [16:08:13] Analytics-Kanban, RESTBase-API: create RESTBase endpoints [34 pts] {slug} - https://phabricator.wikimedia.org/T107053#1715506 (kevinator) Open>Resolved [16:08:47] Analytics-Kanban: Deploy the Analytics RESTBase {slug} [13 pts] - https://phabricator.wikimedia.org/T113991#1715508 (kevinator) Open>Resolved [16:08:52] (PS2) Madhuvishy: Fix inconsistent mobile uniques reports due to partial job runs [analytics/refinery] - https://gerrit.wikimedia.org/r/244604 (https://phabricator.wikimedia.org/T114406) [16:09:23] Analytics-Kanban, Analytics-Wikistats: Feed Wikistats traffic reports with aggregated hive data {lama} [8 pts] - https://phabricator.wikimedia.org/T114379#1715513 (kevinator) [16:09:24] Analytics-Kanban: Spike: understand wikistats enough to estimate replacing pageview data source {lama} [8 pts] - https://phabricator.wikimedia.org/T114660#1715512 (kevinator) Open>Resolved [16:10:35] Analytics-Backlog, Analytics-Cluster, Analytics-Kanban, Patch-For-Review: logrotate camus logs on analytics1027 [3 pts] - https://phabricator.wikimedia.org/T110598#1715519 (kevinator) Open>Resolved [16:10:35] (CR) Madhuvishy: Add libjars optional arg to Camus python wrapper script (1 comment) [analytics/refinery] - https://gerrit.wikimedia.org/r/244599 (owner: Madhuvishy) [16:11:53] Analytics-Cluster, Analytics-Kanban, Patch-For-Review: Move camus properties out of refinery and into puppet [5 pts] - https://phabricator.wikimedia.org/T115114#1715521 (kevinator) Open>Resolved [16:12:30] Analytics-Cluster, Analytics-Kanban, operations, Patch-For-Review: Fix active namenode monitoring so that ANY active namenode is an OK state. [8 pts] - https://phabricator.wikimedia.org/T89463#1715523 (kevinator) Open>Resolved [16:16:28] Analytics-Kanban: Gain permission to delete articles on wikitech and mediawiki (needed for doc cleanup) [3 pts] - https://phabricator.wikimedia.org/T114672#1715532 (kevinator) Open>Resolved [16:18:52] Analytics-Kanban, Wikimedia-Logstash, Patch-For-Review: Make Logstash consume from Kafka:eventlogging_EventError {Oryx} [8 pts] - https://phabricator.wikimedia.org/T113627#1715541 (kevinator) Open>Resolved [16:19:50] (PS2) Madhuvishy: Add libjars optional arg to Camus python wrapper script [analytics/refinery] - https://gerrit.wikimedia.org/r/244599 [16:20:18] Analytics-Kanban: Update camus-wmf to be deployed by maven (missing jars otherwise) {hawk} [8 pts] - https://phabricator.wikimedia.org/T114657#1715545 (kevinator) Open>Resolved [16:21:53] Analytics-Cluster, Analytics-Kanban, operations, Monitoring, Patch-For-Review: Replace uses of monitoring::ganglia with monitoring::graphite_* [5 pts] - https://phabricator.wikimedia.org/T90642#1715547 (kevinator) Open>Resolved [16:22:24] madhuvishy: how about [16:22:31] i just realized :D [16:22:33] changing [16:22:37] oh ok, heh [16:22:42] not sure to what... [16:22:44] was gonn ajust say [16:23:07] {3} ... .format( ... "-libjars " + libjars if libjars else '' ) [16:23:09] or seomthing like that [16:23:41] aah, i thought of doing something like what you did in puppet [16:25:41] Analytics-EventLogging, Analytics-Kanban: {stag} EventLogging on Kafka - https://phabricator.wikimedia.org/T102225#1715551 (kevinator) Open>Resolved a:kevinator This project & 2015-16 Fiscal Q1 goal is DONE as of Sept 30 2015! Goals page updated: https://www.mediawiki.org/wiki/Wikimedia_Engineeri... [16:26:52] (PS3) Madhuvishy: Add libjars optional arg to Camus python wrapper script [analytics/refinery] - https://gerrit.wikimedia.org/r/244599 [16:27:07] aye madhuvishy that is kinda like that [16:27:21] madhuvishy: that is fine [16:27:47] i would probably do it the way I mentioned in python, i can't do it like that in puppet because i can't do conditionals or variable reassignment so easily in puppet [16:27:51] but this way is totally fine [16:28:09] hmmm, not sure about those double quotes [16:28:15] "{3}" [16:28:25] not needed may be? [16:28:29] i think that would make the two opts be passed as a single arg to java [16:28:39] since {3} is now [16:28:43] milimetric: I have values for _tid [16:28:45] -libjars aaa,bbb,ccc [16:28:57] milimetric: BUT, they are the same (and supposed to be different) [16:29:21] milimetric: I assume we move forward with that anyway (field not used and all) [16:30:13] you're using UUIDs.timeBased() ? [16:30:20] yes I do [16:30:20] and that just keeps generating the same value? [16:30:22] :( [16:30:26] wth java [16:30:28] :) [16:30:37] * joal nod [16:30:44] select * from "local_group_default_T_pageviews_per_project".data where "_domain" = 'analytics.wikimedia.org' and project = 'en.wikipedia' and access = 'all-access' and agent = 'all-agents' and granularity = 'daily' limit 10; [16:30:59] has difference values [16:31:11] oh cool! [16:31:35] But select * from "Test_Project"."data"; as same values :( [16:31:46] wait so where do the different values come from? [16:31:50] So basically: We are not usre [16:32:07] milimetric: I get your go ? [16:32:21] sure... :) [16:32:27] ok cool :) [16:32:29] we'll fix it later, no problem [16:32:45] i mean, it'll still be a PK, because we're assuming the rest of the PK is unique [16:32:47] milimetric: will be hard to fix don't you think ? [16:32:48] so I don't see it causing a problem [16:32:58] Yeah, same for me [16:33:04] joal: also, it's not entirely meaningless, right, because each load job will have a different value [16:33:04] ok, I go for that :) [16:33:13] so if we have one bad load, we can justly quickly select all the records from it :) [16:33:39] good point [16:33:42] Ok, let's go [16:33:45] sweet! [16:33:48] dooooo itttt [16:33:59] * milimetric going to find some lunch, rob a bank, bbl [16:34:25] ja madhuvishy i think you should remove those quotes around {3} [16:34:29] aside from that +1! [16:34:30] :) [16:35:20] madhuvishy: Are you ok with the comments I made on your CRs ? [16:35:45] joal: which one? [16:35:58] sorry i pushed too many patches yesterday [16:36:02] :) [16:36:31] (PS4) Madhuvishy: Add libjars optional arg to Camus python wrapper script [analytics/refinery] - https://gerrit.wikimedia.org/r/244599 [16:37:03] I reviewed two, the camus module one and another, I can't remember :) [16:37:29] joal: the other one was the hive query parenthesis, I fixed that [16:37:34] camus module I haven't seen [16:37:36] ah, yes [16:39:36] joal: looking now [16:49:04] joal: just looking at this - https://gist.github.com/jobar/9c471d68cc9e04be3b9b [16:49:15] are these tests for the JSON one or binary one? [16:49:20] json [16:50:36] i might not have mentioned it, but thanks for all the work on camus! i didn't imagine it would be so involved [16:53:49] Analytics-Kanban, Wikimedia-Logstash, Patch-For-Review: Make Logstash consume from Kafka:eventlogging_EventError {Oryx} [8 pts] - https://phabricator.wikimedia.org/T113627#1715611 (bd808) Dashboard at https://logstash.wikimedia.org/#/dashboard/elasticsearch/eventlogging-errors [16:55:24]