[00:37:47] (03PS5) 10Milimetric: Implement Topic Selector [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/382636 (https://phabricator.wikimedia.org/T167676) [06:27:23] 10Analytics, 10DBA: Drop MoodBar tables from all wikis - https://phabricator.wikimedia.org/T153033#3674604 (10Marostegui) 05Open>03Resolved [08:17:12] 10Analytics, 10Operations, 10Ops-Access-Requests: analytics-privatedata-users access for Jeff Green - https://phabricator.wikimedia.org/T177602#3674732 (10elukey) a:03Jgreen @Dzahn I am pretty sure that Jeff needs access to webrequest data due to the fact that they maintain a kafka consumer in the fundrais... [08:33:44] I just restarted the wikidata job that failed, but afaics the issue seems to be [08:33:47] java.lang.SecurityException: class "javax.servlet.FilterRegistration"'s signer information does not match signer information of other classes in the same package [08:33:55] I cannot see any container log, only the above one for the appid [08:34:24] I have already seen this error popping up once in a while [08:44:10] Hi a-team [08:44:16] elukey: I have seen i as well [08:44:28] elukey: I wonder what it actually means :( [08:45:12] joal: o/ [08:45:59] joal: I've read briefly on stack overflow that it might be due to jars containing signatures in the manifest mixed with ones not carrying them (or maybe with different ones) [08:46:04] not sure if related though [08:46:11] because it doesn't seem to be consistent [08:46:15] it happens once in a while [08:46:33] elukey: the fact that it happens inconsistently is what makes me unhappy [08:49:29] joal: completely separate, but atm most of the datanode partitions on the analytics hadoop workers are ~90% used [08:49:48] elukey: Wow ... That is a bit unexpected [08:50:14] elukey: when we were discussing with ottomata and nuria_ about space, we were thinking there was more than 2 month left [08:50:20] well we are not deleting data recently, the space consumption went up a lot [08:50:30] we have ~200GB free now in hdfs [08:50:30] elukey: I'll have a quick look a check what we can drop [08:50:43] elukey: really ? [08:50:47] elukey: No even a Tb? [08:50:50] lemme check [08:51:13] https://grafana.wikimedia.org/dashboard/db/analytics-hadoop?panelId=25&fullscreen&orgId=1 [08:51:31] elukey: hdfs dfs -df -h [08:51:38] hdfs://analytics-hadoop 1.8 P 1.5 P 221.4 T 87% [08:51:52] sorry [08:51:52] Filesystem Size Used Available Use% [08:51:52] hdfs://analytics-hadoop 1.8 P 1.5 P 221.4 T 87% [08:51:59] Feels better :) [08:52:00] ahahah yes sorry I meant 200TB not GB [08:52:09] yeah, this makes a difference :) [08:52:13] I know I know [08:52:38] elukey: I'll chek space state [08:54:49] in the meantime, I am rebooting kafka jumbo hosts for kernel updates [08:56:56] k [08:59:18] 10Analytics-Cluster, 10Analytics-Kanban, 10monitoring, 10User-Elukey: Decide on casing convention for JMX metrics in Prometheus - https://phabricator.wikimedia.org/T177078#3674811 (10fgiunchedi) I've re-read the thread and I think I have a proposal to move things forward. The reason I've insisted on all `... [09:07:41] joal: good to merge https://gerrit.wikimedia.org/r/#/c/383332/ ? [09:12:12] elukey: not yet ! Code has not been deployed [09:13:06] joal: didn't Dan deployed the refinery yesterdaY? [09:13:19] elukey: I don't know, was away yesterday [09:13:36] elukey: but from what I have seen in hdfs, I don't think [09:13:53] elukey: oh, actually I'm wrong -- Code seems to be here :) [09:14:06] so good to go elukey [09:14:07] joal: you know everything, it is irrelevant that you were away :D [09:14:11] sorry for back and forth [09:14:16] :D [09:14:37] ah joal druid100[456] have been moved to the production network! [09:14:43] elukey: Yay ! [09:14:46] This is great [09:15:08] the firewall rules of the analytics vlan are updated, we can contact from hadoop the overlord port and the broker port [09:15:21] elukey: This is hugely great [09:15:32] elukey: Does that mean we could move forward with some testing? [09:16:49] (03CR) 10Joal: "Little glitch but not far to be ready for me :)" (033 comments) [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/382659 (https://phabricator.wikimedia.org/T175268) (owner: 10Fdans) [09:17:12] so there are some details to mention [09:17:54] 1) having LVS in front of what is now the druid analytics cluster (druid100[123]) is technically a bit difficult and will be postponed for the moment (so poor pivot will need to deal with the current config) [09:19:10] 2) we need to install the new druid cluster on druid100[456] and then open ports in Hadoop Ferm's rule to allow workers to get requests from druid100[456] [09:19:21] 3) finally put LVS in front of druid100[123] [09:19:57] --^ 3) in front of druid100[456] ?? [09:20:01] elukey: --^ [09:20:05] correct sorry [09:20:16] k, just to be sure I was following :) [09:20:30] 2) and 3) should be a matter of 1 or 2 days of work if nothing comes up [09:20:49] so I'd say that early next week you'll be able to properly test this [09:20:57] (or even after 2) only ) [09:21:11] ok, it'll allow us to finalize data-loading-job and AQS requester [09:21:22] That's great :P) [09:21:37] Thanks for moving this along with ottomata, elukey :) [09:23:07] yesterday it was a fun learning day, I had to poke with the network vlans and I finally got some interesting details :) [09:23:37] elukey: Are there any info you think I should know? [09:24:21] joal: all the details in https://phabricator.wikimedia.org/T177511, but probably nothing that directly affects your work (only if you are curious about networking config in our routers) [09:25:00] (03PS5) 10Fdans: Add stub of new contributing and content metrics [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/382659 (https://phabricator.wikimedia.org/T175268) [09:25:14] * joal reads [09:25:23] the TL;DR is that the analytics vlan is in reality four VLANs (one for each row) and the production network runs in other "clusters" of VLANs [09:25:33] for example, public and private vlans [09:25:53] (public == all the hosts with public IP, private the rest) [09:26:31] elukey: wow, this makes things a bit complicated :) [09:27:06] Thanks for the patch fdans :) [09:27:43] so we were unable to set up LVS for druid100[123] since the vlan on which the lvs hosts are running (public IIUC) is not allowed to contact the analytics vlan [09:27:45] (vlans) [09:28:07] so no health checks available, etc.. [09:33:53] That's what I had understood from ottomaa yes [09:34:47] * joal looks in fdans direction and hopes to get a smile back [09:35:57] joal: thank youuuuuuu for the review [09:36:05] (sorry, I'm the worst at IRC in the morning) [09:36:12] * joal feels better :) [09:36:45] fdans: I was not after a tahnk you, but wss wondering if me going for review before evewn a hello on irc was bad communication for you :) [09:37:06] 10Analytics-Kanban, 10Analytics-Wikistats, 10Patch-For-Review: Create Druid public cluster such AQS can query druid public data - https://phabricator.wikimedia.org/T176223#3674896 (10elukey) [09:37:08] 10Analytics-Kanban, 10User-Elukey: LVS for Druid - https://phabricator.wikimedia.org/T177511#3674895 (10elukey) 05Open>03Resolved [09:38:12] 10Analytics-Kanban, 10Analytics-Wikistats, 10Patch-For-Review: Create Druid public cluster such AQS can query druid public data - https://phabricator.wikimedia.org/T176223#3617909 (10elukey) [09:38:34] 10Analytics-Kanban, 10Analytics-Wikistats, 10Patch-For-Review: Create Druid public cluster such AQS can query druid public data - https://phabricator.wikimedia.org/T176223#3617909 (10elukey) [09:38:53] joal: noooo I think code reviews are way more async, but I should be a bit more communicative via IRC [09:39:19] also, because I get the notifications on my phone, a lot of times I don't even have it open in my computer, and that's not cool [09:39:22] no worries fdans, I'm usually not that good on irc either :) [09:39:42] As long as everything is good, no reason to make it bad [10:38:23] 10Analytics-Cluster, 10Analytics-Kanban, 10monitoring, 10User-Elukey: Decide on casing convention for JMX metrics in Prometheus - https://phabricator.wikimedia.org/T177078#3675008 (10elukey) +1, looks good to me. As said above I don't particularly like mixing lowercase with camelcase but I find your soluti... [10:40:32] I am upset by fdans behavior [10:40:36] just saying :P [10:41:44] WELL SO AM I elukey [10:43:28] don't need to show this passive aggressive irc behavior fdans [10:44:02] :P [10:46:18] elukey: https://img.buzzfeed.com/buzzfeed-static/static/2014-06/11/16/enhanced/webdr07/anigif_enhanced-5784-1402517772-1.gif [10:47:08] hahahah wonderful [10:49:21] there are some interesting blog posts in https://hacks.mozilla.org/2017/08/inside-a-super-fast-css-engine-quantum-css-aka-stylo/ [10:49:35] not sure if you guys are following hacks.m.o, discovered this morning via twitter [10:50:38] and also learned about https://servo.org/ [10:54:55] I love technical explanations like these [10:56:24] * elukey lunch! [11:08:17] FYI, I'm restarting apache on bohrium to pick up a security update [11:37:18] ack! [11:42:21] 10Analytics, 10Analytics-EventLogging, 10DBA: Potentially decrease db1046's InnoDB buffer pool - https://phabricator.wikimedia.org/T125829#1998057 (10Marostegui) Is this still worth keeping considering the fact that we are going to replace this host soon anyways? [11:43:23] 10Analytics, 10Analytics-EventLogging, 10DBA: Potentially decrease db1046's InnoDB buffer pool - https://phabricator.wikimedia.org/T125829#1998057 (10elukey) I would not touch db1046 from now if possible :) [11:56:04] joal: cp3007 (cache misc esams) is running varnish 5.1 [11:56:27] I am checking webrequest logs from stat1004, all good so far, buuut I wanted to let you know :) [11:56:37] Cool elukey ! [11:56:44] elukey: no huge change in VK needed? [11:57:55] joal: for 5.1 yes, for 5.2 they rewrote the whole shared memory logging (AGAIN) and I'll need to make changes to vk [11:57:59] but next quarter [11:58:27] Ah, was mistaking minor version :) [11:58:30] ok [12:26:46] 10Analytics, 10User-Elukey: Refactor kafka_config.rb and and kafka_cluster_name.rb in puppet to avoid explicit hiera calls - https://phabricator.wikimedia.org/T177927#3675238 (10elukey) [12:27:21] 10Analytics, 10Operations, 10User-Elukey: Refactor kafka_config.rb and and kafka_cluster_name.rb in puppet to avoid explicit hiera calls - https://phabricator.wikimedia.org/T177927#3675250 (10elukey) [12:32:36] urandom: https://github.com/prometheus/jmx_exporter/blob/master/collector/src/main/java/io/prometheus/jmx/JmxCollector.java#L397-L403 [12:32:39] love it [12:35:41] 10Analytics-Cluster, 10Analytics-Kanban, 10monitoring, 10User-Elukey: Decide on casing convention for JMX metrics in Prometheus - https://phabricator.wikimedia.org/T177078#3675289 (10Gehel) All good for me! We can always revisit later if we find real life issues... [13:03:45] 10Analytics, 10Analytics-EventLogging, 10DBA: Potentially decrease db1046's InnoDB buffer pool - https://phabricator.wikimedia.org/T125829#3675361 (10jcrespo) 05Open>03declined [13:14:14] mooorning! :) [13:16:45] o/ [13:16:51] \o [13:20:52] elukey: [13:20:53] http://home.apache.org/~guozhang/kafka-1.0.0-rc0/RELEASE_NOTES.html [13:20:54] has [13:21:03] https://issues.apache.org/jira/browse/KAFKA-4763 [13:21:09] Handle disk failure for JBOD (KIP-112) [13:21:16] woa 1.0?? [13:22:19] ya [13:22:22] crzy [13:22:29] just RC now [13:22:35] 0.11.0.1 not even in confluent dist yet! [13:22:41] so, probably will still be a couple of months [13:23:04] nice to see that we are crossing the 1.0 mark [13:23:19] I mean, it was expected due to how widespread kafka is now [13:24:40] ottomata: if you have patience would you mind if I attempt to code the puppet patch to bring up the druid public cluster? [13:25:19] elukey: yup [13:25:20] 10Analytics, 10Operations, 10Traffic, 10User-Elukey: Refactor kafka_config.rb and and kafka_cluster_name.rb in puppet to avoid explicit hiera calls - https://phabricator.wikimedia.org/T177927#3675554 (10ema) p:05Triage>03Normal [13:25:30] elukey: was about to work on that :)have this partly https://gerrit.wikimedia.org/r/#/c/380804/ [13:25:46] elukey: you closed the LVS for druid one, does that mean LVS is set up? [13:25:52] elukey: didn't go looking for the code, but i did notice that it picks up changes automagically [13:27:19] ottomata: ahh okok if you have already the code I'll let you do it nevermind :) The ticket that I closed (I thought) it was only to decide how to do load balancing for druid public/private, and we decided to postpone it for druid private no? [13:27:54] I can create a new subtask more specific to public druid [13:28:35] or reopen that one [13:28:37] as you wish [13:28:55] (03CR) 10Ottomata: "Ooook, but think about when we start auto importing eventlogging data into druid. We're not going to use hyphens then, no? Then we'll ha" [analytics/refinery] - 10https://gerrit.wikimedia.org/r/383172 (https://phabricator.wikimedia.org/T175162) (owner: 10Joal) [13:29:36] elukey: either way :) we can use the public druid one for the LVS step too [13:32:42] done! [13:32:55] 10Analytics-Kanban, 10Analytics-Wikistats, 10Patch-For-Review: Create Druid public cluster such AQS can query druid public data - https://phabricator.wikimedia.org/T176223#3675602 (10elukey) [13:33:28] 10Analytics-Cluster, 10Analytics-Kanban, 10monitoring, 10Patch-For-Review, 10User-Elukey: Decide on casing convention for JMX metrics in Prometheus - https://phabricator.wikimedia.org/T177078#3675608 (10Ottomata) Hahahah my grumpiness paid off? Sounds good yall :) [13:33:48] 10Analytics-Cluster, 10Analytics-Kanban, 10monitoring, 10Patch-For-Review, 10User-Elukey: Decide on casing convention for JMX metrics in Prometheus - https://phabricator.wikimedia.org/T177078#3675609 (10Ottomata) You are teaching me a bad lesson, you just wait until the next naming bikeshed, I'll never s... [13:34:53] ottomata: the prometheus jmx exporter finds out automatically when the mtime of the file is changed and reloads itself, without any service restart required [13:34:57] love it [13:36:31] 10Analytics, 10Operations, 10Traffic, 10User-Elukey: Refactor kafka_config.rb and and kafka_cluster_name.rb in puppet to avoid explicit hiera calls - https://phabricator.wikimedia.org/T177927#3675614 (10Ottomata) This would really only require passing `kafka_clusters` as well as `kafka_cluster_name` to the... [13:36:37] oh nice! [13:36:38] awesome [13:37:52] 10Analytics-Kanban, 10Analytics-Wikistats, 10Patch-For-Review: Create Druid public cluster such AQS can query druid public data - https://phabricator.wikimedia.org/T176223#3675620 (10elukey) [13:53:37] hey all [13:59:46] hyaaa [14:01:39] hellooo [14:04:54] joal / fdans what's left on https://phabricator.wikimedia.org/T175268 (the metric stubs)? Want me to take over the review? [14:19:55] elukey: q [14:19:59] about private properties for druid [14:20:19] i notied that ./hieradata/eqiad/druid.yaml still has the profile::druid::common::properties: in it [14:20:25] bu tit is also in ./hieradata/role/eqiad/druid/analytics/worker.yaml [14:20:31] we can probalby remove the one in eqiad/druid.yaml, ya? [14:20:33] also [14:21:11] I think so yes [14:21:15] i've put the druid worker.yaml hiera in common/ not eqiad/ since if we did set up corresponding druid cluster roles in codfw, they'd likely have all the same settings...maybe... [14:21:28] so, can I move ./hieradata/role/eqiad/druid/analytics/worker.yaml in private to role/common? [14:22:18] oh yes I stopped touching those things after the last refactoring as promised :D [14:22:28] please go ahead and change those settings [14:22:30] haha ok [14:22:49] i'm goign to keep the private_properties hiera you added reluctantly, only to not have to face the wrath of hiera_hash + _joe_ :p [14:23:23] milimetric: if joal okays the latest changes I think we're good to go! [14:24:03] fdans: that's what I mean, if he's too busy I can take a look at his original review and the changes, lemme know [14:31:35] haha uhh elukey what is the frontend puppetmaster? [14:31:41] i thought it was puppetmaster1001 [14:31:47] i was about to commit to /etc/puppet/private [14:31:57] oh! maybe i just need to use /srv/private [14:32:18] puppetmaster1001 is one of the frontends [14:32:25] ottomata: yes [14:32:40] /srv/private is the right one :) [14:38:48] ottomata,joal - ops sync? [14:41:06] OO [14:41:06] ok [14:43:48] 10Analytics, 10Analytics-Wikistats: Label breakdown categories and other non-obvious concepts - https://phabricator.wikimedia.org/T177950#3675829 (10Milimetric) [14:46:48] 10Analytics, 10Analytics-General-or-Unknown: Add flagged revision status statistics to Wikistats 2.0 - https://phabricator.wikimedia.org/T177951#3675852 (10Milimetric) [14:49:58] (03PS40) 10Mforns: JsonRefine: refine arbitrary JSON datasets into Parquet backed hive tables [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/346291 (https://phabricator.wikimedia.org/T161924) (owner: 10Joal) [14:50:05] (03CR) 10jerkins-bot: [V: 04-1] JsonRefine: refine arbitrary JSON datasets into Parquet backed hive tables [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/346291 (https://phabricator.wikimedia.org/T161924) (owner: 10Joal) [14:50:48] (03CR) 10Mforns: JsonRefine: refine arbitrary JSON datasets into Parquet backed hive tables (037 comments) [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/346291 (https://phabricator.wikimedia.org/T161924) (owner: 10Joal) [14:53:49] (03PS1) 10Fdans: Maintain active breakdowns when time range is changed [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/383589 (https://phabricator.wikimedia.org/T177646) [14:55:14] oops, completely forgot about ops-sync ottomata/elukey [14:57:28] (03CR) 10Fdans: [V: 032 C: 032] "I like the solution of adding the enabled property to each question. I'm a bit weary of getting the metric key by kebab-casing the name, b" [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/382636 (https://phabricator.wikimedia.org/T167676) (owner: 10Milimetric) [15:08:40] (03CR) 10Ottomata: JsonRefine: refine arbitrary JSON datasets into Parquet backed hive tables (033 comments) [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/346291 (https://phabricator.wikimedia.org/T161924) (owner: 10Joal) [15:27:18] (03PS1) 10Fdans: Add "report a bug" link to the footer [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/383597 (https://phabricator.wikimedia.org/T177642) [15:28:16] nuria_: proceeding with https://phabricator.wikimedia.org/T171629#3674250 then, I'll drop those tables from db1046/47/store1002 [15:28:24] (el_sync stopped, etc.. [15:29:47] elukey: okeis [15:29:59] elukey: also purging right? [15:30:05] elukey: needs to be stop i mean [15:30:10] (03PS2) 10Joal: Fix druid datasources for proj-family uniques jobs [analytics/refinery] - 10https://gerrit.wikimedia.org/r/383172 (https://phabricator.wikimedia.org/T175162) [15:30:23] elukey: what was the convo with alex about special network constant hosts [15:31:01] (03PS8) 10Joal: Update mediawiki-history-reduced oozie job [analytics/refinery] - 10https://gerrit.wikimedia.org/r/379000 (https://phabricator.wikimedia.org/T174174) [15:31:25] milimetric: do you have a minute for a smalltalk on editor metric? [15:31:39] nuria_: purging is stopped since I was waiting to drop the huge tables that cause a looong wait time (like Edit_*_* etc..) [15:33:11] ottomata: so basically he explicitly asked me not to do it anymore since he wants to clean up all those ips in there and force people to use hiera.. I guess that we could get an exception for this use case but better to let him know in my opinion [15:33:58] elukey: the only other table i was going to move is PagecontentSavecomplete but i had not plan to move any other ones [15:34:01] mforns: whenever you have time can you review https://gerrit.wikimedia.org/r/#/c/383185/ ? [15:34:10] elukey, sure [15:35:16] nuria_: the ones that I am taking about are the following, probably in the task that got closed by dbas for dbstore1002's space exhaustion [15:35:19] MobileWikiAppToCInteraction_10375484_15423246 140.57 [15:35:22] Edit_11448630 137.02 [15:35:24] Edit_13457736_15423246 130.85 [15:35:55] it is not a big deal if we keep them [15:36:04] but it will take a long time to make them clean [15:38:03] elukey: i can move those but we will need to purge them just the same and it is easier to purge data in mysql than avro, the tables i have moved thus far have public data [15:38:15] elukey: so purging is not needed (except 1) [15:38:27] elukey, done [15:38:37] elukey: that might be true of edit table too , let me see [15:38:51] (03CR) 10Ottomata: [C: 031] Fix druid datasources for proj-family uniques jobs [analytics/refinery] - 10https://gerrit.wikimedia.org/r/383172 (https://phabricator.wikimedia.org/T175162) (owner: 10Joal) [15:39:11] elukey: also: https://gerrit.wikimedia.org/r/#/c/383601/ [15:39:16] nuria_: quick question on druid names [15:39:52] nuria_: We have the per-domain monthly datasource that is not to be used in Pivot (because month is not available as a UI group) [15:40:06] joal: right [15:40:07] Do you htink I can rename that one [15:40:20] joal: yes, right? as there is no bookmarks to it [15:40:30] I agree, just wanted to triple check [15:41:23] There'll be 4 left with hyphens [15:42:13] (03PS3) 10Joal: Fix druid datasources for uniques jobs [analytics/refinery] - 10https://gerrit.wikimedia.org/r/383172 (https://phabricator.wikimedia.org/T175162) [15:46:43] elukey: let me know, if we feel we must we can import those three tables, if we can do without we can re-start purging, i looked at whitelist and purging for those is easy [15:46:45] hm elukey any idea why i might be getting this? https://puppet-compiler.wmflabs.org/compiler02/8286/druid1004.eqiad.wmnet/change.druid1004.eqiad.wmnet.err [15:46:51] Could not find data item profile::druid::common::private_properties [15:47:01] elukey: easy meaning it can be done while importing as it is just dropping columns [15:47:06] i have it checked into private repo at hieradata/role/common/druid/public/worker.yaml [15:48:15] ottomata: ah yes the pcc takes the stuff from the fake private labs repo [15:48:20] elukey: can we also merge whitelist changes? [15:48:35] nuria_: just dropped the tables, merging it in a bit [15:48:41] thanks elukey [15:48:44] elukey: so I need to add it there? [15:48:47] how/can I do that? [15:48:48] elukey: okeis [15:48:54] ottomata: do we already know what will be the LVS name for fruif in prod? [15:49:13] joal: it should be druid-public.svc.analytics.eqiad.wmnet (i htink) with broker port [15:49:13] so [15:49:20] druid-public.svc.analytics.eqiad.wmnet:8082 (right?) [15:49:43] sorry [15:49:43] no [15:49:45] ottomata: you need to clone https://gerrit.wikimedia.org/r/#/admin/projects/labs/private, move the hiera config in there, review, merge and pcc will be happy [15:49:47] ottomata: This is for querying from AQS - Same name with different port for oozie to launch indexation? [15:49:52] druid-public.svc.eqiad.wmnet [15:49:53] no analytics [15:50:01] hmMMm [15:50:02] makes sense ottomata [15:50:09] interesting, hadn't thought of that, but yes, that's how I planned [15:50:21] so i need to add LVS svc for indexing on coordinator port? [15:50:25] (or overlord?) [15:50:34] I means any hadoop worker can contact druid on port 8090 [15:50:37] nope, it will be the same domain but different port [15:50:39] thank you [15:53:27] elukey: do i have to puppet-merge fake it somewhere? [15:53:31] https://gerrit.wikimedia.org/r/#/c/383602/ [15:53:44] nono just +2 +2 and you are done [15:53:59] now pcc should be happier [15:54:03] mforns: Heya - do you have a minute about endpoints names again? [15:54:12] joal, yes! [15:54:16] cave? [15:54:20] Yes ! [15:54:23] omw [16:00:03] elukey: Edit_11448630 is a live table so it cannot be imported, i could do the other two if needed [16:00:20] elukey: , I pinged alex, but i think we should go ahead with this one. that kind of stuff can be refactored later...and we aren't adding any new IPs :p [16:00:22] OH [16:00:27] i need to change the IPs though, eh? [16:00:29] oh yaaaa [16:01:20] 10Analytics-Kanban: Archive tables to hadoop: MobileWikiAppToCInteraction_10375484_15423246 and Edit_13457736_15423246 - https://phabricator.wikimedia.org/T177960#3676121 (10Nuria) [16:01:32] ottomata: yep :D [16:03:40] nuria_: only if you have time and patience, otherwise it is fine anyway.. archiving is not the most fun thing to do :) [16:03:50] ok elukey ready for second review [16:03:50] https://gerrit.wikimedia.org/r/#/c/380804/ [16:03:54] added ferm rules to hadoop nodes too [16:06:23] joal, yea man, I think net and absolute are really cool [16:06:32] nice mforns :) [16:08:57] ottomata: qq - is profile::zookeeper::firewall::srange: '(($DRUID_ANALYTICS_HOSTS $ANALYTICS_NETWORKS))' a bit redundant? I mean, the DRUID_ANALYTICS_HOSTS are in ANALYTICS_NETWORKS no? [16:13:29] going to write some notes [16:13:43] yeahhhhhhh [16:13:45] i thought about that [16:13:47] what do you think? [16:13:50] think we shoudl just leave it out? [16:14:04] DRUID_ANALYTICS_HOSTS? [16:15:08] elukey: ^ [16:20:40] ottomata: writing some silly comments sto the cr, so you'll know all my thoughts [16:20:43] :D [16:20:47] :D [16:22:21] done! [16:30:36] 10Analytics, 10Operations, 10Ops-Access-Requests, 10Patch-For-Review: analytics-privatedata-users access for Jeff Green - https://phabricator.wikimedia.org/T177602#3676231 (10Jgreen) 05Open>03Resolved We concluded analytics-privatedata-user makes sense, so I can use hive to come up with hourly hit coun... [16:57:34] * elukey off! [16:57:51] ottomata: will try to check in later on, gotta go now! [16:57:53] * elukey afk! [16:59:12] 10Analytics, 10Analytics-Wikistats: Beta Release: Resiliency, Rollback and Deployment of Data - https://phabricator.wikimedia.org/T177965#3676320 (10Nuria) [17:03:54] elukey: agree with all your coments will ipmlement [17:05:19] (03PS41) 10Mforns: JsonRefine: refine arbitrary JSON datasets into Parquet backed hive tables [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/346291 (https://phabricator.wikimedia.org/T161924) (owner: 10Joal) [17:05:25] (03CR) 10jerkins-bot: [V: 04-1] JsonRefine: refine arbitrary JSON datasets into Parquet backed hive tables [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/346291 (https://phabricator.wikimedia.org/T161924) (owner: 10Joal) [17:05:47] (03CR) 10Mforns: JsonRefine: refine arbitrary JSON datasets into Parquet backed hive tables (033 comments) [analytics/refinery/source] - 10https://gerrit.wikimedia.org/r/346291 (https://phabricator.wikimedia.org/T161924) (owner: 10Joal) [17:06:16] joal, finished the changes to JsonRefine, it's ready for review [17:09:31] wikimedia/mediawiki-extensions-EventLogging#703 (wmf/1.31.0-wmf.3 - 2cecfd7 : Timo Tijhof): The build has errored. [17:09:31] Change view : https://github.com/wikimedia/mediawiki-extensions-EventLogging/compare/wmf/1.31.0-wmf.3 [17:09:31] Build details : https://travis-ci.org/wikimedia/mediawiki-extensions-EventLogging/builds/286620217 [17:31:02] 10Analytics, 10Analytics-EventLogging, 10Cognate, 10Collaboration-Team-Triage, and 15 others: Possible WMF deployed extension PHP 7 issues - https://phabricator.wikimedia.org/T173850#3676505 (10Smalyshev) [17:31:13] 10Analytics, 10Analytics-EventLogging, 10Cognate, 10Collaboration-Team-Triage, and 15 others: Possible WMF deployed extension PHP 7 issues - https://phabricator.wikimedia.org/T173850#3541977 (10Smalyshev) [17:31:27] 10Analytics, 10Analytics-EventLogging, 10Cognate, 10Collaboration-Team-Triage, and 15 others: Possible WMF deployed extension PHP 7 issues - https://phabricator.wikimedia.org/T173850#3541977 (10Smalyshev) CirrusSearch one seems to be false positive. [17:36:14] 10Analytics, 10Analytics-General-or-Unknown: Add flagged revision status statistics to Wikistats 2.0 - https://phabricator.wikimedia.org/T177951#3676525 (10Zache) Thank you for remembering this and it is important . [17:40:46] 10Analytics, 10Analytics-EventLogging, 10Cognate, 10Collaboration-Team-Triage, and 15 others: Possible WMF deployed extension PHP 7 issues - https://phabricator.wikimedia.org/T173850#3676527 (10Reedy) Many are likely to be false positives, and that's not too much of an issue as long as people have looked o... [17:44:19] 10Analytics-Kanban: Add monthly unique devices dataset to Druid - https://phabricator.wikimedia.org/T163327#3676532 (10Nuria) Ping @JKatzWMF this dataset is not available on pivot but will be available in superset cc @JMinor [17:44:25] 10Analytics-Kanban: Add monthly unique devices dataset to Druid - https://phabricator.wikimedia.org/T163327#3676534 (10Nuria) 05Open>03Resolved [17:44:40] 10Analytics-Kanban, 10Patch-For-Review: Chose how to deal with "Infinity" value for Banners - https://phabricator.wikimedia.org/T175248#3676535 (10Nuria) [17:44:47] 10Analytics-Kanban, 10Patch-For-Review: Chose how to deal with "Infinity" value for Banners - https://phabricator.wikimedia.org/T175248#3587723 (10Nuria) 05Open>03Resolved [17:45:01] 10Analytics-Kanban, 10Analytics-Wikistats: Wikistats unique devices metrics needs some copy that says "monthly" - https://phabricator.wikimedia.org/T176240#3676537 (10Nuria) 05Open>03Resolved [17:45:20] 10Analytics-Kanban, 10Analytics-Wikistats, 10Patch-For-Review: Pageview retrieval does not work if one of the fails requests - https://phabricator.wikimedia.org/T176261#3676538 (10Nuria) 05Open>03Resolved [17:45:40] 10Analytics-Kanban, 10Analytics-Wikistats, 10Patch-For-Review: Add top articles by pageviews metric - https://phabricator.wikimedia.org/T175266#3676539 (10Nuria) 05Open>03Resolved [17:46:01] 10Analytics-Cluster, 10Analytics-Kanban, 10Patch-For-Review: rsync-published-datasets cron should not launch multiple rsync processes - https://phabricator.wikimedia.org/T174756#3676540 (10Nuria) 05Open>03Resolved [17:48:15] ottomata: Forgot to tell you: New superset UI is better, thanks for deploying (and data was not lost AFAIK)! [17:48:34] Nice! [17:48:35] great [17:57:33] (03CR) 10Nuria: [V: 032 C: 032] Fix druid datasources for uniques jobs [analytics/refinery] - 10https://gerrit.wikimedia.org/r/383172 (https://phabricator.wikimedia.org/T175162) (owner: 10Joal) [18:00:47] 10Analytics, 10Proton, 10Readers-Web-Backlog, 10Patch-For-Review, 10Readers-Web-Kanban-Board: Implement Schema:Print purging strategy - https://phabricator.wikimedia.org/T175395#3676556 (10ovasileva) @Tbayer - I'm actually curious to see what the split will be between desktop and mobile users for printin... [18:06:04] (03CR) 10Mforns: "Code looks good! (except typo)" (031 comment) [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/383597 (https://phabricator.wikimedia.org/T177642) (owner: 10Fdans) [18:11:31] (03CR) 10Mforns: "Oh! BTW, the link magic is awesome, the task form already has the wikistats and analytics tags!" [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/383597 (https://phabricator.wikimedia.org/T177642) (owner: 10Fdans) [18:14:25] 10Analytics, 10Analytics-Cluster, 10Operations, 10Research-management: GPU upgrade for stats machine - https://phabricator.wikimedia.org/T148843#3676603 (10dr0ptp4kt) An update: we looked into this. In short, we got stuck at the point of drivers. It appears that there //may// be a way to pure OpenCL-orient... [18:14:28] ottomata, one question about the cron job for Popups refine: as the JsonRefine code is not merged yet, where should that code live? Like.. in a dir under my home folder? [18:18:12] HMMM, let's see what joal thinks, but i think if you get joseph's comments resolved and he approves, we can merge. we can do spark 2 + other TODOs + productionization in another patch [18:18:19] otherwise, ya i guess in your homedir [18:24:53] ottomata, ok! [18:25:07] nuria_: will you present program 10 for QR? [18:25:19] lzia: on meeting can talk in abit [18:25:28] k, nuria_. :) [19:27:04] mforns, ottomata: Will review the scala patch tomorrow [19:32:12] leila: Hello ! [19:32:23] ow hi joal [19:32:49] leila: just realized I had not answered your question on priority queue :) [19:33:21] leila: Not secret code needed :) We relly on our users trust ;) [19:33:30] joal: ;) [19:33:57] leila: yarn application --movetoqueue APP_ID --queue priority [19:34:16] leila: Where APP_ID is, well, you app id :) [19:34:17] joal: thanks. I'll put it in the email as well, for archive happiness. thanks. [21:49:31] (03PS9) 10Joal: Update mediawiki-history-reduced oozie job [analytics/refinery] - 10https://gerrit.wikimedia.org/r/379000 (https://phabricator.wikimedia.org/T174174) [21:49:33] (03PS12) 10Joal: Add mediawiki-history-metrics endpoints [analytics/aqs] - 10https://gerrit.wikimedia.org/r/379227 (https://phabricator.wikimedia.org/T175805) [22:23:56] 10Analytics, 10ChangeProp, 10EventBus, 10MediaWiki-JobQueue, 10Services (doing): Generalized rate limiting, deduplication, and job scheduling module - https://phabricator.wikimedia.org/T173447#3677866 (10GWicke) Basic implementation: https://github.com/wikimedia/budgeteer. There is a basic integration o... [22:25:24] 10Analytics, 10ChangeProp, 10EventBus, 10MediaWiki-JobQueue, 10Services (doing): Generalized rate limiting, deduplication, and job scheduling module - https://phabricator.wikimedia.org/T173447#3677870 (10GWicke) [22:29:56] 10Analytics, 10ChangeProp, 10EventBus, 10MediaWiki-JobQueue, 10Services (doing): Generalized rate limiting, deduplication, and job scheduling module - https://phabricator.wikimedia.org/T173447#3677899 (10GWicke) a:05GWicke>03None [22:50:49] 10Analytics-Kanban, 10Discovery, 10Operations, 10Discovery-Analysis (Current work), and 2 others: Can't install R package Boom (& bsts) on stat1002 (but can on stat1003) - https://phabricator.wikimedia.org/T147682#3677987 (10debt) [22:52:35] 10Analytics, 10Analytics-Wikistats: Wikistats: support annotations - https://phabricator.wikimedia.org/T178015#3678012 (10Milimetric) [22:59:40] (03PS1) 10Milimetric: Exclude bots from pageviews [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/383725 [23:09:59] 10Analytics, 10Analytics-Wikistats: Make Wikistats data easily embedable on-wiki - https://phabricator.wikimedia.org/T178016#3678030 (10Milimetric) [23:12:31] 10Analytics, 10Analytics-Wikistats: Beta Release: Wikistats: support annotations - https://phabricator.wikimedia.org/T178015#3678046 (10Nuria) [23:13:30] 10Analytics, 10Analytics-Wikistats: Add edit/upload distinction to mediawiki history pipeline - https://phabricator.wikimedia.org/T178017#3678050 (10Milimetric) [23:13:34] 10Analytics, 10Analytics-Wikistats: Beta Release: Wikistats: support annotations - https://phabricator.wikimedia.org/T178015#3677997 (10Nuria) Annotation functionality for graphs can be ported from dashiki [23:17:51] (03PS1) 10Milimetric: Tweak order and alignment of lists for readability [analytics/wikistats2] - 10https://gerrit.wikimedia.org/r/383728 [23:19:37] 10Analytics, 10Analytics-Wikistats: Consider adding breadcrumbs to Wikistats 2 - https://phabricator.wikimedia.org/T178018#3678081 (10Milimetric) [23:22:49] 10Analytics, 10Analytics-Wikistats: Add overall ORES scores to Wikistats - https://phabricator.wikimedia.org/T178019#3678099 (10Milimetric) [23:36:01] nuria_: dashiki annotations are on top of dygraphs, but we can definitely do it the same way, like have them on-wiki. That said, we still haven't set up the mediawiki-Dashiki extension to properly handle non-dashboard pages (like annotation pages)