[01:37:48] Analytics, Analytics-Cluster, Fundraising Tech Backlog, Fundraising-Backlog, operations: Verify kafkatee use for fundraising logs on erbium - https://phabricator.wikimedia.org/T97676#1729945 (ellery) @awight Thanks for double checking, I understand the difference. See this notebook for an ana... [05:23:08] (PS13) Madhuvishy: [WIP] Add cassandra load job for pageview API [analytics/refinery] - https://gerrit.wikimedia.org/r/236224 (owner: Joal) [09:12:51] a-team good morning! [09:46:36] joal, yt? [11:05:38] Hi mforns [11:05:43] Whssup ? [11:05:43] hi joal :] [11:05:59] mmm I was having some problems with hive query [11:06:08] but it seems to be ok now [11:06:15] although [11:06:19] one question: [11:06:35] is it ok to set hive.mapred.mode=nonstrict? [11:06:56] because otherwise my query does not work [11:07:02] IIRC, this will allow you not to set limit when using order, is that correct ? [11:07:26] it also allows to make the cartesian product in a join [11:07:36] which in my case is not expensive [11:07:51] because one of the two joined tables has only one record [11:08:09] hmm [11:08:28] if you want, I can batcave you the query and maybe you have a better idea [11:09:04] PLease send the query, I'll collect my thoughts in the mean time :) [11:09:16] ok [11:09:51] mmm, pastebin is slow.. [11:10:59] joal, http://pastebin.com/Tg7CFyMK [11:11:39] the thing I found tricky is to compute the percentage [11:11:49] Right mforns [11:11:52] Get it :) [11:12:00] PROBLEM - Difference between raw and validated EventLogging overall message rates on graphite1001 is CRITICAL: CRITICAL: 40.00% of data above the critical threshold [30.0] [11:12:10] I'll show you how I'd do that :) [11:12:15] ok, cool [11:12:24] homf, EL first ? [11:12:46] let me look quickly [11:13:41] RECOVERY - Difference between raw and validated EventLogging overall message rates on graphite1001 is OK: OK: Less than 25.00% above the threshold [20.0] [11:13:52] mmm [11:14:15] seems the false positives are back? [11:14:30] looks like so indeed ! [11:14:42] ok, I think we can continue [11:15:06] mforns: cave ? [11:16:26] sure! [11:16:33] joal, ^ [11:16:43] :) [11:39:10] hi, I'm trying to setup hadoop kafka to test the new mediawiki monolog avro formatter [11:39:35] is there a role that installs kafka in vagrant? [11:49:58] thanks joal! [11:58:11] Analytics-Cluster, Analytics-Kanban, Easy: PM sees reports on browsers (Weekly or Daily) [8 pts] - https://phabricator.wikimedia.org/T88504#1730469 (mforns) A couple questions: - Do we want the report to be broken down by project? - Do we want to filter by project and show Wikipedia only? Or English Wi... [12:10:59] (CR) Joal: Archive hourly pageviews by article in wsc format (1 comment) [analytics/refinery] - https://gerrit.wikimedia.org/r/246149 (https://phabricator.wikimedia.org/T114379) (owner: Milimetric) [12:38:12] Analytics-Tech-community-metrics, Possible-Tech-Projects: Improving MediaWikiAnalysis - https://phabricator.wikimedia.org/T89135#1730509 (NiharikaKohli) @tmalhotra, this is another project you can look at for contributing. [12:43:51] Analytics-Tech-community-metrics, Possible-Tech-Projects: Improving MediaWikiAnalysis - https://phabricator.wikimedia.org/T89135#1730531 (ashitaprasad) @jgbarah Thanks for the reply adn giving me a head start. I will explore more about the project and select a microtask and inform you soon. [14:21:34] PROBLEM - Difference between raw and validated EventLogging overall message rates on graphite1001 is CRITICAL: CRITICAL: 26.67% of data above the critical threshold [30.0] [14:23:15] RECOVERY - Difference between raw and validated EventLogging overall message rates on graphite1001 is OK: OK: Less than 25.00% above the threshold [20.0] [14:32:00] joal: hi [14:32:06] hey mforns [14:32:07] those guys did the restbase config for us! [14:32:09] :) [14:32:18] how milimetric sorry :) [14:32:23] awesoem :) [14:32:28] https://github.com/wikimedia/restbase/pull/379 [14:32:34] so, now we need a URL [14:32:38] that will always work [14:32:39] joal, hi [14:32:45] (with data) [14:32:53] is anything loaded? [14:32:55] sorry mforns, mistake with milimetric :) [14:33:01] oh ok :] [14:33:02] milimetric: noe, still failing :( [14:33:30] should we insert a fake piece of data? Like 1970: pageviews = 0 [14:33:42] milimetric: our fight with oozie is a tough one :( [14:33:57] ok, I'll help with that fight in a sec, but what do you think, fake data? [14:33:57] There might be some data in by-project [14:33:59] it's just for monitoring [14:34:16] I'm thinking in case we delete or update that, it'd be nice to have something we know will always be there [14:34:59] milimetric: a way to insert / remove fake data would be cool for front-end testing purposes that's for sure ! [14:35:15] no, I don't mean to add a way to insert data [14:35:37] oh, meaning fake data for a fake date for instamce A [14:35:37] I mean, I'd just go into cqlsh and insert something in per-project like 1970-01-01, 0 pageviews [14:36:03] they need a reliable response to a URL for monitoring purposes [14:36:08] milimetric: why not, but so far the only easy way when a clean is needed is to truncate [14:36:26] what's wrong with delete where year = 2015 [14:36:39] too slow? [14:36:42] I think it doesn't work --> no primary key [14:36:48] But I have not tried [14:37:07] ok, hm... [14:37:24] then maybe we can make a delete script that also always makes sure this fake data's in there [14:38:01] milimetric: why not [14:38:10] ok then [14:39:04] Analytics-Backlog, Analytics-Kanban, Wikimedia-Developer-Summit-2016: Developer summit session: Pageview API overview - https://phabricator.wikimedia.org/T112956#1730690 (Qgil) Hi, this Summit proposal has several problems currently: * It seems to be a conventional presentation, which is a type of s... [14:39:44] milimetric: just tested removing without primary key --> doesn't work [14:39:52] So a script will be mostly needed [14:40:09] this seems like a pretty big problem in cassandra :) [14:40:17] like... i can't remove data?! :) [14:40:34] but, aren't we using a constant timeuuid now? Won't that mean we can always just insert / update? [14:40:46] oh, but if we have bad data, we have to delete [14:40:46] milimetric: correct ! [14:40:46] ok [14:46:59] dcausse: still there? [14:50:51] joal: figured out what my problem was yesterday [14:51:00] nuria: what was it ? [14:51:06] path stuff ? [14:51:31] ebernhardson: yt? [14:51:59] nuria: too old hive-site, we used to have a setting that did not let you insert on external tables [14:52:10] joal: "only" took 1 whole day to figure it out [14:52:16] joal: man .... [14:52:28] nuria: I'm veru sorry to have sent you on the wrong direction :( [14:52:36] nuria: We talked about the hive stuff ... [14:52:44] joal: nah, it is oozie, testing cycles are too long [14:52:54] nuria: same on my side --> issue comes from jar it seems [14:52:58] nuria: yep [14:53:03] I don't know how feasible [14:53:06] :( [14:53:13] joal: ah it's not precedence then? [14:53:15] Man, difficult end of week [14:53:24] nuria: not only [14:53:32] dcausse: give us a sec [14:53:36] Analytics-Cluster: vagrant role that installs kafka in vagrant so users can test monolog in mediawiki - https://phabricator.wikimedia.org/T115703#1730739 (Aklapper) Hi @Nuria. Please associate at least one [[ https://phabricator.wikimedia.org/project/query/active/ | project ]] with this task, otherwise nobod... [14:53:40] joal: what else? [14:53:53] joal: jar is bad? [14:54:03] jar wasn't containing the right version of guava [14:54:16] But now it is, and I still get issues [14:54:23] Too bizarre [14:54:25] joal: precedence ? [14:54:35] joal: or functionally it doesn't work? [14:54:42] nope --> testing outside of oozie [14:54:50] joal: if it was not the right version... how dod it work with the plain java task? [14:55:16] nuria: I have seen that jar working, for sure ! [14:55:28] nuria: I'll review changesets [14:55:42] joal: ya.. now is when you start doubting yourself .. like " i swear i saw this working!" [14:55:50] :D [14:56:01] I hate you nuria :-P [14:57:10] dcausse: back, i have created a ticket for the kafka role in vagrant: https://phabricator.wikimedia.org/T115703#1730739 [14:57:21] nuria: thanks! [14:57:36] dcausse: do you have a monolog role or is monolog working on vagrant? [14:58:50] dcausse: it's just a config I think. I'm still stuck on it. I installed kafka manually but I'm not yet able to send anything there [14:59:05] nuria: ^ [14:59:37] will check with ebernhardson when he arrives [14:59:37] dcausse: ok, we cannot help you with the monolog side [14:59:44] yes sure [14:59:59] I had problems with hadoop and yarn [15:00:11] yarn refuses to start [15:00:16] dcausse: i would focus on getting to kafka first [15:00:24] ok [15:00:40] dcausse: i am not sure hadoop role on vagrant works anymore , last time i used it was a year back [15:01:11] Analytics, Discovery, EventBus, MediaWiki-General-or-Unknown, and 7 others: EventBus MVP - https://phabricator.wikimedia.org/T114443#1730753 (Eevans) >>! In T114443#1726139, @Eevans wrote: > I've expanded upon @gwicke's prototype a bit, progress here: https://github.com/wikimedia/restevent Actual... [15:01:17] not a big deal for now, if I can send messages to kafka it will be good already :) [15:01:24] dcausse: should i file a ticket about that too? were you able to make it work? [15:01:43] no I gave trying to start yarn [15:01:49] s/gave/gave up/ [15:03:32] dcausse: one thing is that you need a lot of memory, probably about twice as much as your usual vagrant [15:03:38] dcausse: i filed ticket: https://phabricator.wikimedia.org/T115707 [15:03:48] yes I set up 8G for vagrant [15:04:06] Analytics-Backlog, Analytics-Cluster: Vagrant hadoop role needs maaintenance - https://phabricator.wikimedia.org/T115707#1730780 (Krenair) [15:04:19] joal: ok, so where would you like a script that wipes the cassandra tables? [15:04:34] btw, this is the fake data I ended up going with: [15:04:34] insert into "local_group_default_T_pageviews_per_project"."data" ("_domain", project, access, agent, granularity, timestamp, "_tid", "_del", views) values ('analytics.wikimedia.org', 'en.wikipedia', 'all-access', 'all-agents', 'hourly', '1970010100', minTimeuuid('1970-01-01'), null, 0); [15:05:07] milimetric: I don't know :( [15:05:38] milimetric: we have no code on those machines except the restbase one [15:05:46] Analytics-Backlog, Analytics-Cluster: Vagrant hadoop role needs maintenance - https://phabricator.wikimedia.org/T115707#1730786 (Nuria) [15:07:07] joal: well, are you going to wipe the data again? [15:07:15] like, do you plan to [15:07:46] milimetric: can't be sure :( [15:07:58] I think I don't, but ... [15:08:13] let's talk in batcave, maybe we can solve all the problems :) [15:08:40] omw [15:12:20] Analytics-Backlog, Analytics-Cluster: vagrant role that installs kafka in vagrant so users can test monolog in mediawiki - https://phabricator.wikimedia.org/T115703#1730789 (Nuria) [15:25:19] milimetric: oozie worked ! [15:25:29] YEEEEEEEEEEEEEEEEEEEEES [15:25:30] :) [15:25:31] awesome [15:25:40] nuria: got the error ! [15:25:51] It's a tricky one ! [15:26:02] joal: yesss? [15:26:03] Man ... batcave for a minute nuria and milimetric ? [15:26:07] oops, phoen [15:26:08] brb [15:26:11] nuria: here [15:26:11] sure, 1 min [15:26:46] back [15:26:47] cave ? [15:26:55] ebernhardson: hello, sounds like you modified the schema [15:27:11] (in da cave) [15:28:01] ebernhardson: take a look at: https://github.com/wikimedia/analytics-refinery-source/blob/master/refinery-camus/src/main/avro/CirrusSearchRequestSet.avsc [15:28:21] nuria: yes, i finally finished up my other stuff and had a chance to update our stuff with the new timestamp field and the various defaults [15:28:28] ebernhardson: likely want to send us a new gerrit patchset with teh new schema and tests? we can talk in a few [15:28:48] ebernhardson: or maybe this version of teh schema (tried and tested) works for you [15:29:00] nuria: sure that makes sense, i was having a hard time finding where it went. The main change was you said you needed a timestamp field? [15:29:29] ebernhardson: the schema i sent you has everything that we need [15:29:35] ebernhardson: tried & tested [15:29:53] oh ok, i thought you had asked for a timestamp field somehow. I'll swap over to this one then (it should be almost the same) [15:30:21] ebernhardson: then we are set, do get depot and make sure you have mvn so you can execute tests and such [15:30:57] cool. When this deploys next week does anything special need to happen (create the channel w/ partitions in kafka?) [15:37:15] ebernhardson: the topic creation you mean? [15:37:30] nuria: yea, otto had mentioned if it auto-creates it only gets 1 partition [15:37:50] ebernhardson: ah i cannot say, we have to wait for otto to come back for that [15:37:53] ok [15:38:15] ebernhardson: but topic exists cause we tested the full round friday [15:38:33] ebernhardson: i have also filed tickets to add kafka to vagrant [15:39:55] ebernhardson: we do need a timestamp field [15:40:06] Analytics-Backlog, Analytics-Cluster, MediaWiki-Vagrant: vagrant role that installs kafka in vagrant so users can test monolog in mediawiki - https://phabricator.wikimedia.org/T115703#1730866 (bd808) [15:40:40] ebernhardson: if there's no timestamp the messages in kafka will be partitioned into the hour in which they are imported, which can get messy for your data [15:40:50] sorry I missed your ping yesterday [15:42:41] ahh, ok :) [15:42:54] ebernhardson: sorry i missed that too [15:43:03] no biggie on timing, one day is as good as the next with all deployments canceled ;) [15:43:06] ebernhardson, madhuvishy : then we need to update our schema? [15:43:08] (and its no deploy friday anyways) [15:44:03] nuria: yeah [15:45:21] ok, madhuvishy can you file a ticket for that ? [15:45:31] or rather i can do it [15:45:37] nuria: ya, andrew put a comment on the old ticket [15:45:41] https://phabricator.wikimedia.org/T113521 [15:45:53] but let me do that [15:46:09] madhuvishy: ok, otherwise i can do it [15:54:14] (PS1) Mforns: [WIP] Add oozie job to compute browser usage reports [analytics/refinery] - https://gerrit.wikimedia.org/r/246851 (https://phabricator.wikimedia.org/T88504) [15:54:47] (CR) Mforns: [C: -1] "Still WIP." [analytics/refinery] - https://gerrit.wikimedia.org/r/246851 (https://phabricator.wikimedia.org/T88504) (owner: Mforns) [15:56:39] Analytics-Backlog, Discovery, Discovery-Cirrus-Sprint: Update CirrusSearchRequestSet schema to have a timestamp field - https://phabricator.wikimedia.org/T115715#1730939 (madhuvishy) NEW a:EBernhardson [15:56:54] ebernhardson: I made one and assigned it to you [15:56:59] nuria: ^ [15:57:58] thank you! [15:59:43] madhuvishy: ok sweet, i'll submit a patch to camus (little later today, heading into interview now) [16:02:53] nuria: standup? :) [16:03:19] arghhh [16:20:43] Analytics-Kanban: improve timeuuid writing {slug} [5 pts] - https://phabricator.wikimedia.org/T115353#1731029 (JAllemandou) [16:22:20] Analytics-Kanban: special character stripping on cassandra loading (tabs) {slug} [5 pts] - https://phabricator.wikimedia.org/T115356#1731035 (JAllemandou) [16:22:58] dcausse: I've set up kafka on my own and used it for testing on vagrant before [16:23:03] Analytics-Kanban, Patch-For-Review: Incident: EventLogging mysql consumer stopped consuming from kafka {oryx} - https://phabricator.wikimedia.org/T115667#1731038 (Milimetric) a:Milimetric>mforns [16:23:26] dcausse: while you wait for the role, if you need some help, let me know what you're running into [16:23:32] Analytics-Kanban, Patch-For-Review: Incident: EventLogging mysql consumer stopped consuming from kafka {oryx} - https://phabricator.wikimedia.org/T115667#1729697 (Milimetric) a:mforns>Milimetric [16:23:33] madhuvishy: thanks! :) [16:23:49] just try to send something for the moment :) [16:24:50] dcausse: like nuria said, don't worry about testing the camus end of things, its more complicated, and we've tested that part. I can show you how to test if the consumption will work using python if you want [16:33:38] madhuvishy: if the schema change again will you have to do some work on your side? [16:34:06] dcausse: no, we just need to change the schema and rebuild -> redeploy [16:34:09] but that's expected [16:34:15] I have a mismatch in a field name, wondering if I should update the schema or our code [16:34:27] simpler to change our code I guess [16:34:45] dcausse: we are gonna update the schema anyway [16:34:56] ok [16:34:58] dcausse: see https://phabricator.wikimedia.org/T115715#1730939 [16:35:18] we have to add a timestamp field [16:35:45] ok [16:37:40] the first send failed because the topic was inexistent (I guess) but now it works [16:37:48] looks like topic are created automagically [16:37:58] dcausse: yeah! they are [16:38:06] nice [16:38:57] I don't have any consumers behind this topic, is there a way to deiplay the number of messages there? [16:40:23] dcausse: yeah there are a bunch of tools to see this stuff [16:40:56] dcausse: http://kafka.apache.org/documentation.html#basic_ops [16:41:03] madhuvishy: thanks [16:41:44] dcausse: bin/kafka-console-consumer.sh --zookeeper localhost:2181 --topic test --from-beginning [16:41:52] will run a consumer [16:42:03] nice [17:01:49] Analytics, Services, operations, Patch-For-Review: Automatic monitoring not working for AQS - https://phabricator.wikimedia.org/T115588#1731198 (Dzahn) that fixed it :) https://icinga.wikimedia.org/cgi-bin/icinga/extinfo.cgi?type=2&host=aqs1001&service=Restbase+endpoints+health [17:02:01] Analytics, Services, operations, Patch-For-Review: Automatic monitoring not working for AQS - https://phabricator.wikimedia.org/T115588#1731199 (Dzahn) Open>Resolved [17:03:18] Analytics, Services, operations: Automatic monitoring not working for AQS - https://phabricator.wikimedia.org/T115588#1728063 (Dzahn) [17:10:54] Analytics-Backlog, The-Wikipedia-Library, Wikimedia-General-or-Unknown: Implement Schema:ExternalLinkChange - https://phabricator.wikimedia.org/T115119#1731246 (Sadads) [17:13:37] Analytics-EventLogging, Editing-Department, Improving access, Reading Web Planning, discovery-system: Schema changes - https://phabricator.wikimedia.org/T114164#1731254 (leila) @Jdlrobson, for eye-balling data quality, being able to see page title is very helpful. For joining the data with othe... [17:19:04] Analytics-EventLogging, Editing-Department, Improving access, Reading Web Planning, discovery-system: Schema changes - https://phabricator.wikimedia.org/T114164#1731265 (leila) @csteipp, >>! In T114164#1713239, @csteipp wrote: > In general, we should limit the number of eventlogging schemas t... [17:20:26] Analytics-EventLogging, Editing-Department, Improving access, Reading Web Planning, discovery-system: Schema changes - https://phabricator.wikimedia.org/T114164#1731275 (leila) >>! In T114164#1706669, @JKatzWMF wrote: > I think this field would be invaluable if it provides a reasonable way for... [17:26:33] Analytics, Discovery, EventBus, MediaWiki-General-or-Unknown, and 7 others: EventBus MVP - https://phabricator.wikimedia.org/T114443#1731284 (GWicke) >>! In T114443#1730753, @Eevans wrote: > # Already leverages a (really slick) [[https://meta.wikimedia.org/wiki/Category:Schemas_%28active%29?status... [17:29:19] kevinator, still in batcave :] [17:30:29] I'm tempted to lurk and see what's going on in the batcave.... but I should let you guys work it out without "management interference" [17:45:33] joal: let me know if you want to catch up on the cassandra stuff [17:47:17] nuria, do you have another minute for batcave? [17:48:05] nuria, never mind, let's talk on monday :] [17:49:17] Analytics-Backlog, Discovery, Discovery-Cirrus-Sprint: Update CirrusSearchRequestSet schema to have a timestamp field - https://phabricator.wikimedia.org/T115715#1731359 (EBernhardson) Couple other changes to apply at same time: * index field should be indices (its an array) * integer fields should def... [17:50:00] mforns_gym: i do [17:50:05] mforns_gym: k you let me know [17:55:36] Analytics, Discovery, EventBus, MediaWiki-General-or-Unknown, and 7 others: EventBus MVP - https://phabricator.wikimedia.org/T114443#1731399 (ori) >>! In T114443#1731284, @GWicke wrote: > See T88459#1604768. tl;dr: It's not necessarily clear that saving very little code (see above) for EL schema f... [18:16:54] Analytics, Discovery, EventBus, MediaWiki-General-or-Unknown, and 7 others: EventBus MVP - https://phabricator.wikimedia.org/T114443#1731526 (GWicke) >>! In T114443#1731399, @ori wrote: >>>! In T114443#1731284, @GWicke wrote: >> See T88459#1604768. tl;dr: It's not necessarily clear that saving ver... [18:20:27] (PS9) Joal: Add refinery-cassandra module [analytics/refinery/source] - https://gerrit.wikimedia.org/r/232448 (https://phabricator.wikimedia.org/T108174) [18:21:49] madhuvishy: currently finishing the modifs, hopefully will work ! [18:22:52] (PS10) Joal: Add refinery-cassandra module [analytics/refinery/source] - https://gerrit.wikimedia.org/r/232448 (https://phabricator.wikimedia.org/T108174) [18:23:27] Analytics, Discovery, EventBus, MediaWiki-General-or-Unknown, and 7 others: EventBus MVP - https://phabricator.wikimedia.org/T114443#1731564 (Eevans) >>! In T114443#1731284, @GWicke wrote: >>>! In T114443#1730753, @Eevans wrote: >> # Already leverages a (really slick) [[https://meta.wikimedia.org/... [18:24:27] joal: cool! [18:33:17] milimetric: have you seen this? http://gethue.com/bay-area-bikeshare-data-analysis-with-search-and-spark-notebook/ [18:33:23] had no idea [18:49:13] Analytics, Discovery, EventBus, MediaWiki-General-or-Unknown, and 7 others: EventBus MVP - https://phabricator.wikimedia.org/T114443#1731699 (mobrovac) [18:51:50] (PS14) Joal: [WIP] Add cassandra load job for pageview API [analytics/refinery] - https://gerrit.wikimedia.org/r/236224 [19:02:49] (PS15) Joal: Add cassandra load job for pageview API [analytics/refinery] - https://gerrit.wikimedia.org/r/236224 [19:08:24] madhuvishy: that's cool, yeah, but I think they haven't quite caught up to ipython notebooks [19:08:41] yeah, it seems to use something called spark notebooks [19:08:50] yep, everyone's got a notebook now [19:10:02] that UI is not very intuitive, and it's packed with "features" [19:10:17] I prefer cartodb, tableau, etc. [19:10:47] the OSS stuff is about 5 years behind in terms of BI interfaces [19:14:36] (PS16) Joal: Add cassandra load job for pageview API [analytics/refinery] - https://gerrit.wikimedia.org/r/236224 [20:26:44] Analytics, Discovery, EventBus, MediaWiki-General-or-Unknown, and 7 others: EventBus MVP - https://phabricator.wikimedia.org/T114443#1731947 (GWicke) > For starters, it means that we have alternatives for environments where Kafka is overkill (small third-party installations, dev environments, mw-v... [20:29:04] PROBLEM - Analytics Cassanda CQL query interface on aqs1001 is CRITICAL: Connection refused [20:30:55] RECOVERY - Analytics Cassanda CQL query interface on aqs1001 is OK: TCP OK - 0.001 second response time on port 9042 [20:38:25] !log restarted cassandra on aqs100[1,2,3] [20:38:27] Logged the message at https://www.mediawiki.org/wiki/Analytics/Server_Admin_Log, Master [20:46:33] a-team good weekend! see you on monday [21:17:31] Analytics-Backlog, Discovery, Discovery-Cirrus-Sprint, WorkType-NewFunctionality: Update CirrusSearchRequestSet schema to have a timestamp field - https://phabricator.wikimedia.org/T115715#1732032 (ksmith) [21:57:31] (PS1) EBernhardson: Update CirrusSearchRequestSet avro schema [analytics/refinery/source] - https://gerrit.wikimedia.org/r/246990 (https://phabricator.wikimedia.org/T115715)