[09:19:12] elukey: o/ [09:20:11] o/ [09:22:02] elukey: I managed to have the deploy repo setup, now fighiting with gerrit to push the thing in a branch [09:23:21] nice! thnaks! [09:23:27] *thanks [09:24:24] I guess that git config deploy.remote deploy_repo_remote_name is not sufficient [09:27:17] elukey: I don't have push rights on the deploy repo on gerrit :( [09:27:27] So I can't push my new branch [09:29:33] (PS1) Joal: Update aqs to 4d9f516 [analytics/aqs/deploy] - https://gerrit.wikimedia.org/r/310763 [09:29:57] elukey: I managed that --^ [09:30:06] woooooo [09:30:37] elukey: but I'm not sure of what I did :( [09:30:54] can you list the thing that you did in here? [09:31:03] just to brainstorm [09:31:07] elukey: not really, too much of a mess [09:31:10] :) [09:31:13] I'll do :)( [09:32:18] I got last patches for nuria on aqs repo, getting new-aqs-cluster branch [09:34:40] I merged that into master locally and pushed that master to a gitthub repo, to have a new repo having updated code on master [09:35:25] Then I patched local aqs-deploy git config for the submodule to reference the gitthub repo, and deployed with server.js [09:35:35] it worked, creating a branch [09:36:45] (Abandoned) Joal: Update aqs to 4d9f516 [analytics/aqs/deploy] - https://gerrit.wikimedia.org/r/310763 (owner: Joal) [09:37:18] then noew I'm trying to push the patch to aqs-deploy:new-aqs-cluster (and not master) [09:37:34] but since I don't have push rights on gerrit, I don't know how to do [09:37:38] elukey: --^ [09:38:39] mmmm I didn't get the whole thing :D [09:38:53] tell what needs to be detailed :) [09:39:10] I am trying to get the github repo thing [09:39:42] elukey: Using server.js deploy implies usiong master [09:40:21] Since we don't want to that on gerrit (we use new-aqs-cluster), I created a tmp repo where code is in master instead of new-aqs-cluster [09:41:26] elukey: now, I realize that my things might completely fail, since git will be expecting to be able to access a certain commit in submodule, not having it abailable ... [09:41:52] ah okok now it is clearer :) [09:42:06] but Marko yesterday mentioned that server.js could use a different branch [09:42:13] even if the code yesterday suggested otherwise [09:42:52] elukey: I think mobrovac suggested scap could take a branch, not server.js [09:44:09] I was referring to https://www.mediawiki.org/wiki/ServiceTemplateNode/Deployment#Local_git [09:44:21] re-reading [09:45:34] elukey: no mention of branch in there I think [09:46:56] git config deploy.remote deploy_repo_remote_name [09:47:06] this one seemed the one to use [09:47:33] but not sure if you can specify a branch [09:47:44] because it mentions only origin [09:47:58] just to know: did you try to play with it by any chance? [09:48:25] elukey: nope [09:48:46] but I read from it in the code :) [09:49:00] and it doesn't do what we need [09:49:01] sigh [09:49:21] SO [09:49:32] I know that what I am about to say is not elegant [09:49:34] but [09:49:40] :) [09:50:04] what if we deploy on aqs100[456] master, and hack manually the DTCS settings? [09:50:34] is there also other things that differs from the current cluster? [09:50:39] I might have missed some points [09:50:48] but IIRC the only difference is that one [09:51:06] not sure about the 0/NULL diff though [09:51:13] elukey: I'd suggest something different - Let's move to master assuming everything will fine in traitionning to the new cluster [09:51:28] elukey: that means, no deploy for old aqs anymore [09:52:13] this could be another option as well [09:52:27] and we'll roll back new-aqs stuff if we need to deploy old-aqs anew [09:53:01] elukey: 2 things to deploy for new-aqs: compaction setting and 0/null conversion [09:53:40] is the 0/null conversion already in the current cluster? [09:53:52] (sorry very basic questions but I'd need to clear out some things) [09:54:12] elukey: not yet, but it's in the new-asqs-cluster branch [09:55:25] and this is because the new data loaded on aqs100[456] has this difference meanwhile the one in aqs100[123] has not [09:55:30] correct [09:55:40] ok now it makes sense :D [09:55:43] SOOOOO [09:55:57] +1 on merging in master and deploy only to aqs100[456] [09:56:09] simpler and quicker [09:56:23] and with scap we can also do scap deploy --rev [09:56:27] elukey: agreed - Let's wait on nuria_ to triple check [09:56:51] so in case of fire we can just do scap deploy --rev ROLLBACK --limit 100[123] [09:57:13] well no it doesn't make sense [09:57:15] elukey: k [09:57:16] sorry [09:57:17] ahhahaha [09:57:21] anyhooowww [09:57:23] huhu [09:57:25] let's wait nuria_ [09:57:29] +1 [09:57:31] I am getting crazy with nodejs stuff [09:58:41] yeah, we should continue the talk we started few days ago about package management systmes for not-java languages [10:06:49] elukey: joal: right, sorry, service-runner doesn't take a src repo branch [10:06:51] but! [10:06:54] we can make it do so [10:07:49] if not super difficult it would help in these cases [10:08:29] mobrovac: agreed, not super sdifficult :) [10:08:54] sdifficult? [10:08:55] :D [10:09:07] ...and i shall make it so [10:09:35] give me half an hour to go through some stuff and i'll be on it [10:09:38] shouldn't take long [10:10:01] you rock mobrovac :) [10:10:16] i have a heart of gold! [10:10:17] hahaha [10:10:20] :D [10:10:20] ahahahah [10:30:30] joal: elukey: so, you want to build a specific branch of the deploy repo starting from a src repo branch different from master? [10:30:48] mobrovac: you have it :) [10:30:54] kk [10:31:23] mobrovac: specific branch in repo, I can easily deal with from sync-deploy [10:31:34] but building from another branch than master will be helpful [10:31:46] (the other one as well, but less crucial_ [10:32:57] ok, i [10:33:12] i'll make it possible to do src/branchA -> deploy/branchB [10:33:35] awesome mobrovac :) [10:33:37] Thanks [10:51:58] joal: elukey: uh, gerrit is being a dick about creating remote branches [10:52:04] so, here's what we are going to do [10:52:28] i'll add a feature to service-runner that will allow you to specify the src repo branch [10:52:40] you can then use it to build the deploy repo [10:52:49] but don't use the --review switch [10:53:19] instead, you'll have to create the remote branch manually and then submit the build patch for the deploy repo against that branch manually [10:53:42] it sucks a bit, but it at least allows you to get around the current problems [10:54:44] hm, actually, lemme check something [10:56:19] disregard that, i'll add the deploy branch option too [10:56:45] joal: elukey: but, you have to remember to create the branch manually in gerrit before building the deploy repo [10:56:52] otherwise the process will fail [10:58:21] this seems awesome [10:58:34] glad we could come to an agreement [10:58:35] :) [10:59:36] the only drawback is that we are using an ancient version of service-runner in package.json [10:59:48] so we'll need to upgrade it if we want to use the awesomeness [10:59:49] right? [11:01:54] Analytics, MediaWiki-extensions-WikimediaEvents, The-Wikipedia-Library, Wikimedia-General-or-Unknown, Patch-For-Review: Implement Schema:ExternalLinksChange - https://phabricator.wikimedia.org/T115119#2640146 (Samwalton9) >>! In T115119#2630233, @Nuria wrote: > @Samwalton9: data is being coll... [11:02:48] elukey: yup [11:02:53] no way arround that [11:03:07] i can help with that [11:04:04] hi team! [11:04:49] hey mforns [11:04:54] hi joal ! [11:04:55] Thanks a mil mobrovac :) [11:05:58] hey joal, how is pivot data hell? [11:06:31] mforns: didn't pivot that much this morning, fought with git a bit [11:07:52] joal, me neither, tried executing enwiki [11:08:43] Did it work mforns ? [11:09:02] joal, it's taking some time, because I needed the page history and user history before [11:09:02] mforns: I assume not, givewn you dais try :) [11:09:15] Ah right :) [11:10:36] mforns: sorry for unreadable typos :) [11:11:27] mforns: I assume going for enwiki means simplewiki data using month-splitting were the same as other? [11:13:00] * elukey lunch! [11:22:50] joal, after a couple trials, yes* (* = some "latest" values were different, because the split by month broke the state chains...) [11:23:15] mforns: makes sense [11:23:31] joal, but I thought to move on, because Erik does not need the "latest" values, [11:23:49] joal, and we can compute those in the page/user reconstruction code easily. [11:24:02] or else we have to do a medium refactor in the denormalize code [11:24:44] np mforns [11:25:12] I mean, I think we'll only send simplewiki to Erik [11:26:13] joal, sure! [11:30:36] Now the decision to be taken is: Do we send data to Erik as-is, or do we want to improve it before (2 main issues are page history cut in the middle, and anonymous users user text not well recorded) [11:30:40] mforns: --^ [11:31:49] joal: elukey: tested, works, will release a new version of service-runner soon [11:32:44] joal, maybe we can dig a bit more into the page history cut, and timebox it, and see if we find a quick fix or partial fix [11:33:14] k mforns, will start on that then (after my break though ;) [11:33:20] a-team, taking a reak ! [11:33:29] joal, I'll spend my time until standup looking into this [11:33:42] k, will ping you whan I get back [11:33:45] cya! [11:38:25] mobrovac: you rock [11:39:44] mobrovac: one thing that I wanted to ask you (since you love your colleagues) would be to create a service-node-template for to wrap basic nodejs services that don't need cassandra, api, etc.. [11:39:56] or maybe some doc about the bare minimum skeleton [11:40:33] service-node-template doesn't use or need cassandra [11:41:06] it is a skeleton around tjhe express framework with added bonuses, like logging, metrics collection, error handling etc [11:41:43] sorry I referenced cassandra because I found var uuid = require('cassandra-uuid'); in one of the lib/ files [11:41:50] anyhow, you got my point [11:48:59] elukey: euh, no not sure i do [11:49:27] the things that are in the template are minimal things that we expect prod services to have [11:53:44] mobrovac: all right so probably I need to dive a bit more into that service template :) [11:53:47] will let you know [11:59:18] elukey: ok, new version published [12:00:00] elukey: you don't need to update service-runner in package.json, just do npm i service-runner@2.1.6 in the root of the src repo [12:00:43] this will install the new service-runner temporarily and you will be able to use the new features [12:01:17] ah nice! [12:01:21] then, create the needed branches in gerrit for src and deploy repos and initialise them properly (pull them locally and do git review -s) [12:01:52] finally, you need to tell the build script the branches' names with git config srcbranch|deploybranch [12:01:57] and you are ready to build [12:02:17] updated https://www.mediawiki.org/wiki/ServiceTemplateNode/Deployment#Local_git with the info on srcbranch and deploybranch [12:02:29] let me know if you encounter any problems [12:03:02] sounds easy, will wait for joal to test it [12:04:10] mobrovac: the only thing that doesn't work is the npm i in the root of src [12:04:18] npm tells me "invalid" [12:04:49] you are in the root of the src repo, not the src submodule of the deploy repo, right? [12:05:15] yes [12:05:32] where node_modules is [12:05:54] └─┬ service-runner@2.1.6 invalid [12:05:55] └─┬ limitation@0.1.9 [12:05:55] └── kad@1.3.6 [12:06:07] hm so it did install it? [12:06:29] what does grep version node_modules/service-runner/package.json say? [12:06:59] WHAT - 2.1.6 [12:07:07] at this point I am confused [12:07:45] it's probably because in your package.json you have a version that doesn't match 2.1.6 so it complained about that [12:08:04] never mind that now [12:08:13] I've read it as "mess with dependencies, not installing it" [12:08:27] it will still install the service-runner version you have in package.json [12:08:28] anyhow [12:08:44] I got your point, thanks :) [12:11:15] for the build process, that is ^ [12:11:20] s/for/during/ [12:11:26] anyhow, all good, you can continue [12:11:27] :D [12:11:39] thanks! [12:35:02] Analytics-Tech-community-metrics, Developer-Relations (Jul-Sep-2016), Documentation: Create basic/high-level Kibana (dashboard) documentation - https://phabricator.wikimedia.org/T132323#2640352 (Aklapper) I updated the screenshot and the terms in https://www.mediawiki.org/wiki/Community_metrics#User_... [12:48:14] (PS1) Joal: Update aqs to 4d9f516 [analytics/aqs/deploy] (new-aqs-cluster) - https://gerrit.wikimedia.org/r/310816 [12:48:38] mobrovac: That patch of yours is awesome :D [12:48:55] elukey: you're good to go --^ [12:49:08] Back to being away ;) [12:53:02] it worked i see :) [12:53:03] awesome! [13:01:56] (CR) Elukey: [C: 1] "I haven't checked node_modules dependencies for obvious reasons but the src SHA looks good :)" [analytics/aqs/deploy] (new-aqs-cluster) - https://gerrit.wikimedia.org/r/310816 (owner: Joal) [13:02:22] joal: --^ [13:02:24] LGTM [13:02:51] now the next step is to merge and do the trick with scap [13:03:09] maybe --rev X --limit aqs1004 to test? [13:09:05] * elukey awards a hearth token to mobrovac [13:09:43] maybe without the ending h [13:09:46] details [13:09:55] :) [13:10:10] for deploying, you can perhaps use --environement too [13:10:12] * mobrovac checking the docs [13:16:12] https://doc.wikimedia.org/mw-tools-scap/scap3/quickstart/setup.html documents the environments [13:16:19] but there is no mention of the list of nodes... [13:16:26] so -l ought to be used [13:19:21] there is a --list no? [13:19:25] sorry limit [13:20:30] yes [13:20:33] -l == --limit [13:20:34] :) [13:20:51] also, i looked at the scap code, and you can use --rev branch_ame [13:20:52] name [13:21:02] so you should be good to go here [13:22:14] nice [13:22:28] joal: ? [13:25:25] elukey: you can also have a specific scap.cfg for the new cluster if you want [13:25:47] elukey: I'm here, let's try [13:26:17] eqi aqs1004 [13:26:19] oop [13:31:50] mobrovac: thanks, will check that.. but atm I just want a way to make this deployment.. After that, the plan is to put the new cluster in prod and nuke the old cluster :) [13:32:08] sure, sure, elukey [13:32:12] just mentioning .,.. [13:32:15] (CR) Elukey: [C: 2 V: 2] Update aqs to 4d9f516 [analytics/aqs/deploy] (new-aqs-cluster) - https://gerrit.wikimedia.org/r/310816 (owner: Joal) [13:32:41] mobrovac: you are very kind sir, thanks ;) [13:38:43] elukey: moving forward? [13:39:36] I was finishing a thing in deployment-prep, checking tin now [13:39:49] elukey: no rush, I was wondering :) [13:44:49] so what I did [13:44:59] 1) git fetch && git checkout new-aqs-cluster [13:45:03] all good and sound [13:45:12] 2) git checkout master [13:45:29] now following what has been written, I should be able to do: scap deploy --rev new-aqs-cluster --limit aqs1004.eqiad.wmnet [13:45:49] seems correct elukey [13:46:05] all right executing [13:46:41] I forgot submodule [13:46:42] grr [13:46:51] :( [13:48:18] it seems doing promote and restart now [13:50:10] all right promote failed and I rolled back [13:50:13] checking logs [13:51:16] elukey: don't git checkout master, stay on new-aqs-cluster and git submodule update --init from there [13:52:59] mobrovac: yes yes it was my next action, 13:52:32 Finished Deploy: analytics/aqs/deploy (duration: 00m 01s) [13:53:10] all good :) [13:53:24] yay :) [13:53:55] joal: do you want to sanity check with me aqs1004? [13:54:03] sure elukey [13:56:06] mmmm I don't find anything in /srv/deployment/analytics/aqs/deploy/src [13:56:32] in fact alerts in ops :D [13:56:49] right ... [13:56:58] elukey: checkout issues? [13:57:11] or maybe submodule issue [13:57:22] * mobrovac looking [13:57:57] the submodule wasn't checked out there [13:58:59] mobrovac: ? [13:59:18] ? [13:59:46] this one --> "the submodule wasn't checked out there" [13:59:48] :D [14:00:04] on aqs1004 [14:00:08] the src dir is empty [14:00:21] ah yes ok [14:00:29] I thought you were referrint to tin [14:00:41] ah! [14:00:42] maybe the --ref doesn't work as expected? [14:00:58] elukey: try scap deploy --force --rev new_aqs_cluster [14:01:01] what if I try scap deploy --limit aqs1004 from new-aqs-cluster? [14:02:04] i see in the logs "/srv/deployment/analytics/aqs/deploy-cache/revs/56ab863456b409b1f27075597bc9c0b2f69a5da3 is already live" [14:02:05] so it didn't do anything [14:02:06] so you need to force the deployment [14:02:23] but the bizarre thing is - why does it say that the dir already exists? [14:04:10] mobrovac: mmm I did but scap didn't like it a lot [14:04:49] Unable to checkout '4d9f5160687f8dc3df3401453d2da5e861c19db7' in submodule path 'src' [14:06:24] elukey: a workaround would be to create a dummy commit in the new_aqs_cluster branch (like change the readme or touch a stupid file or whatever) and then try to deploy the new SHA1 [14:06:45] sometimes scap gets confused with re-deploying the same rev or sth [14:06:54] anyhow, better to have a clean one [14:07:35] Analytics-Tech-community-metrics, Developer-Relations (Jul-Sep-2016), Documentation: Create basic/high-level Kibana (dashboard) documentation - https://phabricator.wikimedia.org/T132323#2640523 (Aklapper) Open>Resolved https://www.mediawiki.org/w/index.php?title=Community_metrics&oldid=223914... [14:08:32] mobrovac: we can try [14:09:17] ottomata: we have event logging events in hadoop, right? [14:09:25] joal: do you have time to file a new code review? [14:09:27] but there's no hive table? [14:09:29] or I just can't find it [14:10:28] milimetric: no hive table [14:10:29] heheh [14:10:31] i was about to do something similar [14:10:38] no, it would be a night mare to maintain hive tables for that :) [14:10:47] i'm trying to get a spark shell now to try uaparser [14:10:53] was going to add an example to the wikitech page [14:11:00] buuut, am still waiting for room on hadoop! [14:11:05] ottomata: ok, cool, to answer tilman's question right? [14:11:11] ya [14:11:18] elukey: hmm, the scap state does not seem clean on aqs1004 [14:11:19] elukey: I can do that [14:11:27] ok, I'll let you do that then, and post a simple example of using it from Hive [14:11:33] ok cool [14:11:39] but explain why it's hard to maintain hive tables [14:11:47] because the data is [14:11:49] 'free form' [14:11:51] yes there are schemas [14:11:58] but the schemas vary in backwards incompatible ways [14:12:07] and there are hundreds of them [14:12:09] mobrovac: we have that revision on the new-aqs-cluster branch even in the src repo [14:12:19] coild it be that scap tries to look in master? [14:12:19] elukey: i'd advice forcing a deploy of 2c471aa2abbfc21d31eb7edbfb8df8387d6d922d first (the current master), then introduce the dummy commit and then try to deploy new_aqs_cluster [14:12:27] we'd have to manually maintain each hive table [14:13:03] mobrovac: so checkout master and then scap deploy --rev 2c471aa2abbfc21d31eb7edbfb8df8387d6d922d --limit aqs1004? [14:16:32] elukey: git checkout master && git submodule update --init && scap deploy -f --limit aqs1004 [14:16:45] yes ok [14:16:59] ottomata: in case of 'mutually compatible' schemas, spark could merge data :) [14:17:19] ok done [14:17:39] Analytics-Tech-community-metrics, Developer-Relations (Jul-Sep-2016), Documentation: Create basic/high-level Kibana (dashboard) documentation - https://phabricator.wikimedia.org/T132323#2640549 (Qgil) Thank you very much! [14:17:45] now I can see src populared [14:17:47] mforns: I'm watching elukey fight with aqs, but can still isten to you while I'm not needed [14:17:49] *populated [14:18:06] joal, batcave? [14:18:14] sure mforns [14:18:16] k [14:18:59] joal: ja, spark can deal, but hive not as well [14:19:08] ottomata: yes [14:24:00] joal: I am wondering if the fact that we have the new-aqs-cluster branch also in the src repo might cause this [14:25:00] elukey: I wouldn't think it cause troubl [14:25:38] elukey: I would tend to think the same you did: does scap look in branches for submodules? [14:26:09] yes this is what I meant [14:26:15] sorry maybe it wasn't clear [14:26:22] elukey: maybe it was related to having tried a deploy without submodule update [14:26:24] i checked on tin the deploy repo [14:26:32] and .git/modules/src/config shows only master [14:26:34] it happened to me before and generated a mess [14:26:52] ah nice [14:28:05] but "Unable to checkout '4d9f5160687f8dc3df3401453d2da5e861c19db7' in submodule path 'src'" looks really like it doesn't find the commit [14:33:40] :( elukey [14:35:56] Analytics-Kanban, Patch-For-Review: Make reportupdater support passing the values of an explode_by using a file path - https://phabricator.wikimedia.org/T132481#2199910 (Milimetric) reviewing now. btw, to ensure changes get merged in a certain order, you can just chain them in gerrit (make them on top o... [14:36:40] elukey: since you have root, on aqs1004 do: sudo rm -rf /srv/deployment/analytics/aqs/deploy-cache/revs/56ab863456b409b1f27075597bc9c0b2f69a5da3 before the next deploy [14:36:45] to clean up the mess [14:37:15] and then try deploying new_aqs_cluster again [14:38:37] mobrovac: current seems to be revs/2c471aa2abbfc21d31eb7edbfb8df8387d6d922d, ok it should be the right thing to do [14:39:19] yes, 56ab is not used currently, so just clean it up because it doesn't have the submodule checked-out [14:40:13] done [14:42:06] (CR) Milimetric: [C: 2 V: 2] Support passing the exploded values by file path [analytics/reportupdater] - https://gerrit.wikimedia.org/r/306966 (https://phabricator.wikimedia.org/T132481) (owner: Mforns) [14:42:32] mobrovac: so now I retry the deployment [14:42:46] scap deploy --rev new-aqs-cluster --force --limit aqs1004.eqiad.wmnet [14:44:48] wow this time it worked [14:45:06] so it was probably me forgetting to submodule init before deploying? [14:45:11] as joal mentioned [14:45:14] (CR) Milimetric: [C: 2] Make use of the new explode by file feature [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/307243 (https://phabricator.wikimedia.org/T132481) (owner: Mforns) [14:45:19] (Merged) jenkins-bot: Make use of the new explode by file feature [analytics/limn-language-data] - https://gerrit.wikimedia.org/r/307243 (https://phabricator.wikimedia.org/T132481) (owner: Mforns) [14:45:31] thanks milimetric :] [14:45:54] milimetric, are you merging and deploying? [14:45:54] thanks mobrovac [14:46:05] hope that I didn't delay your trip :) [14:46:11] mforns: reportupdater deploys automatically right, it's still ensure: latest? [14:46:34] I'll double check [14:47:18] ok, I git pulled just in case [14:48:24] elukey: tested usual requests, works like a charm :) [14:48:25] (CR) Milimetric: [C: 2 V: 2] Make use of the new explode by file feature [analytics/limn-ee-data] - https://gerrit.wikimedia.org/r/307244 (https://phabricator.wikimedia.org/T132481) (owner: Mforns) [14:49:13] elukey: Also, tested 0/null conversion: works :) [14:49:17] elukey: uh just in time, just in time [14:49:20] now, i have to run! [14:49:29] (CR) Milimetric: [C: 2 V: 2] Make use of the new explode by file feature [analytics/limn-multimedia-data] - https://gerrit.wikimedia.org/r/307273 (https://phabricator.wikimedia.org/T132481) (owner: Mforns) [14:49:40] milimetric, the 6 RU patches were to be deployed like 1) support feature, 2) all query repositories using the feature, 3) disable deprecated old way [14:49:53] mobrovac: o/ [14:49:59] joal: gooood! [14:50:00] mforns: yeah, I read that [14:50:06] mforns: that's what should happen [14:50:07] milimetric, ok ok [14:50:09] can I proceed with 1005 and 1006 ? [14:50:26] I merged reportupdater, it should've pulled it before doing the next job, but I pulled it manually just in cse [14:50:34] now I'm merging all the configs [14:50:47] mforns: how come in some places you changed it to wiki_db and in others you left it as "wiki"? [14:50:53] awesomel milimetric thanks a lot! :] [14:51:22] milimetric: sorry, y'day and today time simply passed by without me having time to look at pivot [14:51:29] i'll get to it monday, scout's honour! [14:51:42] mobrovac: you don't have to look at pivot at all really [14:51:56] mobrovac: it's just in typescript, and needs to be compiled to JS before it runs [14:52:02] i do need to look at the gulp buid though [14:52:08] build [14:52:22] mobrovac: oh ok, sure, no huge rush, but the sooner the better [14:52:30] kk [14:52:34] mobrovac: I can always work around it by building gulp and checking that into the repo [14:52:42] mobrovac: so if you don't have time to think/look let me know [14:52:54] let you know monday [14:52:57] * mobrovac out [14:53:02] have a nice night [14:54:21] milimetric, I tried to leave the var names as they were, I didn't change wiki to wiki_db IIRC, did I? [14:54:26] (CR) Milimetric: [C: 2 V: 2] Disable the deprecated option by_wiki [analytics/reportupdater] - https://gerrit.wikimedia.org/r/306968 (https://phabricator.wikimedia.org/T132481) (owner: Mforns) [14:55:02] mforns: in one of the patches you did, the multimedia one [14:55:27] (CR) jenkins-bot: [V: -1] Disable the deprecated option by_wiki [analytics/reportupdater] - https://gerrit.wikimedia.org/r/306968 (https://phabricator.wikimedia.org/T132481) (owner: Mforns) [14:55:40]