[02:19:54] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools-dev.diskspace._var.byte_avail.value (10.00%) [03:21:03] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools-dev.diskspace._var.byte_avail.value (80.00%) [04:23:33] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools-dev.diskspace._var.byte_avail.value (90.00%) [05:19:37] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools-dev.diskspace._var.byte_avail.value (100.00%) [06:07:41] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools-dev.diskspace._var.byte_avail.value (90.00%) [06:16:38] !log toollabs killed a "vi" belonging to user icelabs and running for two days saturating the I/O network bandwidth, and rm'ed a 3.5T .final_mg.txt.swp [06:16:38] toollabs is not a valid project. [06:18:48] !log tools killed a "vi" process belonging to user icelabs and running for two days saturating the I/O network bandwidth, and rm'ed a 3.5T(!) .final_mg.txt.swp [06:18:53] Logged the message, Master [06:23:42] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools-dev.diskspace._var.byte_avail.value (100.00%) [06:35:21] RECOVERY - ToolLabs: Low disk space on /var on labmon1001 is OK: OK: All targets OK [07:48:50] !ping [07:48:50] !pong [08:10:25] hi. when I try to stop lab jobs using jstop, they end up going into state dr(instead of deletion) [08:10:58] what can I do? [10:31:27] Hello, new to toollabs [10:32:09] I am exploring around how to install my dependencies. [10:32:34] npm install fails. [10:33:16] note: i also created a stackoverflow tag: #wikimedia-labs [10:52:25] Anyone with toolabs-npm experience?? [11:50:40] yug: ping me when you got more info :) but in my case npm install was okay most times [11:52:16] yug: And thank you for making the tag but that doesn't make lazy people around chatty :) [11:59:14] yug: http://stackoverflow.com/questions/26650243/is-npm-install-available-into-wikimedia-tool-labs yes, the comment is right, I always did the trick [13:07:11] 3Wikimedia Labs / 3deployment-prep (beta): HHVM fcgi restart during scap runs cause 503s (and failed tests) - 10https://bugzilla.wikimedia.org/72366#c19 (10Antoine "hashar" Musso (WMF)) When the change "make scap restart HHVM" ( https://gerrit.wikimedia.org/r/#/c/163078/ ) has been deployed on the beta clust... [13:22:27] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools.diskspace._var.byte_avail.value (10.00%) WARN: tools.tools-dev.diskspace._var.byte_avail.value (100.00%) [13:34:57] Wait, you got npm working fine? (Hooked) [13:35:37] ebraminio: need to know more, see you on stackoverflow [13:36:45] ebraminio: StackO comment suggest to reinstall npm. [13:37:05] yug: no, it says you should disable ca [13:37:26] yug: npm config set ca "" [13:39:40] yug: next question is how to run node on labs? probably using fcgi [13:41:23] yug: If you know how to do that in details, please let know :) [13:42:30] I team up with an india guy who will handle the fcgi. [13:43:54] I also have node experience on my machine. [13:45:23] yug: so would be nice if you guys write a basic guide for how to do node fcgi on labs [13:46:24] yug: I've stock on that and currently running my script as a command :) [13:46:38] I asked my teamate already. Hope he can do. [13:47:42] ebraminio: writing the cgi-NodeJs documentation [13:49:15] yug: Great! Thanks a lot [13:51:04] The is no page Help:Nodejs [13:51:28] yug: you can use http://wikitech.wikimedia.org/wiki/User:Yug/cgi-NodeJs I would say [13:56:15] hey yug [13:56:24] yug: arun is sitting in front of me, says he doesn't know what FCGI is :) [14:02:40] yug: So probably the way you will go is making an instance [14:02:49] ebraminio: not necessarily. [14:03:05] ebraminio: so there are two problems here. one is that yug tried to do 'sudo npm install' which won't work on tools [14:03:28] ebraminio: the other is that npm on ubuntu precise has problems and won't be fixed. you can use it properly from trusty, which you can access by sshing to trusty.tools.wmflabs.org in stead of tools-login.wmflabs.org [14:03:44] YuviPandano: no it is not sudo http://stackoverflow.com/questions/26650243/is-npm-install-available-into-wikimedia-tool-labs [14:03:48] not -g even [14:03:51] YuviPanda: Shit. You are working on a cgi, just add the F. [14:03:54] ebraminio: indeed, that's the second problem I was talking about [14:04:18] ebraminio: and he did try sudo, since I got an email saying someone attempted to run 'sudo npm install topojson' :) [14:04:32] ebraminio: and the other person yug is working with is sitting in front of me and told me this is what happened :) [14:05:24] YuviPanda: Ok :) [14:05:46] YuviPanda: So do you know how to properly run a node.js code? [14:06:03] YuviPanda: I admit it... i made the sudo npm install topojson. [14:06:05] ebraminio: I haven't tried, but let's try now! :) [14:06:06] YuviPanda: Honestly I confused on doing https://github.com/samcday/node-fastcgi-application [14:06:56] ebraminio: ah, hmm. that might be an option, yeah... Although I dunno if that's the best way to do it. [14:07:29] ebraminio: looking for docs, moment [14:07:42] YuviPanda: i do run node scripts with passing env.variable. Run node "properly" ... is an interesting question. [14:07:44] yug: do you need a nodejs *server*, or do you just need to run a script or two? [14:08:57] I run scripts, no server needed pn short term. [14:09:11] ah ok [14:09:21] yug: is your code on github somewhere? [14:10:16] the endpoint is to covert topojson into svg via nodejs script with d3js and Jsdom modules. [14:10:32] yes, git is online. [14:10:37] link? [14:11:04] YuviPanda: Just to let you know, I want to run this code http://fa.wikipedia.org/wiki/User:Ebraminio/commonsdupe [14:11:39] YuviPanda: currently am running it with a php node calling code [14:12:04] but like to do that on better way, not even cgi [14:12:32] YuviPanda: http://goo.gl/ZVsVRS [14:12:57] YuviPanda: not clean, doc to rewrite. [14:14:22] yug: you need to add a package.json file in your repo for npm to work. https://www.npmjs.org/doc/files/package.json.html [14:14:56] Node modules in http://goo.gl/llEAEM [14:15:14] a single script for now. [14:15:39] YuviPanda: yes, package.json is coming [14:15:55] yug: cool :) poke me when you've done that and we'll figure out a way to get this to run [14:16:58] anyone seem GoingBatty on IRC? [14:19:14] Betacommand: no, I haven't... [14:19:55] YuviPanda: Yes. When do you need my code to migrate / demo on labs? [14:20:12] yug: well, I'm going to be around for an hour or so at best :) [14:20:25] no way~~~~~ [14:20:31] hmm? [14:21:33] YuviPanda: Ok, i can do, but likely to have breaking parts. [14:22:11] yug: well, you shouldn't block on me as such. arun started an email thread going as well, so do keep that going as well :) [14:32:23] andrewbogott: around? Trusty tools instances have had broken puppet for weeks, I'm attempting to fix them. Looks like we'll have to modify the gdal package by hand... [14:33:06] What is gdal? [14:33:31] andrewbogott: libgdal is a nice library that has a lot of geo features, is used by a bunch of tools [14:34:03] andrewbogott: I suppose just changing the depends to libmariadbclient-dev should work... [14:37:22] !log tools cleaned out pacct and atop logs on tools-dev [14:37:22] YuviPanda: I need more context to understand this :) What's the instance name of a trusty tool node? [14:37:24] Logged the message, Master [14:37:44] andrewbogott: ah, ok. tools-trusty. [14:38:12] hm, context and also breakfast. back soon... [14:38:29] andrewbogott: ah, ok :) [14:39:54] Aw so much red :( [14:40:06] andrewbogott: yeaaah. [14:40:20] andrewbogott: the violets are just deleted instances.. [14:44:51] RECOVERY - ToolLabs: Low disk space on /var on labmon1001 is OK: OK: All targets OK [14:44:59] andrewbogott: also, why is toollabs hitting the internet (Ubuntu's repos) instead of WMF repos? [14:46:00] YuviPanda: I don't know! Are other labs instances doing that too? [14:46:08] andrewbogott: not sure, checking... [14:46:12] I guess they shouldn't... [14:47:01] andrewbogott: ah, hmm. they seem t o be hitting both [14:52:57] yug: am glad that you got help :) [14:57:30] YuviPanda: looking at sources.list and sources.list.d and I'm confused... [14:57:39] hmm? [14:57:59] They point to security.ubuntu.com/ubuntu which is right... [14:58:17] production hosts seem to have ubuntu.wikimedia.org in their sources.list, which labs instances do not have [14:58:36] but both prod and labs have apt.wikimedia.org in sources.list.d [14:59:01] So… seems like labs instances are maybe wrong and should include ubuntu.wikimedia.org [14:59:15] BUT I don't see a primary ubuntu repo on labs instances at all -- do you? [15:02:14] andrewbogott: what do you mean by 'primary ubuntu repo'? [15:02:51] well… maybe the cloud archive serves that role. [15:03:07] I just mean, the repo that contains normal ubuntu packages, not security updates or wmf custom things [15:03:33] In any case… I don't think there's an actual problem here, right? Things work, and if we add custom packages to carbon they seem to get picked up [15:03:58] andrewbogott: true, but is... weird? and I'd want labs to mirror prod... [15:04:32] I bet that having cloud instances point to cloud repo is on purpose [15:04:38] aaah [15:04:40] hmm, right [15:04:44] maybe some packages are different... [15:06:52] anyway… explain to me about gdal vs. maria? I see most complaints about libvips [15:07:34] andrewbogott: right, so libgdal depends on libmysqlclient-dev, but we have libmariadbclient-dev, which conflicts with libmysqlclient-dev [15:07:39] so we can't install libgdal [15:08:39] Hm, why does it work on precise? [15:09:21] andrewbogott: ah, hmm. it does work on precise, I'm not sure why [15:12:05] on precise libgdal1-dev doesn't depend on libgdal-dev [15:12:22] oh, wait, 'reverse depends' [15:12:24] I'm confused [15:13:00] ebraminio: Thanks, seems Yuvi kick start us pretty well. Good to have him. [15:13:36] yug: Yes. But he forget me which is fine :) [15:14:37] YuviPanda: so maybe the problem is that libmariadbclient-dev should know that it's an adequate substitute for libmysqlclient-dev [15:14:58] andrewbogott: yeah [15:15:09] ebraminio: awww :( one step at a time. first is to get npm working at all.. [15:15:14] https://github.com/samcday/node-fastcgi-application/issues/4 [15:15:16] ebraminio: plus I'm technically on vacation [15:15:25] Is that maria package homemade? [15:15:35] andrewbogott: don't think so [15:15:46] andrewbogott: http://packages.ubuntu.com/hu/trusty/i386/libmariadbclient-dev [15:15:58] hm… paravoid, a little packaging help here if you are still working? [15:16:14] YuviPanda: No problem :) yug: I asked on https://github.com/samcday/node-fastcgi-application/issues/4 probably he will provide some help [15:19:55] andrewbogott: hey [15:20:06] akosiaris: the issue is that we have a package that depends on libmysqlclient-dev but if we install libmysqlclient-dev then libmariadbclient-dev removes itself. [15:20:34] This makes me think that something is wrong in libmariadbclient-dev -- shouldn't it offer to fulfill the libmysqlclient-dev requirement? [15:20:43] sounds like the package needs to be rebuilt to depend on either [15:20:44] Or do dependencies not work like that? [15:21:20] they could... if libmariadbclient-dev "Provides" libmysqlclient-dev [15:21:24] but I doubt it does [15:21:29] trusty or precise ? [15:21:32] I suppose trusty ? [15:21:32] Trusty. [15:21:49] I think this is my question -- is the proper solution to fix the maria package, or fix the thing that depends on libmysqlclient? [15:22:16] (Which, I don't actually know what depends on it, it comes up pretty far down the chain) [15:22:35] which package is the one that causes the problem ? [15:22:55] I'm looking... [15:23:12] Package: libmariadbclient-dev Provides: libmysqlclient-dev [15:23:18] so nothing to be fixed there [15:23:22] libgdal-dev can't install because it wants to install libmysqlclient-dev but we already have libmariadbclient-dev installed [15:23:30] oh, it does provide... [15:23:33] then why...? [15:24:00] Yeah, I was wrong, it's a direct dependency: libgdal-dev Depends: libmysqlclient-dev [15:24:03] The following packages have unmet dependencies: [15:24:03] libgdal-dev : Depends: libmysqlclient-dev [15:24:04] E: Unable to correct problems, you have held broken packages. [15:25:35] held ? [15:25:44] which VM ? [15:25:49] tools-trusty.eqiad.wmflabs [15:25:50] akosiaris: tools-trusty [15:30:47] libmariadbclient-dev 5.5.39-0ubuntu0.14.04.1 stopped having Provides: libmysqlclient-dev [15:30:57] now it only has Breaks: and Replaces: [15:31:22] 5.5.36-1 used to have all three [15:32:27] That'd do it [15:32:38] * akosiaris testing something [15:33:39] so ... I downgraded 1) libmariadbclient18 [5.5.39-0ubuntu0.14.04.1 (now, trusty-security, trusty-updates) -> 5.5.36-1 (trusty)] [15:33:40] 2) mariadb-client [5.5.39-0ubuntu0.14.04.1 (now, trusty-security, trusty-updates) -> 5.5.36-1 (trusty)] [15:33:40] 3) mariadb-client-5.5 [5.5.39-0ubuntu0.14.04.1 (now, trusty-security, trusty-updates) -> 5.5.36-1 (trusty)] [15:33:40] 4) mariadb-client-core-5.5 [5.5.39-0ubuntu0.14.04.1 (now, trusty-security, trusty-updates) -> 5.5.36-1 (trusty)] [15:33:51] and libgal-dev was installable [15:34:11] upgrade should be broken however [15:34:34] ok -- so we need to report this to whoever it is that packages mariadbclient-dev. And... [15:34:38] I am looking at libmariadb's changelog [15:34:40] wait? Or build our own [15:39:36] andrewbogott: no way we build our own [15:39:49] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=739841 [15:40:03] it was supposedly fixed... checking... [15:43:27] andrewbogott: and then reverted https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=759309 [15:44:04] This is quite the emotional roller-coaster [15:45:53] "now it is possible to install packages libmariadbclient and package libmysqlclient-dev at the same time" -- not true, is it? [15:45:56] 3Wikimedia Labs / 3deployment-prep (beta): HHVM fcgi restart during scap runs cause 503s (and failed tests) - 10https://bugzilla.wikimedia.org/72366#c20 (10Bryan Davis) When we talked with Emir from Facebook this spring he outlined how they handle restarts for each node: * Load balancer polls /status.php on... [15:46:11] oh... [15:49:15] akosiaris: I've read that thread and I sort of follow the argument that at build-time it doesn't make sense for the one package to substitute for the other. But then why 'replace'? [15:49:30] http://anonscm.debian.org/cgit/pkg-mysql/mariadb-5.5.git/commit/?id=bc637404c6d7bd30803f08013615d1fec160b049 [15:49:59] sigh... we are caught right in the middle right now... [15:50:39] andrewbogott: might want to read this to clear up the various concepts https://www.debian.org/doc/debian-policy/ch-relationships.html [15:50:50] that is if you don't already know them [15:51:12] I thought I knew them :) [15:54:20] I just don't follow the argument made to remove the 'provides.' [15:55:36] "libmariadbclient-dev and libmysqlclient-dev are not co-installable, but it is my understanding that this is not a problem as such, as it is only a build dependency, not a runtime dependency" [15:56:21] they say that you may choose to build with either but not both [15:56:53] which begs the question [15:57:07] why is libmariadbclient-dev installed ? [15:57:15] Yeah, maybe we don't need it. I'll look [15:57:18] aptitude why libmariadbclient-dev [15:57:18] Unable to find a reason to install libmariadbclient-dev [15:57:26] maybe puppet ? [15:57:46] probably a tools user specifically requested it. I'm looking... [15:58:09] It does not seem like an issue we can report and hope to have it fixed though. They seem to already know about it [15:58:13] YuviPanda: just made 'npm list' to check dependencies. Nothing on g level, as expected. [15:59:19] akosiaris: looks like it's been in there for a while but may not have been specially requested. [15:59:52] Coren: do you have any idea if it's ok for us to replace libmariadbclient-dev with libmysqlclient-dev in dev_environ? [16:00:05] YuviPanda: after 'cd wikiAtlas.../back/' , doing 'npm list' show back my 3 npm packages, but none of their own dependencies. [16:00:21] It would solve some dependency issues. But it's unclear why we install libmariadbclient-dev in the first place [16:01:00] andrewbogott: https://gerrit.wikimedia.org/r/#/c/63560/ [16:01:24] seems like apt was not stopped... simply stalled [16:02:17] I wonder if the change to libmariadbclient-dev in that patch was really motivated or just "while we're at it" [16:14:40] andrewbogott: akosiaris it might have been a 'just while we are at it', since -dev shouldn't really affect the client itself [16:14:53] YuviPanda: I'm setting up a test instance to see if it matters [16:14:58] andrewbogott: ok! [16:15:47] YuviPanda: could be... I can't quite make out the exact nature of the problem Coren was trying to solve back then [16:15:57] yeah... [16:20:36] andrewbogott: We installed libmariadbclient-dev because some mariadb extensions are not available through libmysqlclient-dev [16:20:46] Though I honestly do not know if anyone is using it, per se. [16:21:02] (It also seemed more reasonable to use the mariadb client things with mariadb, but meh) [16:21:11] -dev is needed for the extensions? It's not enough for just libmariadbclient? [16:21:29] Well, not if you want to build against it. [16:22:18] (Also, I think you'll get a dependency conflict if you install libmysqlclient-dev because it'll want libmysqlclient which may or may not be satisfied with libmariddbclient. [16:23:03] We've been reading a somewhat long discussion about competing dependencies… in general they assert that the two are compatible, although I don't know if that means that they're compatible for building purposes. [16:23:20] What we know for sure is that libmysqlclient-dev and libmariadbclient-dev can't coexist [16:25:01] Without knowing what people are building it's hard to test this though :( [16:25:28] The immediate issue is that libgdal-dev won't work without libmysqlclient-dev [16:28:14] Coren, YuviPanda: So, what to do? [16:30:04] Well, if libmariadbclient-dev doesn't provide libmysqlclient-dev, that's an upstream issue; but doesn't solve our immediate problem. I think it's safe to say that if there are anyone building against it we can help when with a different instance at need. "Downgrade" to libmysqlclient{,-dev} imo. [16:31:01] Coren: I could also switch on distro and only downgrade for Trusty [16:31:07] Since that's where the problem is [16:32:35] That's just begging for trouble down the road. If anyone actually relies on libmariadb vs libmysql we need to know. [16:37:59] Coren, YuviPanda: https://gerrit.wikimedia.org/r/#/c/170073/ [16:41:10] andrewbogott: commented [16:42:53] On the commandline it definitely removes the other. I will test... [16:49:29] andrewbogott: Doesn't it ask for confirmation? [16:49:43] I think not, but let me try again [16:51:01] Coren: It does… do you know what args will imitate what puppet does when invoking apt? [16:51:55] Not really, but I know it won't do conflict resolution at all. You're really better off with the ensure=>absent stanza to be removed in a few hours. [16:52:31] ok [16:55:44] Coren: https://gerrit.wikimedia.org/r/#/c/170073/2 [17:06:45] Bah, the dependency chains on Trusty vs. Precise are too different, my patch doesn't work on Precise [17:15:17] Coren: https://gerrit.wikimedia.org/r/#/c/170089/ :( [17:15:45] o_O [17:15:51] Eeeew. [17:33:04] Coren: regarding https://lists.wikimedia.org/pipermail/labs-l/2014-September/002946.html -- Does that mean that Labs only finally moved to Maria in September? (I'm updating the September engineering report.) [17:33:28] YuviPanda: I believe that one issue with apt on tools/trusty is now resolved (badly). there are yet more packaging issues which I am leaving to you for now :) [17:34:30] andrewbogott: 2 of 3 servers were done before; turns out the third was indeed not done before then (much to my surprise and explosion of things) [17:35:05] "The last of the ToolLabs replica servers was upgraded to MariaDB 10" <- accurate? [17:36:26] Coren: ^ [17:36:50] Ayup. [17:55:54] (03PS1) 10Alexandros Kosiaris: Introduce passwords::postgres [labs/private] - 10https://gerrit.wikimedia.org/r/170100 [17:57:54] (03CR) 10Dzahn: [C: 031] ".to paaaarty!" [labs/private] - 10https://gerrit.wikimedia.org/r/170100 (owner: 10Alexandros Kosiaris) [17:59:35] andrewbogott: heh, I'll check. [18:02:18] mutante: :-) [18:03:08] (03CR) 10Alexandros Kosiaris: [C: 032 V: 032] Introduce passwords::postgres [labs/private] - 10https://gerrit.wikimedia.org/r/170100 (owner: 10Alexandros Kosiaris) [18:04:23] akosiaris: password generator: git clone labs/private ; apt-get install randomize-lines ; echo "Your password shall be: $(grep -o "'.*'" ./modules/passwords/manifests/init.pp | rl -c 1)" [18:05:54] mutante: Users are recommended to use the shuf command instead which should be [18:05:54] available by default. This package may be considered deprecated. [18:06:04] apt-cache show randomize-lines [18:06:23] btw, I like the idea :-) [18:10:50] akosiaris: :) [18:10:53] ok [18:13:20] Coren: wanna change the webserver config of toollabs? [18:13:45] mutante: What to, and what for? [18:14:16] Coren: disable SSLv3 [18:14:56] like this one for "dynamicproxy" [18:14:58] https://gerrit.wikimedia.org/r/#/c/169949/2/modules/dynamicproxy/templates/urlproxy.conf [18:28:26] 3Wikimedia Labs / 3tools: program created by proprietary compiler allowed on labs? - 10https://bugzilla.wikimedia.org/72253#c4 (10Marc A. Pelletier) A point of data is that section 2 of http://opensource.org/osd-annotated reads: "The program must include source code, and must allow distribution in source co... [18:35:41] 3Wikimedia Labs / 3tools: program created by proprietary compiler allowed on labs? - 10https://bugzilla.wikimedia.org/72253#c5 (10Marc A. Pelletier) On the third hand, the code being suitably licensed means that its function can easily be reimplemented in some other language (a derived work), or some interpr... [18:50:50] Coren, what's the status of per-project mail routing? do we have that now? [18:51:00] It's been a while since I last checked.. [18:52:50] No news. It's on the "desired features" list. [18:53:14] PROBLEM - ToolLabs: Low disk space on /var on labmon1001 is CRITICAL: CRITICAL: tools.tools-dev.diskspace._var.byte_avail.value (11.11%) [18:53:54] Coren, is there a bug I can CC on? [18:54:20] Krenair: I think there is; lemme try to find it. [19:03:53] Krenair: I can't find it. Either nobody ever made one yet, or bugzilla's horrid search abilities defetaed me. [19:05:44] 3Wikimedia Labs / 3deployment-prep (beta): References to non-existent cswiki in beta's centralauth db - 10https://bugzilla.wikimedia.org/72773 (10Kunal Mehta (Legoktm)) 3NEW p:3Unprio s:3major a:3None http://cs.wikipedia.beta.wmflabs.org/ does not exist mysql> select count(*) from localnames where l... [19:05:56] 3Wikimedia Labs / 3deployment-prep (beta): References to non-existent cswiki in beta's centralauth db - 10https://bugzilla.wikimedia.org/72773#c1 (10Kunal Mehta (Legoktm)) Also breaking http://meta.wikimedia.beta.wmflabs.org/wiki/Special:GlobalRenameQueue/closed [19:11:13] 3Wikimedia Labs / 3deployment-prep (beta): Beta cluster centralauth accounts points to no longer existing wikis - 10https://bugzilla.wikimedia.org/63396#c8 (10Alex Monk) *** Bug 72773 has been marked as a duplicate of this bug. *** [19:11:13] 3Wikimedia Labs / 3deployment-prep (beta): References to non-existent cswiki in beta's centralauth db - 10https://bugzilla.wikimedia.org/72773#c2 (10Alex Monk) 5NEW>3RESO/DUP *** This bug has been marked as a duplicate of bug 63396 *** [19:25:44] andrewbogott: re: other puppet errors, should be easier, I hope :) [19:26:25] !log dynamicproxy restarting nginx on dynamicproxy-gateway - disabled SSLv3 [19:26:25] dynamicproxy is not a valid project. [19:26:30] hrmm [19:26:46] mutante: proxyproject or projectproxy [19:26:47] !search_project dynamicproxy-gateway [19:26:57] !log projectproxy restarting nginx on dynamicproxy-gateway - disabled SSLv3 [19:26:57] projectproxy is not a valid project. [19:27:06] !log proxyproject restarting nginx on dynamicproxy-gateway - disabled SSLv3 [19:27:06] proxyproject is not a valid project. [19:27:12] wat [19:27:45] mutante: lol, project-proxy [19:30:23] !log project-proxy restarting nginx on dynamicproxy-gateway - disabled SSLv3 [19:30:26] Logged the message, Master [19:30:30] :) [19:58:27] YuviPanda: Started back on clean space, going quite smoothly. Some gdal sugar missing, will find a way. [19:58:54] yurikR: are you yurik working on vega? [19:59:02] yug, yep [19:59:13] more like dragging my feet :) [19:59:22] hello. [19:59:35] we all are this way actually. [19:59:43] hehe [20:00:10] yug: yeah, gdal needs a bit more love I Think [20:00:18] Are you coming to Amsterdam wikidata hackathon? November 14-16? [20:00:42] yug: oh, no, I'm not :( [20:00:52] yug, i think i should :) [20:01:08] up to tfinc to org that :) [20:02:01] YuviPanda: globally good, i didn't expected to move that fast on day 1 back on work. The last day of clean up i spent in August also pay off. All smooth, all paths are correct. [20:02:34] yug: :) [20:02:35] nice [20:03:18] yurikR: need to kick start your travel grant request???? I could help, i just posted mine. [20:03:35] yug, sure! [20:03:54] yurikR: Let's go! [20:04:33] yug, do you have a link to wikidata? [20:04:34] see you on meta, post me a one word message on user_talk:yug [20:04:49] yes: our maps!!!! [20:05:40] done [20:05:51] Arun and me made a function to translate D3js map labels from wikidata queries :) [20:07:03] yug, we should really push for lua-based data manipulation ;) [20:07:22] this way you can have mw.setRawResult('some-json-text') [20:08:10] so lua would pull data from wikidata, convert it, and output it in a format read for the graph consumption [20:18:39] Lua is above my current skillset. As wiki work incrementally, i do my outmost to demo structured svg & data biding for maps. More powerful tech will build upon us, either our code either by the awareness we will increase. yurikR [20:19:30] lua is esay, but this capability (setraw) is not (will not?) ready yet [20:19:32] yurikR: Ok! Check your meta, there is a seed of a travel grant request. [20:19:38] yug, thx!!! [20:19:51] i think i should show it to tfinc first [20:21:02] Good luck. We only have 2~3 opportunities a year to meet up and synchronises visions, let's hunt this one. [20:21:07] also, yug, i suspect that i can't apply to the grants as a wmf empl [20:21:20] where are you geographically? [20:21:30] yurikR: I also suspect that ^^ [20:21:41] Paris, France. [20:22:18] yeah, makes it harder to visit SF [20:22:27] i actually might be in europe during that time [20:31:46] If you come to France, please notice me/us (yug, lyhana8, Planemad) [20:32:24] We love your work. yurikR [20:32:47] :) thanks! [20:34:50] Wait: are you working at the office? [20:35:06] WMF in LA. [20:35:08] ?? [20:35:32] WMF is in SF, but i live/work in NY [20:35:43] although i am in NC at the moment [20:38:32] Ok. They check out grant rewuest on friday morning i believe. [20:54:41] how can i best replicate jenkins test runners(specifically phpunit) in labs. I have some phpunit tests that intermittently fail only when run by jenkins so I want to recreate the environment so i can debug the issue [20:54:57] i've already tried setting up my local with sqlite and letting it run overnight, but i can't trigger these intermittent failures locally [21:02:12] 3Wikimedia Labs / 3deployment-prep (beta): Rename all occurences of "deployment-prep" to "beta-cluster" - 10https://bugzilla.wikimedia.org/72694#c4 (10Antoine "hashar" Musso (WMF)) Recreating the project from scratch definitely takes a few days. Migrating the instances to a new labs project would most proba... [21:20:34] Hi all ! [21:20:52] Newbie question: I'm having some trouble logging in [21:21:32] jheald: any more details? i might be able to help, but depends on the issue :) [21:21:38] ssh drops out with "Permission denied (publickey)" [21:21:51] It's the first time I've tried to log in [21:22:17] I think I have the accounts set up -- there's two messages on my talk page [21:22:33] & I've generated and uploaded an rsa key [21:22:36] jheald: and its the same one you uploaded to openstack? [21:22:37] ok [21:23:40] jheald: are you connecting to bastion, or forwarding through it? [21:24:29] & created a .ssh/config file [21:25:49] ok so should be forwarding ... sadly i'm not sure :( thats basically the correct process [21:26:02] sorry -- wireless dropped out [21:26:41] so: should I now be able to ssh tools-login.wmflabs.org ? [21:26:58] or ssh Jheald@tools-login.wmflabs.org ? [21:27:17] oh duh, i'm thinking regular labs and your logging into tools :) [21:27:40] in general though, yes that sounds like the correct process. but i've never logged into tools just openstack labs instances [21:27:43] the other part of wmflabs :) [21:27:54] sorry ebernhardson, I lost IRC for a couple of minutes, so I don't know what you first said [21:28:25] yurikR: if not willing to apply, just delete the request. It will be cleaner. [21:28:43] yug, i'm emailing my boss [21:28:44] jheald: just that your process sounds correct, setting up the ssh/.config, uploading the keys [21:29:05] jheald_: i'm really not sure why its not working, but i've also only used the other (non-tools) half of labs [21:29:15] thx [21:29:22] yurikR: ok, good to give a try. [21:29:31] yep :) [21:34:39] yurikR: I leave, see ya. [21:34:50] yug, thx for the heads up! [21:34:53] good night :) [21:35:04] just emailed tfinc, will see [21:36:43] Do I need an "instance shell account name" to ssh to tools-login ? Is that different from my usual "Jheald" ? [21:44:33] jheald_: You picked it when you created your account. [21:45:12] jheald_: You can double check what it is on your wikitech's preference page; as "shell account name" IIRC [21:45:28] thanks. lower case was needed. created the account six weeks ago, and had forgotten [21:45:52] managed to log in now, thanks all round [22:07:10] (03PS2) 10Legoktm: Refactor into a proper class [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/169309 [22:11:21] (03CR) 10Legoktm: [C: 032] Refactor into a proper class [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/169309 (owner: 10Legoktm) [22:14:28] 3Wikimedia Labs / 3deployment-prep (beta): HHVM fcgi restart during scap runs cause 503s (and failed tests) - 10https://bugzilla.wikimedia.org/72366#c21 (10Antoine "hashar" Musso (WMF)) Thank you for the high level synopsis, that seems to be a good base to add graceful/graceful-stop to the init/upstart scrip... [22:49:34] (03PS1) 10Legoktm: Add basic tests for the tarball generator [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/170240 [22:49:39] (03CR) 10jenkins-bot: [V: 04-1] Add basic tests for the tarball generator [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/170240 (owner: 10Legoktm) [22:50:17] when your tests fail tests >.> [22:50:30] (03PS2) 10Legoktm: Add basic tests for the tarball generator [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/170240 [23:04:02] (03CR) 10Legoktm: [C: 032] "Will figure out how to get jenkins to run these later" [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/170240 (owner: 10Legoktm)