[00:03:44] <hasteur>	 !ping
[00:03:45] <wm-bot>	 !pong
[09:33:42] <ebraminio>	 Hi, can I request "Sys Admin" right or someone can make an instance for me on wikitech wiki? (https://wikitech.wikimedia.org/wiki/Help:Instances)
[09:47:58] <hashar>	 !log deployment-prep mobile varnish frontend cache is not starting anymore : /usr/lib/x86_64-linux-gnu/varnish/vmods/libvmod_netmapper.so: cannot open shared object file: No such file or directory  {{bug|55662}}
[09:48:04] <labs-morebots>	 Logged the message, Master
[09:51:18] <hashar>	 !log deployment apt upgrading all varnish caches to bump varnish from 3.0.3plus~rc1-wm14 to     3.0.3plus~rc1-wm17
[09:51:19] <labs-morebots>	 deployment is not a valid project.
[09:52:37] <hashar>	 zeljkof: I am rebooting the caches of beta
[09:52:59] <zeljkof>	 hashar: great, thanks
[09:53:09] <hashar>	 !log deployment-prep rebooting all varnish caches ( deployment-cache-text1 deployment-cache-upload04 deployment-cache-bits03 deployment-cache-mobile01 )
[09:53:13] <labs-morebots>	 Logged the message, Master
[09:53:22] <hashar>	 zeljkof: but I think they will all be dead after that
[09:53:27] <hashar>	 zeljkof: we are missing a dependency
[09:53:48] <zeljkof>	 hashar: can you fix that too? :)
[09:55:49] <hashar>	 zeljkof: http://en.m.wikipedia.beta.wmflabs.org/ back up apparently
[09:56:04] <zeljkof>	 hashar: great
[09:56:06] <zeljkof>	 thanks
[10:00:06] <hashar>	 zeljkof: upload is broke though :(
[10:00:19] <zeljkof>	 :)
[10:00:32] <zeljkof>	 at least the site is up
[10:12:57] <hashar>	 !log deployment-prep purged varnishhtcpd on deployment-upload04 to make it start again.
[10:13:02] <labs-morebots>	 Logged the message, Master
[10:13:58] <hashar>	 zeljkof: I think everything is back in order now
[10:14:36] <zeljkof>	 hashar: great, thanks
[13:18:20] <Steinsplitter>	 hashar: Hosted tools (http://tools.wmflabs.org/#Hosted%20tools) how you generate this list?
[13:29:10] <hashar>	 Steinsplitter: I have no clue what is generating that
[13:29:31] <hashar>	 Steinsplitter: that might be a LDAP query
[13:29:37] <Steinsplitter>	 ah
[13:29:45] <hashar>	 Steinsplitter: or an API query against http://wikitech.wikimedia.org
[13:30:06] <Steinsplitter>	 yes.  mabye petan knows
[13:50:14] <Coren|Away>	 Steinsplitter: The source is in git (operations/toollabs) under /www
[13:53:28] <Steinsplitter>	 Coren|Away: thx
[14:01:57] <hashar>	 Steinsplitter: got my email regarding upstart jobs ?
[14:02:18] <hashar>	 Steinsplitter: I was also wondering why you talk about node.js when the latex web service is apache/perl 
[14:03:54] <Steinsplitter>	 hashar: I think you're confusing me with someone else? 
[14:04:02] <Steinsplitter>	 i dos not have recived a mail
[14:04:43] <hashar>	 ahh
[14:04:45] <hashar>	 damn
[14:04:49] <hashar>	 physikwert
[14:04:51] <hashar>	 sorry
[14:04:53] <Steinsplitter>	 np
[14:04:55] <YuviPanda>	 :D
[14:05:01] <hashar>	 w
[14:05:12] <hashar>	 we need to get rid of most people, I can't keep track of everyone nick/names
[14:30:56] <ebraminio>	 Hi, can I request "Sys Admin" right on wikitech wiki or someone can make an instance for me? (https://wikitech.wikimedia.org/wiki/Help:Instances)
[14:54:20] <YuviPanda>	 ebraminio: poke andrewbogott or Coren|Away or Ryan_Lane
[14:55:50] <ebraminio>	 YuviPanda: thank you.
[14:55:50] <ebraminio>	 andrewbogott or Coren|Away or Ryan_Lane: Hi, can I request "Sys Admin" right on wikitech wiki or someone can make an instance for me? (https://wikitech.wikimedia.org/wiki/Help:Instances)
[14:55:51] <andrewbogott>	 ebraminio: what project?
[14:57:00] <ebraminio>	 andrewbogott: just want to a little instance and after that may run a little service with nodejs if is possible
[14:57:06] <ebraminio>	 *test
[14:57:24] <andrewbogott>	 ebraminio: OK, there are various approaches here.
[14:57:42] <andrewbogott>	 We can make you a 'project' which is (potentially) a cluster of instances, and give you rights to create/delete instances within the project.
[14:57:51] <andrewbogott>	 Or we can just give you an account in tool-labs...
[14:57:58] <andrewbogott>	 which is what most people turn out to actually want
[14:58:08] <andrewbogott>	 since it gives you shell access and such w/out the headache of administrating entire systems.
[14:58:18] <ebraminio>	 I have tool-labs account
[14:58:58] <andrewbogott>	 I thought so :)  What do you need to do that can't be done via tool-labs?
[15:00:01] <ebraminio>	 Using https://wikitech.wikimedia.org/wiki/Help:Proxy and opening a host socket port
[15:01:18] <andrewbogott>	 A non-web socket, I take it?
[15:02:19] <ebraminio>	 a HTTP socket would be sufficient also
[15:02:43] <andrewbogott>	 Oh, well… aren't they open for tool-labs already?  They all have public IPs and hostnames as far as I know.
[15:03:17] <andrewbogott>	 Sorry if I'm being thick -- proliferation of projects imposes some maintenance overhead so generally we try to steer people into using existing ones if possible.
[15:03:20] <ebraminio>	 andrewbogott: as far as I know and I tested, no
[15:04:11] <andrewbogott>	 Hm… if possible I'd like you to check in with Coren|Away and verify that this can't be accomplished in tool-labs.  If he agrees then you can come back and bug me and I'll make you a project, OK?
[15:04:34] <YuviPanda>	 ebraminio: you want websockets?
[15:05:26] <YuviPanda>	 andrewbogott: I'll see if I can spend time today writing v2 of the proxy, that'll work for toollabs
[15:05:38] <ebraminio>	 YuviPanda: no, websocket is different from http server port, as said just wanted to test if can I setup nodejs REST service
[15:05:42] <andrewbogott>	 Ah, aren't you supposed to be sitting on your hands?
[15:05:57] <YuviPanda>	 andrewbogott: yeah, which is why i wasn't online the last two days
[15:06:02] <YuviPanda>	 but... it's so hard :(
[15:06:22] <YuviPanda>	 ebraminio: ah, hmm, right.
[15:06:31] * andrewbogott  thinks YuviPanda should take up football or marathon running or something else foot-related

[15:07:10] <Coren|Away>	 YuviPanda: FYI: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help/NewWeb
[15:07:21] <YuviPanda>	 andrewbogott: have flat feet...
[15:07:22] <Coren|Away>	 YuviPanda: It would probably be downright *trivial* to tie this to your proxy
[15:07:37] <andrewbogott>	 ok, something elbow-related then
[15:07:56] <YuviPanda>	 Coren|Away: how does it assign ports?
[15:08:23] <Coren|Away>	 YuviPanda: There's a happy fun daemon that assigns ports and registers them; this is what would need to talk to your proxy.  :-)
[15:08:30] <YuviPanda>	 Coren|Away: bingo
[15:08:33] <YuviPanda>	 Coren|Away: where's the code?
[15:08:39] <YuviPanda>	 labs/toollabs?
[15:08:43] <Coren|Away>	 (Right now, it simply tells the webproxy)
[15:08:45] <YuviPanda>	 or ops/pupet?
[15:08:58] <YuviPanda>	 Coren|Away: how does it 'tell' the webproxy?
[15:09:08] <YuviPanda>	 Coren|Away: rewrite apache webproxy's config file and do a graceful?
[15:09:24] <Coren|Away>	 No, it's a flat file that webproxy uses for rewriterules.
[15:09:28] <YuviPanda>	 aha
[15:09:29] <YuviPanda>	 right
[15:09:38] <YuviPanda>	 same thing as my proxyt does but flatfile instead of redis
[15:10:00] <Coren|Away>	 YuviPanda: Code is still very much in flux and not yet puppetized; you can look at the daemon on tools-webgrid-01 /usr/local/sbin/portgranter
[15:10:13] <YuviPanda>	 Coren|Away: ah, sweet
[15:10:36] <YuviPanda>	 Coren|Away: let me add the feature to route by URLs
[15:10:41] <Coren|Away>	 Wait, I lied.  It *is* in puppet now.
[15:11:12] <Coren|Away>	 in toolserver:webnode.pp
[15:11:32] <Coren|Away>	 toollabs::webnode.pp
[15:11:36] * Coren|Away  is still sleepy.

[15:12:36] <Coren|Away>	 The daemon is in perl, but it's simple enough.  Look at the if($update) { } part.  All it does right now is lock then push a flat file on NFS; it should be insanely trivial to tell your proxy instead.
[15:13:20] <ebraminio>	 andrewbogott: I tested a sample host port again and find out that I was wrong about public IP, however I like to have an instance to test things there if is possible
[15:13:39] <YuviPanda>	 Coren|Away: hah ok
[15:13:44] <YuviPanda>	 Coren|Away: i'm going to eat food now, will look at it later
[15:16:57] * Coren  remains insanely happy with that new scheme.

[15:17:14] <Coren>	 *much* easier to control resources, and turns out to be lighter weight to boot.
[15:24:10] <andrewbogott>	 ebraminio: As I said, if you aren't working with an existing project then it's kind of a pain to give you your own.  So I'd advise you to team up with an existing project, or do your work in tool-labs, or try harder to convince me that you need your own :)
[15:24:40] <andrewbogott>	 ebraminio: Part of the background here is that we're going to migrate to a new data center soon, and moving projects may result in downtime.  Tool-labs will get priority there, so will remain more stable.
[15:24:45] <andrewbogott>	 List of other existing projects is here:  https://wikitech.wikimedia.org/wiki/Special:Ask/-5B-5BResource-20Type::project-5D-5D/-3F/-3FMember/-3FDescription/mainlabel%3D-2D/searchlabel%3Dprojects/offset%3D0
[15:29:54] <ebraminio>	 andrewbogott: no problem :) you wanna the fact I wanted setup a test wiki without two level path https://tools.wmflabs.org/..../.... and attach a clean domain to it with A Record and that isn't reasonable I guess
[15:32:25] <Yogi_Bear>	 Invitation - Only WIKIMEDIA http://www.tatuuu.com.br 
[15:54:57] <Danny_B>	 any maintenance again? my ssh is frozen again
[16:02:24] <andrewbogott>	 Danny_B: What instance?
[16:07:41] <petan>	 Steinsplitter??
[16:14:54] <Steinsplitter>	 petan: {{resolved}} 
[16:27:25] <andrewbogott>	 milimetric: I can actually make the switch for you, if you're available to watch.
[16:27:54] <andrewbogott>	 Want to do this now?  I have to go to lunch in 30 or so, but can come back and troubleshoot all afternoon if it comes to that.
[16:28:21] <milimetric>	 sure andrewbogott, let's do it
[16:29:37] <andrewbogott>	 OK, I'll make all the changes, then you can tell me if it broke or not :/
[16:35:15] <milimetric>	 will do, just ping me and I can test
[16:35:19] <milimetric>	 andrewbogott: ^^
[16:35:44] <andrewbogott>	 ok.  I just tried and mysql wouldn't start, so I have some more work to do on my end.
[16:35:50] <andrewbogott>	 Meanwhile… is it working again?  I reverted.
[16:36:33] <andrewbogott>	 milimetric: ^ ?
[16:36:41] <milimetric>	 checking
[16:38:03] <milimetric>	 yep andrewbogott, all happy
[16:38:13] <andrewbogott>	 ok.  Well, at least we know I can safely test :)
[16:38:26] <andrewbogott>	 milimetric: Are you theoretically working today or on holiday?
[16:40:03] <milimetric>	 theoretically on holiday :)
[16:40:19] <andrewbogott>	 Hm… I should leave you be then
[16:40:21] <milimetric>	 nono
[16:40:33] <milimetric>	 hey, I got an idea if you want to run a bunch of tests and not depend on me though
[16:40:49] <andrewbogott>	 ok!
[16:40:50] <milimetric>	 i can set you up with wikimetrics
[16:40:54] <milimetric>	 should be super simple
[16:41:11] <andrewbogott>	 Does it hit the db more-or-less constantly?
[16:41:12] <milimetric>	 just log in and I'll give you some cohorts, then show you what to do
[16:41:15] <andrewbogott>	 Or is it only used for batch jobs?
[16:41:29] <milimetric>	 it hits mostly the labs-db databases
[16:41:38] <milimetric>	 but also a metadata store local on that instance
[16:41:43] <andrewbogott>	 I'm logged in as andrewbogott@gmail.com
[16:41:59] <milimetric>	 actually, so if you can log out / log in, that means it's working
[16:42:31] <milimetric>	 that hits the user table in the wikimetrics local database
[16:42:52] <andrewbogott>	 even if I log in via google?
[16:42:56] <milimetric>	 yes
[16:43:09] <andrewbogott>	 ok, well, that's simple enough :)
[16:43:21] <milimetric>	 it stores your email and uses your generated id to assign you reports / cohorts, etc.
[16:43:34] <andrewbogott>	 I will email you in case of disaster or otherwise when I finish up.  Thanks.
[16:43:35] <milimetric>	 yep, and don't worry - i backed this up
[16:43:39] <milimetric>	 just don't format the instance :)
[16:43:55] <milimetric>	 no prob, ping me on here or email
[16:44:07] <milimetric>	 I'm just relaxing at home
[16:46:09] <milimetric>	 ok, I copied the backup locally so you can feel free to format the drive too, it'd be easy to bring it back up
[16:46:39] <milimetric>	 well, maybe not "feel free" :D
[17:37:13] <jonas_agx>	 Hello, is there any instance for Spanish wiki?
[17:37:26] <jonas_agx>	 Something like the ptwiki?
[17:54:29] <YuviPanda>	 hi arrbee
[17:54:33] <arrbee>	 hey YuviPanda 
[17:54:40] <YuviPanda>	 arrbee: you should probably poke andrewbogott_afk or Ryan_Lane about getting the instance :)
[17:54:49] <YuviPanda>	 arrbee: andrew should be back in a few mins, he stepped out for lunch
[17:55:01] <arrbee>	 YuviPanda: sure thing. Thanks for the pointer. :)
[17:55:29] <YuviPanda>	 arrbee: do you know if there's already a language engineering project on labs?
[17:56:02] <arrbee>	 Umm.. I am not sure. I can find out.
[17:56:51] <arrbee>	 YuviPanda: ^^
[17:56:58] <YuviPanda>	 arrbee: ah, okay
[17:57:08] <YuviPanda>	 arrbee: if they don't, we should probably get you guys one :)
[17:57:23] <arrbee>	 YuviPanda: That would probably be nice.
[17:57:48] <YuviPanda>	 arrbee: yeah
[17:59:10] <arrbee>	 YuviPanda: We don't.
[17:59:31] <YuviPanda>	 arrbee: yeah, think we should get you one.
[17:59:35] <YuviPanda>	 should wait till andrewbogott_afk comes back
[17:59:49] <arrbee>	 YuviPanda: Sure. 
[18:21:42] <YuviPanda>	 !ping
[18:21:42] <wm-bot>	 !pong
[19:03:20] <arrbee>	 andrewbogott: Hello… I was looking to get an instance on labs to host a test case management system (for use by the language team). I am not sure what 'project' it could be a part of. Would you have any suggestions? 
[19:07:05] <YuviPanda>	 andrewbogott: if the langeng team doesn't have a project, I think we should create one
[19:07:09] <YuviPanda>	 arrbee: ^
[19:07:31] <andrewbogott>	 Sure… do you think this should be a generic 'language' project, or a specific one for this particular system?
[19:07:56] <YuviPanda>	 andrewbogott: languages sounds okay to me
[19:08:13] <YuviPanda>	 arrbee: please do tell rest of language engineering too, that they can / should use this for testing / staging :)
[19:08:13] <andrewbogott>	 arrbee, what is your username in labs?
[19:08:21] <andrewbogott>	 Um… wikitech username, I mean
[19:08:59] <arrbee>	 YuviPanda: andrewbogott : Would it be like a catch-all project for anything that we could suggest to other people as well (like GSoC students) to use for projects related to our team?
[19:09:04] <arrbee>	 andrewbogott: its 'arrbee'
[19:09:08] <YuviPanda>	 arrbee: indeed.
[19:09:20] <YuviPanda>	 arrbee: one 'project' is a unit of access control, you can grant access to different people
[19:09:22] <YuviPanda>	 and create instances
[19:09:32] <arrbee>	 YuviPanda: Okay. That would be nice. 
[19:11:07] <andrewbogott>	 arrbee, you should be all set, project name 'language'
[19:11:35] <arrbee>	 andrewbogott: Thank you. Is there an email coming in too?
[19:11:40] <andrewbogott>	 nope
[19:13:05] <arrbee>	 andrewbogott: So I should just follow whats written in this page : https://wikitech.wikimedia.org/wiki/Help:Instances
[19:13:08] <arrbee>	 or is there something more I should read up before tinkering around?
[19:13:41] <andrewbogott>	 I think that page is enough -- you'll want to think about security groups before you create an instance.
[19:14:00] <arrbee>	 andrewbogott: Okay. Shall do.
[19:14:01] <andrewbogott>	 And remember that it's always OK to just delete a failed instance and create a new one.  Don't get attached :)
[19:14:10] <arrbee>	 :)
[19:14:28] <arrbee>	 andrewbogott: YuviPanda : Thanks much for the help. Really appreciate it.
[19:14:33] <YuviPanda>	 :D
[19:14:37] <YuviPanda>	 labs <3
[19:51:11] <hedonil>	 Some instance database masters around here?
[20:07:19] <petan>	 Ryan_Lane: ping
[20:07:26] <petan>	 can you fix gluster on bots-labs?
[20:09:29] <petan>	 poor wm-bot is caching all irc to ram
[20:09:33] <petan>	 * logs
[20:09:55] <petan>	 there is still some 600mb of ram free so no rush :P but it would be nice
[20:10:06] <petan>	 or mutante can u do that?
[20:10:16] <petan>	 mutante: https://bugzilla.wikimedia.org/show_bug.cgi?id=55690
[20:27:07] <Ryan_Lane>	 petan: sure
[20:27:08] <Ryan_Lane>	 also
[20:27:11] <Ryan_Lane>	 why not switch to nfs?
[20:27:23] <Ryan_Lane>	 and... I'd prefer not to create a single-use project
[20:27:49] <Ryan_Lane>	 let's just keep the bots project open for one off bots
[20:27:59] <Ryan_Lane>	 that don't use managed infrastructure
[20:28:14] <Ryan_Lane>	 and instances can be requested when they can't work in tools
[20:28:21] <Ryan_Lane>	 also, why can't your bot work in tools?
[20:29:48] <Ryan_Lane>	 grrr. this bots-labs instance isn't running puppet
[20:33:38] * SurprisedPanda  makes joke about puppet and git taking 6 hours

[20:36:14] <Ryan_Lane>	 SurprisedPanda: 6 hours for what?
[20:36:24] <Ryan_Lane>	 puppet in labs runs in <15s when there's no changes
[20:36:35] <Ryan_Lane>	 and about 1-2 mins when there's a catalogue recompilation
[20:36:53] <SurprisedPanda>	 Ryan_Lane: that's how much time it takes to add a package to exec_environ.pp and commit it to gerrit, according to petan :P
[20:37:18] <Ryan_Lane>	 o.O
[20:37:29] <Ryan_Lane>	 maybe to get it merged and everything
[20:37:41] <SurprisedPanda>	 no, to commit it
[20:38:12] <SurprisedPanda>	 legoktm: python-dateutil is installed on tools :)
[20:38:16] <SurprisedPanda>	 as in
[20:38:18] <SurprisedPanda>	 already installed
[20:38:19] <SurprisedPanda>	 i just checked
[20:38:28] <legoktm>	 me too
[20:38:32] <legoktm>	 >>> import dateutil
[20:38:32] <legoktm>	 >>>
[20:38:35] <SurprisedPanda>	 heh
[20:38:35] <SurprisedPanda>	 yeah
[20:38:38] <SurprisedPanda>	 exact same thing i did
[20:39:27] <Ryan_Lane>	 to commit it? :D
[20:39:31] <SurprisedPanda>	 Ryan_Lane: yes!
[20:39:41] <Ryan_Lane>	 I have serious doubts there ;)
[20:39:49] <SurprisedPanda>	 Ryan_Lane: petan said it, not me! :P
[20:39:55] <Ryan_Lane>	 :)
[20:40:12] <SurprisedPanda>	 Ryan_Lane: when i was ranting about him installing packages on labs with apt-get and then not putting them in puppet
[20:42:50] <Ryan_Lane>	 -_-
[21:02:12] <hedonil>	 Ryan_lane: ping
[21:02:18] <Ryan_Lane>	 ?
[21:02:38] <hedonil>	 Having some issues on labsdb1003 - disk full
[21:02:59] <Ryan_Lane>	 Coren: ^^
[21:03:31] <Coren>	 Disk full?!
[21:03:59] <Ryan_Lane>	 there's 3 .bad directories in /a
[21:04:07] <hedonil>	 My task was cancelled, ganglia says 100% disk space used
[21:04:08] <Ryan_Lane>	 eating up 1/3 o the disk
[21:04:28] <Coren>	 Hm.  Indeed.  Those can go now, certainly.
[21:04:48] <hedonil>	 what to do?
[21:04:58] <Coren>	 hedonil: Making room now.
[21:05:59] <Coren>	 We had kept those copies out of a sense of paranoia, but by now anyone who wanted data out of them has gotten it.
[21:06:51] <hedonil>	 May I proceed with my tool? I need approx. 2 TB 
[21:07:24] <Ryan_Lane>	 2TB? :D
[21:07:28] <Coren>	 2T?
[21:07:37] <Ryan_Lane>	 that's a *lot* of data
[21:07:42] <Ryan_Lane>	 what are you putting into it? :)
[21:07:42] <Coren>	 No, you can't.  We'll need to make special arrangement for that much data.
[21:07:58] <Ryan_Lane>	 Coren: how are those hardware DBs coming along? :)
[21:08:07] <Ryan_Lane>	 they'll live in eqiad and not pmtpa, right?
[21:08:11] <Coren>	 (Not saying we can't do it, but that we'll need to set something up for you)
[21:08:17] * Ryan_Lane  doesn't want to complicate the move any more

[21:08:25] <Coren>	 Yes, eqiad -- no reasonable reason to put them in pmtpa just to move them.
[21:08:28] <Ryan_Lane>	 yep
[21:08:30] <Ryan_Lane>	 cool
[21:09:03] <Coren>	 hedonil: What do you need that much data for anyways?
[21:09:28] <hedonil>	 It's s hitcounter webstatistics
[21:09:32] <Coren>	 I'll need to talk to Sean tomorrow (our DBA) but we can probably set something up for you then.
[21:09:57] <hedonil>	 this would be great!
[21:10:15] <YuviPanda>	 Coren: are dbs in labsdb still transitional?
[21:10:27] <YuviPanda>	 if they are, hedonil's data will be / can be lost anytime
[21:10:40] <Coren>	 YuviPanda: Yes, they are.
[21:10:46] <YuviPanda>	 right
[21:11:02] <YuviPanda>	 Coren: that really should be documented  in !tools-help if it isn't already
[21:11:03] <Coren>	 hedonil: Wait, this data would be precious?
[21:11:11] <Coren>	 YuviPanda: It is, AFAIK.
[21:11:16] <YuviPanda>	 oh ok
[21:11:17] <YuviPanda>	 :D
[21:11:59] <Coren>	 hedonil: Then we really *need* to set something up for you, and dedicate some resources to it.  That'll need support from ops and an okay from wider engineering.
[21:12:24] <Coren>	 hedonil: Please to start a discussion about it on wikitech-l
[21:12:32] <hedonil>	 for now it's not that vital if it gets lost
[21:12:40] <YuviPanda>	 hedonil: plus if you're looking to query the data that powers stats.groke, IIRC analytics is working on a very similar system
[21:12:49] <YuviPanda>	 and there was a large discussion on the analytics-l list
[21:12:53] <YuviPanda>	 you might want to follow that up
[21:12:59] <YuviPanda>	 and also poke people on #wikimedia-analytics
[21:13:43] <hedonil>	 stats.groke doesn't have that detail of data available for my needs
[21:14:23] <YuviPanda>	 no, i talked about the *Raw* data that powers stats.groke
[21:14:26] <YuviPanda>	 not stats.groke itself
[21:14:36] <YuviPanda>	 anyawy, talking to them should definitely help
[21:14:37] <YuviPanda>	 yeah
[21:15:11] <hedonil>	 It's the rawdata from stats i'm importing and aggregating right now
[21:15:40] <YuviPanda>	 we're talking about the same thing
[21:16:33] <hedonil>	 if these data would already be in labs, that would be great
[21:17:17] <YuviPanda>	 true
[21:17:20] <YuviPanda>	 there's even a bug for that
[21:17:21] <YuviPanda>	 i think
[21:18:09] <hedonil>	 stats.groke is a private project? not related to wikimedia?
[21:18:44] <YuviPanda>	 not sure
[21:22:52] <hedonil>	 Coren: Can you tell me the actual physical size of the database?
[21:23:23] <Coren>	 On  that DB?  About 1.8T
[21:27:59] <Ryan_Lane>	 hedonil: it's a private project, yes
[21:28:15] <Ryan_Lane>	 we're looking at making the data accessible without stats.groke
[21:28:48] <hedonil>	 Ryan_Lane: This would be great.
[22:00:47] <wm-bot>	 !log deployment-prep hashar: made /data/project/logs group writable, it belongs to nemobis :/
[22:00:54] <labs-morebots>	 Logged the message, Master
[22:45:28] <andrewbogott>	 milimetric: I've gotten past the issue I had before, but now metrics is throwing 500s.  I'm sure that I broke it, but… can you see what I broke?
[22:56:03] <andrewbogott>	 milimetric: nevermind, I see it...