[00:08:34] <Vogone>	 rschen7754: you're a polar bear
[00:11:08] <Cyberpower678>	 Coren, ping
[00:11:42] <Cyberpower678>	 Is there no user_daily_contribs table?
[00:15:36] <a930913>	 Is there anything that can copy the contents of a file to a wiki page, whenever the file changes?
[00:20:56] <Betacommand>	 a930913: what are you trying to do?
[00:30:48] <Reedy>	 Cyberpower678: Why would there be? There isn't one in MediaWiki
[00:31:12] <Cyberpower678>	 Reedy, there was one on toolserver.
[00:31:25] <Reedy>	 There's a lot of crap on toolserver
[00:31:36] <Cyberpower678>	 It was a useful table.
[00:31:42] <Reedy>	 Why?
[00:31:49] <Cyberpower678>	 For the SUL info tool.
[00:31:49] <Reedy>	 It wasn't a table, presumably a view
[00:31:59] <Reedy>	 What does it have that the revision table doesn't?
[00:32:10] <Cyberpower678>	 Hold on...
[00:44:07] <Betacommand>	 Reedy: https://www.mediawiki.org/wiki/Extension:UserDailyContribs
[00:46:30] <Reedy>	 God we keep some crap around
[00:50:15] <Betacommand>	 Im trying to find what the most recent entry was
[00:51:39] <Reedy>	 most recent entry?
[00:51:54] <Reedy>	 Looks like it's still working fine
[00:52:42] <Betacommand>	 Reedy: I didnt see it on enwiki's extension list
[00:54:07] <Reedy>	 It's got an entry for 2014-01-02 for me
[00:54:26] <Reedy>	 https://en.wikipedia.org/wiki/Special:Version too
[00:55:09] <Betacommand>	 Ah, was looking for it without the space
[00:58:01] <a930913>	 Betacommand: Not get distracted :p
[00:58:36] <Cyberpower678>	 Reedy, nevermind.  I adapted my script to use revision
[00:58:37] <a930913>	 Betacommand: I have a userscript I'm making, that will require lots of changes, so I want to make a local file copy to the userscript page.
[01:00:22] <Betacommand>	 a930913: thats like a 4 line python script, set on a cron and poof
[01:01:51] <a930913>	 Betacommand: Ew.
[01:01:55] * a930913  slaps Betacommand.

[01:02:03] <Betacommand>	 a930913: what?
[01:02:06] <a930913>	 Polling.
[01:02:25] <a930913>	 Inotify or something.
[01:02:45] <Betacommand>	 depends on how complex you want to get
[01:06:24] <a930913>	 Betacommand: Complex enough that someone else has done it before and packaged it ;)
[01:06:43] <Betacommand>	 a930913: not to my knowledge
[01:07:06] <stwalkerster>	 any ideas why I can't "cd /data/project" on instances in the account-creation-assistance project? It just hangs until I close the ssh session.
[01:07:38] <stwalkerster>	 reboots of the instances don't help.
[01:07:59] <a930913>	 What's the lightest library to upload the file in a line or so of code?
[01:09:39] <Betacommand>	 onwiki? I suggest pywiki
[01:13:24] <anomie>	 stwalkerster: Is shared storage set up for the project?
[01:13:43] <stwalkerster>	 anomie, it worked before, so I presume so
[01:13:58] <stwalkerster>	 Yep
[01:14:16] <anomie>	 stwalkerster: I have no idea then.
[01:15:03] <stwalkerster>	 who can I poke who's online?
[01:17:20] <anomie>	 stwalkerster: All the people I might suggest you poke (andrewbogott_afk, Coren, RyanLane) don't seem to be around at the moment
[01:18:02] <stwalkerster>	 Those were the three I was looking for, I just didn't know if there were more I could poke. Thanks anyway
[10:39:59] <dalba>	 hi! I have a wsgi script that I want to run on http://tools.wmflabs.org/ but after starting the webservice I get "403 - Forbidden" error on http://tools.wmflabs.org/yadkard/yadkard.fcgi . Any idea where I might be doing wrong?
[12:28:51] <huh>	 Hi.
[12:31:21] <giftpflanze>	 oh, you are πr²
[12:31:38] <giftpflanze>	 wouldn't have guessed
[12:33:29] <huh>	 Yep, that's me.
[12:51:59] <huh>	 Is it possible to put jstart errors and output in a different directory?
[12:59:21] <Pavel_WMDE>	 anyone here who can help me loggin in to http://tools.wmflabs.org/tusc/
[13:14:11] <huh>	 giftpflanze: hi
[13:14:17] <huh>	 Is it possible to put jstart errors and output in a different directory?
[13:14:56] <Steinsplitter>	 hi Pavel_WMDE
[13:15:03] <huh>	 Pavel_WMDE has left freenode ()
[13:16:03] <Steinsplitter>	 was funktioniert genau nicht? einfach ausfüllen und so? :)
[13:16:10] <giftpflanze>	 let me see
[13:16:44] <giftpflanze>	 this tool page is wrong in so many aspects i don't want to start enumerating
[13:20:25] <giftpflanze>	 huh: if i get you right, the -e and -o options (probably also -j y) of jsub are of interest to you
[13:20:55] <huh>	 giftpflanze: is jstart the same as jsub?
[13:21:04] <huh>	 Where did you learn about those options?
[13:21:09] <huh>	 they're not in man jsub
[13:22:05] <giftpflanze>	 they're qsub options that are passed from qsub (you can read the script)
[13:22:14] <giftpflanze>	 also see https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#jsub_options
[13:23:41] <giftpflanze>	 and jsub and jstart are differentiated by script name
[13:23:52] <giftpflanze>	 so file-wise the same
[13:24:46] <giftpflanze>	 *passed from jsub
[13:25:26] <huh>	 Thanks giftpflanze 
[13:26:22] <giftpflanze>	 i love helping (when i know the answer) :)
[13:29:57] <giftpflanze>	 Coren: may i bring https://bugzilla.wikimedia.org/show_bug.cgi?id=58949 to your attention again? i want my christmas present ;)
[14:02:44] <johang>	 pagecounts dumps on dumps.wikimedia.org are not updating. who maintains that?
[14:26:29] <scfc_de>	 johang: I think the server itself is maintained by Ariel aka apergos.  Don't know where the pagecounts come from.
[14:40:06] <scfc_de>	 Is there a valid case for a service group being a member of itself?!
[14:50:05] <giftpflanze>	 scfc_de: what are you even talking about?
[14:55:47] <scfc_de>	 giftpflanze: You can add users and service groups as members to a service group.  Some service group had the service group itself as a member.  I wonder if there is a use case for that.
[14:56:54] <giftpflanze>	 aren't service group users and service groups different from each other?
[15:02:27] <andrewbogott>	 YuviPanda: Can you explain to me about urlproxy.lua vs. domainproxy.lua?  It's not clear to me that the former is used anyplace...
[15:03:24] <scfc_de`>	 giftpflanze: Yes, but service groups can be "nested".
[15:05:36] <giftpflanze>	 a-ha
[15:11:38] <hedonil>	 scfc_de: a guess - to nest your service-group into your own service-group my have been an attempt to force NFS updare it's tokens
[15:15:32] <scfc_de>	 hedonil: No, in this case the maintainer just fucked up (and removed himself from the service group in the process as well :-)).  But I'm wondering (well, very mildly :-)) if there is a valid use case.
[15:30:45] <hedonil>	 bad news for all pagecount guys, dumps dried up  Jan 5 3AM  http://dumps.wikimedia.org/other/pagecounts-raw/2014/2014-01/
[15:31:32] <hedonil>	 and even stats.grok.se stopped updating Jan 2
[15:36:25] <Cyberpower678>	 @seen petan
[15:36:25] <wm-bot>	 Cyberpower678: Last time I saw petan they were quitting the network with reason: Remote host closed the connection N/A at 1/3/2014 7:40:01 AM (2d7h56m23s ago)
[15:36:37] <Cyberpower678>	 Hmm
[15:39:10] <hedonil>	 giftpflanze: hihi,  <giftbot> regular tasks: ArchivBot hinterherräumen
[15:39:38] <hedonil>	 giftpflanze: hinterherräumen - wonderful wording
[15:40:28] <Cyberpower678>	 poisonplant. :p
[15:40:55] <Cyberpower678>	 YuviPanda, poke
[15:41:42] <giftpflanze>	 hedonil: ^^
[15:42:41] <hedonil>	 giftpflanze: what does giftbot actually in doing  hinterherräumen?
[15:43:46] <giftpflanze>	 archive when archivbot doesn't
[15:44:13] <giftpflanze>	 there are so many bugs because the bot herder doesn't fix bugs
[15:44:33] <giftpflanze>	 so i wrote my own
[15:44:36] <hedonil>	 giftpflanze: so why don't replace ArchivBot entirely if it's faulty
[15:44:45] <giftpflanze>	 i have my own bugs :)
[15:44:54] <Cyberpower678>	 Does anybody know where petan is?
[15:45:19] <giftpflanze>	 which relay on mediawiki api/parser bugs
[15:46:58] <giftpflanze>	 i also took over archiving in wikisource because he doesn't even know that it stopped working (he never looks on the bugs page/responds, it drives me crazy)
[15:47:58] <hedonil>	 giftpflanze: mantra: Never change a running bot ;)
[15:49:20] <giftpflanze>	 running being the keyword here …
[15:49:42] <hedonil>	 hehe
[15:49:50] <giftpflanze>	 when you run bots you have some responsibility
[15:50:50] <giftpflanze>	 some bugs cannot be fixed or need hilarious amounts of work but the rest … just a little time (plus the never ending infrastructure changes)
[15:52:02] <hedonil>	 yeah, the job of bot herder is never done
[15:52:35] <giftpflanze>	 which make me think of just throwing that shit away but who would do it then?
[15:55:35] <hedonil>	 giftpflanze: no, this won't be a satisfying solution, admit it - but for all that it makes fun 
[15:57:04] * hedonil  is wondering why all pagecount sources dried up. Is there something weird going on with the proxies? Or is this the end as we know it?

[15:57:15] <giftpflanze>	 hedonil: can you say that again in german so that i can understand it properly?
[15:57:53] <hedonil>	 giftpflanze: trotz allem macht es spass
[15:58:09] <giftpflanze>	 ah, ok
[15:59:22] <giftpflanze>	 i think i don't have all the time or the appetite to make any new investments into the codebase, but i will see
[16:00:39] <hedonil>	 giftpflanze: take, for example Cyberpower678. he's having so much fun runnig bots and tools. with and without BOM's :P
[16:00:51] <hedonil>	 !BOM
[16:00:51] <wm-bot>	 Did Cyberpower678 use his crappy editor again?
[16:00:57] <giftpflanze>	 hihi
[16:01:04] <hedonil>	 !BOManswer
[16:01:04] <wm-bot>	 of course he did it again!
[16:04:26] <huh>	 What text editor is that?
[16:05:11] <giftpflanze>	 notepad?
[16:08:25] <valhallasw>	 of course, this is only an issue for inferior languages such as PHP ;-)
[16:09:50] <hedonil>	 valhallasw: PHP, "inferior" - tssss
[16:10:44] <giftpflanze>	 a language that i don't want to learn …
[16:10:55] <valhallasw>	 I could also use 'laughable', 'fractal of bad design' or 'AAAAAAAAAAAAAAHHHHHHHHH *bangs head into a wall*'
[16:11:17] <valhallasw>	 (but on the other hand, I also see people making useful tools with it, so I guess it's good we have it)
[16:12:04] <hedonil>	 valhallasw: or even Python, this newfangled collection of spaces :P
[16:12:37] <valhallasw>	 hedonil: yes, indentation, which you were doing already to begin with :p
[16:14:52] <hedonil>	 valhallasw: you can make wonderful code with curly brackets - no need for endless indentation ;)
[16:15:16] <hedonil>	 valhallasw: while(true) { drink(beer); }
[16:15:49] <valhallasw>	 hedonil: you can make wonderful code in python with curly brackets!
[16:16:40] <valhallasw>	 {'result': [ {f.formatter.title: f.APIformat() for f in row } for row in self.rows ] } 
[16:16:44] <valhallasw>	 terse enough? ;-)
[16:18:58] <huh>	 hedonil: from __future__ import braces
[16:20:04] <hedonil>	 huh: hehe "import braces"
[16:23:03] <hedonil>	 Coren: whom to poke concerning this pagecount issue?
[16:23:30] <Coren>	 "pagecount issue"?
[16:24:12] <hedonil>	 Coren: dumps.wikimedia dried up, stats.groks.se doesn't update any longer.
[16:24:54] <Coren>	 Hm.  I don't rightly know whose yard this is in, but I can ask.
[16:25:13] <hedonil>	 Coren: this would be great 
[16:27:09] <giftpflanze>	 Coren: did you see my bug question?
[16:27:26] <Coren>	 giftpflanze: Not that I recall.  Which?
[16:28:00] <giftpflanze>	 Coren: may i bring https://bugzilla.wikimedia.org/show_bug.cgi?id=58949 to your attention again? i want my christmas present ;) ← this one
[16:28:30] <Coren>	 Monday.  It's #2 on my todo list.
[16:28:59] <giftpflanze>	 woohoo :)
[16:29:14] <giftpflanze>	 what is #1?
[16:30:56] <hedonil>	 giftpflanze: replace current ldap with a *really* good solution like MS ActiveDiretory :P
[16:31:56] <giftpflanze>	 ist das dein ernst?
[16:32:17] <hedonil>	 giftpflanze: hmm I think I'm just kidding ;)
[16:35:30] <valhallasw>	 Coren: would it be an idea for the web proxy to auto-gzip text/* application/json and application/xml ?
[16:39:24] <hedonil>	 giftpflanze: I would rather work on my tool than suggesting the use MS software. but I have to know if this pagecount dumps dried up just temporarily or forever.
[16:39:50] <hedonil>	 giftpflanze: if it's forever, I'm screwed and my tool is dead
[16:43:45] <hedonil>	 :|
[16:44:26] <huh>	 Never say never. Oh, wait, I just said it twice.
[16:46:12] <hedonil>	 huh: never never never never ever ;)
[16:46:16] <giftpflanze>	 well, i think it's not forever, just a hiccup
[16:47:46] <hedonil>	 giftpflanze: what really makes me pondering is that even stats.grok.se stopped updating
[17:05:24] <MrZ-man>	 hedonil: I'm pretty sure stats.grok.se uses the same data
[17:05:37] <MrZ-man>	 (and it breaks all the time)
[17:06:14] <hedonil>	 MrZ-man: I've never seen this before
[17:07:07] <MrZ-man>	 almost every other section on https://en.wikipedia.org/wiki/User_talk:Henrik is someone complaining the site stopped updating
[17:08:27] <Matthew_>	 Hmmmmm... take appears to be broken for me on matthewrbowker-dev: "vars.local.php: You need to share a group with the file" What'd I do wrong this time :/ XD
[17:08:52] <hedonil>	 MrZ-man: ah, ok
[17:10:28] <scfc_de>	 Matthew_: What's the full path?
[17:10:36] <hedonil>	 Matthew_: Cyberpower678 stated yesterday that his error appears if you want to take something that doens't exist
[17:10:40] * YuviPanda  looks around

[17:13:26] <Matthew_>	 Whoops, crashed sorry.  
[17:16:44] <hedonil>	 Matthew_: so? with correct filename and path ?
[17:16:58] <Matthew_>	 hedonil: Hmmmm?  Taking the file itself you mean?
[17:17:15] <hedonil>	 Matthew_: yeah
[17:17:24] <Matthew_>	 Nope:
[17:17:25] <Matthew_>	 local-matthewrbowker-dev@tools-login:~/public_html/vars$ take vars.local.php
[17:17:25] <Matthew_>	 vars.local.php: You need to share a group with the file
[17:20:05] <scfc_de>	 Matthew_: Apparently, the "+s" mode of the directory got lost, and so the file you uploaded/created there as a user didn't get the local-matthewrbowker-dev group.
[17:20:56] <Matthew_>	 scfc_de: That makes sense.  I wonder if that one got uploaded when "take" was broken...
[17:23:27] <scfc_de>	 Matthew_: What you can always do as a work around in such cases is, as the tool account, "cp FILE FILE.new && mv FILE.new FILE".  This isn't *exactly* what take does, but is equivalent in many cases.
[17:23:36] <scfc_de>	 Matthew_: Do you want me to fix that file?
[17:23:49] <Matthew_>	 scfc_de: please?  If not I can try...
[17:26:38] <scfc_de>	 Coren: Why doesn't chown work here: "chown local-matthewrbowker-dev /data/project/matthewrbowker-dev/public_html/vars/vars.local.php" => "chown: changing ownership of `/data/project/matthewrbowker-dev/public_html/vars/vars.local.php': Invalid argument".
[17:27:57] <Matthew_>	 o.O I'm just a heap of trouble here aren't I? XD
[17:29:09] <scfc_de>	 Matthew_: Hmmm.  On tools-login, it worked perfectly ... ?!
[17:29:36] <Matthew_>	 scfc_de: Hmmmmm o.O I'm on tools-login... so that's really weird.
[17:30:06] <scfc_de>	 Matthew_: Okay, I fixed the owner/group of the file.  I'll fix the directories next.
[17:32:41] <Matthew_>	 scfc_de: Thank you :)
[17:34:16] <scfc_de>	 Matthew_: Okay, I hope I did everything alright.
[17:34:25] <Matthew_>	 scfc_de: Ok, thanks.  Let me try it :)
[17:34:47] <Matthew_>	 scfc_de: No errors on take, so yay :)
[18:11:24] <giftpflanze>	 it takes some time until ssh keys are updated when adding them?
[18:42:59] <scfc_de>	 giftpflanze: How much time has passed?
[18:45:00] <giftpflanze>	 a few minutes 
[18:52:21] <scfc_de>	 giftpflanze: /public/keys/gifti/.ssh/authorized_keys was last modified 18:05Z, so it appears to have been changed.
[18:53:56] <giftpflanze>	 hm, weird ...
[18:54:34] <giftpflanze>	 but works in the meantime
[19:12:09] <Matthew_>	 scfc_de: Thank you again for your fix, it's working wonderfully :)
[19:13:29] <hedonil>	 anomie: Hi, I registered my tool as OAuth consumer yesterday.
[19:13:50] <hedonil>	 anomie:  your page was very helpful in understanding how this stuff works https://tools.wmflabs.org/oauth-hello-world/enduser.php
[19:14:00] <hedonil>	 anomie: thx
[19:14:28] <hedonil>	 anomie: but I have a problem adding a public key
[19:15:52] <hedonil>	 anomie: did you ever use SSH keys for that or just default shared secret?
[19:18:43] <anomie>	 hedonil: For OAuth? The hello world app uses the shared secret. There's an example of a client using an RSA key at https://www.mediawiki.org/wiki/User:CSteipp/OAuth_Training_Dec2013 (and also the commands necessary to generate the key)
[19:20:47] <hedonil>	 anomie: I'm not this far. I failed just adding the pulic key. did you every try adding one?
[19:21:30] <hedonil>	 anomie:  just converted it to required PEM format with:   ssh-keygen -f .ssh/id_rsa.pub -e -m PEM
[19:21:46] <hedonil>	 anomie: but it isn't accepted
[19:21:52] <anomie>	 hedonil: I have for some test consumers in the past.
[19:22:15] <hedonil>	 anomie: hmm
[19:22:17] <anomie>	 What does the first line of your key file look like?
[19:23:06] <hedonil>	 -----BEGIN RSA PUBLIC KEY-----
[19:23:06] <hedonil>	 4
[19:24:02] <hedonil>	 anomie: http://tools.wmflabs.org/paste/view/56fc030c
[19:38:42] <anomie>	 That was annoying.
[19:39:47] <anomie>	 hedonil: It looks like ssh-keygen formats its public keys differently from openssl, for some reason. The easiest thing to do is probably to just generate your keys with openssl: "openssl genrsa -out appkey.pem 4096" to generate the keypair and then "openssl rsa -in appkey.pem -pubout > appkey.pub" to extract the public key.
[19:40:55] <hedonil>	 anomie: thx. will try that.
[19:46:32] <hedonil>	 anomie: ok, key has been accepted ;)
[20:13:06] <huh>	 Jamesofur|away: ``Jamesofuz''?
[20:24:41] <hedonil>	 Coren: concerning pagecounts, some news from operations
[20:24:48] <hedonil>	 Coren: (21:22:38) apergos: gadolinium is down or at least unreachable, can't get to mgmt console either; that's the host that collects the pagecount files
[20:26:19] <quentinv57>	 hi, we're having some issues with the job scheduling system... sockets doesn't work if the two processes are in a different server
[20:26:27] <quentinv57>	 does anyone know how we could solve this, please ?
[20:45:21] <huh>	 Hi.
[20:50:56] <huh>	 Hello anomie.
[20:51:12] <huh>	 Is it possible to submit jobs to a specific server on tool labs?
[20:54:40] <anomie>	 huh: It's possible with something like -q task@tools-exec-01. Chances are it's not a great idea though, because then the particular node might wind up overloaded or your task wouldn't be run if that node is down.
[20:55:34] <huh>	 quentinv57: ^
[20:56:05] <huh>	 anomie: how do you request something like "tools-exec-cyberbot-02"?
[20:58:02] <anomie>	 huh: You have to be cyberbot, for one thing. It's restricted to that tool (and maybe also other of CP's tools)
[20:58:35] <JohnLewis>	 anomie: He meant something similar not a cyberbot one :)
[20:59:05] <anomie>	 JohnLewis: Oh, ok.
[20:59:46] <anomie>	 huh: Ask Coren.
[20:59:55] <quentinv57>	 anomie, the thing is that we need to have several processes running on the same machine
[21:00:33] <anomie>	 quentinv57: How tight is the coupling between them?
[21:01:19] <quentinv57>	 anomie, I don't really understand, do you mean : what is linking them ?
[21:01:44] <anomie>	 quentinv57: Basically, what do they do that they need to be on the same machine?
[21:02:38] <quentinv57>	 because there is one process reading the feeds of irc.wikimedia.org and sending it through sockets to other processes who treat them
[21:03:06] <quentinv57>	 we could basically change the socket protocol to make it work on different servers, but that's clearly a bad idea because there is a lot of feed
[21:06:05] <anomie>	 The "a lot of feed" part could justify it. Coren might have a better idea though.
[21:06:41] <scfc_de>	 There are also Redis queues; I think the Gerrit bots use them.
[21:08:05] <huh>	 anomie: is it possible to access the replica DB from Labs, but not Tool Labs?
[21:08:30] <anomie>	 huh: Yes, although I don't know the details beyond that you need to set up a hosts file.
[21:08:35] <huh>	 quentinv57: ^
[21:08:50] <scfc_de>	 huh: And you need to replicate the iptables NAT.
[21:10:13] <scfc_de>	 huh: /data/project/.system/{hosts,iptables.conf}
[21:11:00] <anomie>	 huh: Seems to have been discussed in http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-labs/20131126.txt at around 21:18
[21:11:02] <quentinv57>	 anomie, so it may be far better to move from Tools to a specific project
[21:11:21] <anomie>	 quentinv57: No idea. Ask Coren.
[21:11:39] <quentinv57>	 sure, I will, thanks
[21:13:07] <giftpflanze>	 we should document that somewhere
[21:13:29] <giftpflanze>	 not sure where
[21:14:06] * anomie  has been using Linux as his primary OS for about 15 years now, and has picked up a bit about how labs and tool labs work, but doesn't know all the details about the labs setup.

[21:14:49] <valhallasw>	 quentinv57: you can also start all the processes as a single process - e.g. write a shell wrapper that starts everything, and jsub /that/
[21:15:05] <huh>	 What if one of the bots crashes?
[21:15:19] <quentinv57>	 valhallasw, yes, I thought about that, but it's in my opinion not the best solution
[21:15:19] <huh>	 quentinv57: let's at least try it
[21:15:41] <huh>	 or no?
[21:15:52] <anomie>	 valhallasw, quentinv57: Be aware that all the processes will share the same vmem limit in that situation, though.
[21:16:09] <quentinv57>	 huh, I propose we ask Coren wheter it's better to have something like cyberbot or to start a new project
[21:16:22] <huh>	 Coren: around?
[21:16:23] <quentinv57>	 for us it's exactly the same, I just want to know what's better for Labs
[21:17:38] <anomie>	 quentinv57: I imagine the benefits to having a node rather than a project would be that you wouldn't have to sysadmin the virtual machine in addition to maintaining your software.
[21:18:21] <scfc_de>	 valhallasw: Was I correct that the Gerrit bots use some Redis queue?
[21:18:59] <valhallasw>	 scfc_de: no, gerrit-reviewer-bot uses the mail feed
[21:19:08] <valhallasw>	 and reads gmail using pop :-)
[21:19:28] <valhallasw>	 (it's older than the redis queue, otherwise I might have used that)
[21:20:34] <scfc_de>	 valhallasw: Hmmm.  But I am certain that something uses Redis for exactly these "to do" queues, i. e. one process fills them, one process (somewhere else) reads them.
[21:20:46] <quentinv57>	 anomie, sure, I agree, and I surely have no time to maintain the virtual machine
[21:21:03] * huh  is too n00bish to maintain it, maybe JohnLewis could

[21:22:04] <quentinv57>	 anyway if we can focus on maintaining the software it would be easier
[21:22:26] <anomie>	 scfc_de: Tool Labs has a redis instance, although I don't know of anything specific that uses it in that way. But pointing the enwiki rc feed firehose at it might not be the best idea ;)
[21:22:31] <quentinv57>	 anomie, how can we request to have a node for us on tools labs project ?
[21:23:00] <valhallasw>	 quentinv57: open a bug under Labs/Tool Labs
[21:23:09] <anomie>	 quentinv57: Ask Coren. If he doesn't show up before then, he'll probably be in around 14:00 UTC tomorrow.
[21:23:13] <quentinv57>	 okay, thanks :)
[21:23:30] <quentinv57>	 14 UTC, thanks, I'll be there
[21:23:45] <giftpflanze>	 scfc_de: I do
[21:24:08] <giftpflanze>	 or will do
[21:24:30] <giftpflanze>	 I already tested it
[21:25:08] <scfc_de>	 anomie: I think we never tried the firehose :-).  But http://ganglia.wmflabs.org/latest/?r=hour&cs=&ce=&m=load_one&s=by+name&c=tools&h=tools-redis&host_regex=&max_graphs=0&tab=m&vn=&sh=1&z=small&hc=4 shows some use.
[21:25:36] <huh>	 anomie: not just "enwiki rc feed", but all wikis ;)
[21:25:44] * anomie  has AnomieBOT's various tasks dump status info in redis and then uses that to generate http://tools.wmflabs.org/anomiebot/, but no actual job queuing

[21:26:04] <anomie>	 huh: True
[21:27:36] <scfc_de>	 Even if it were all wikis -- it should be even more performant to provide one queue instead of x tools each implementing their own IRC parsers.
[21:28:09] <scfc_de>	 What's the general status of Wikimedia rc feeds?  Wasn't there talk of ... MQ or something like that?
[21:28:41] <valhallasw>	 yeah, boys and girls with their fancy toys
[21:29:19] <valhallasw>	 the nice thing of the IRC feeds is that is has been working for the last 10-or-so years ;-)
[21:29:54] <valhallasw>	 it's not perfect, but you know exactly what to expect, and it's not going to disappear suddenly because someone doesn't maintain their toolset anymore
[21:30:20] <huh>	 It does have some problems that make it a pain to use, however.
[21:30:36] <huh>	 For example the block log entry is localized by some MediaWiki message.
[21:30:51] <huh>	 It could be hard to parse depending on what is in that message on that wiki.
[21:31:04] <valhallasw>	 Very true.
[21:32:20] <giftpflanze>	 Also the IRC line limit is … limiting
[21:32:33] <scfc_de>	 valhallasw: Sure.  But if parsing it means a node per tool ... :-)
[21:33:42] <giftpflanze>	 I'd appreciate a centralised solution
[21:34:12] <giftpflanze>	 it means I don't have to implement pinging the server
[21:34:43] <giftpflanze>	 which causes my tools to break
[21:40:11] <scfc_de>	 Ah!  It was YuviPanda who uses Redis for Gerrit queues: https://github.com/yuvipanda/gerrit-to-redis.  Hmmm.  The not existing access control could be tricky.
[21:41:22] <Mouda>	 Hi All
[21:41:38] <Mouda>	 any one here 
[21:41:42] <Mouda>	 I need help
[21:41:42] <wm-bot>	 Hi Mouda, just ask! There is no need to ask if you can ask
[21:42:06] <valhallasw>	 scfc_de: also streaming events over SSH is a PITA
[21:42:20] <valhallasw>	 scfc_de: at least, if you want to get *all* events
[21:42:26] <Mouda>	 please I want to connect to tools-login server
[21:42:34] <Mouda>	 how can I get access to it
[21:42:53] <valhallasw>	 scfc_de: the SSH connection sometimes hangs for multiple minutes (or even completely, until you reconnect), which means you miss several minutes of events.
[21:43:13] <valhallasw>	 scfc_de: which is fine for an IRC bot, but not for the reviewer bot :-)
[21:43:13] <scfc_de>	 valhallasw: Why SSH?  Or do you mean for Gerrit?
[21:43:25] <valhallasw>	 scfc_de: gerrit-to-redis reads events from SSH
[21:43:45] <scfc_de>	 Mouda: Are you using Linux or Windows?
[21:43:49] <Mouda>	 windows
[21:44:17] <Mouda>	 I connect to bastion but can't open tools-login
[21:44:42] <scfc_de>	 Mouda: Can you connect to tools-login.wmflabs.org?
[21:45:48] <Mouda>	 the sever refused my private key
[21:46:04] <scfc_de>	 Mouda: Ah, I see, you haven't requested access to the Tools project yet.  One moment, please.
[21:46:19] <Mouda>	 ok
[21:46:29] <Mouda>	 please from where I can request it
[21:47:22] <scfc_de>	 You just did :-).  Try again, please.
[21:49:03] <Mouda>	 wow working now
[21:49:09] <Mouda>	 thank u sooooo much
[21:49:40] <scfc_de>	 Mouda: No problem.
[21:57:23] <huh>	 Coren: ping
[21:58:10] <Coren>	 huh: Semi-pong?  (Limited availability)
[21:59:12] <huh>	 Coren, when you're back, could you take a look at https://bugzilla.wikimedia.org/59697?
[22:02:18] <CP678|iPhone>	 @seen petan
[22:02:18] <wm-bot>	 CP678|iPhone: Last time I saw petan they were quitting the network with reason: Remote host closed the connection N/A at 1/3/2014 7:40:01 AM (2d14h22m17s ago)
[22:02:55] <CP678|iPhone>	 zz_YuviPanda: ping
[22:03:25] <CP678|iPhone>	 Rats.
[22:08:04] <quentinv57>	 hey Coren, do you have a minute, please ?
[22:08:35] <Coren>	 quentinv57: More or less.  Do tell.
[22:09:13] <Mouda>	 hello
[22:09:20] <Mouda>	 I have another problem
[22:09:28] <Mouda>	 when I want to run a toll
[22:09:33] <Mouda>	 it tells me 
[22:09:35] <Mouda>	 mouda is not allowed to run sudo on tools-login.
[22:09:48] <huh>	 Mouda: did you just create it?
[22:09:57] <huh>	 You should use ``become nameofthing''
[22:10:01] <quentinv57>	 Coren, we filed a bug (https://bugzilla.wikimedia.org/show_bug.cgi?id=59697) to request a new node : before that I wanted to know with you if it's possible and if it's really the best solution
[22:10:16] <huh>	 If you just created it, you need to log out and back in before "becoming" that tool account
[22:10:18] <Mouda>	 yes I created it on http://tools.wmflabs.org/
[22:10:28] <huh>	 Mouda: see my last post
[22:10:36] <Mouda>	 aha ok i will
[22:10:50] <Coren>	 quentinv57: It probably is.  I'll comment on that bug in more detail Monday.
[22:10:50] <quentinv57>	 Mouda, did you just create it some minutes ago ?
[22:10:57] <Mouda>	 yes
[22:11:20] <quentinv57>	 Coren, okay, thanks (it's not urgent, we can wait)
[22:11:43] <quentinv57>	 Mouda, I think it may take some minutes, try in an hour, the problem may solve itself ;-)
[22:12:10] <Mouda>	 ok thank u soo much
[22:12:23] <huh>	 If you are seeing "a password is required" message when you try to become your tool (i.e., sudo yourtoolaccount), it is likely because you were logged in to your Labs account when you created the tool account. Unix group membership is checked on login only, so an existing session will not have access to the new tool group. Log out and then log in to your Labs account again to fix this problem.
[22:12:38] <huh>	 http://tools.wmflabs.org/?Help
[22:12:42] <huh>	 Mouda: ^ ^ try logging out ;)
[22:15:19] <huh>	 Mouda: does it work now?
[22:24:56] <Mouda>	 yes working now
[22:25:03] <Mouda>	 thank u so much