[08:16:15] !ping [08:16:15] pong [09:01:13] hi! Can I use Labs to create/run Wikepedia bots, or to create http wiki tools ? [09:01:18] yes [09:02:13] nice) [09:02:21] !bots [09:02:22] http://www.mediawiki.org/wiki/Wikimedia_Labs/Create_a_bot_running_infrastructure proposal for bots [09:02:33] !project bots [09:02:33] https://labsconsole.wikimedia.org/wiki/Nova_Resource:bots [09:02:46] thanks [12:36:38] !log wikidata-dev wikidata-dev-3: installed SpamBlacklist extension [12:38:34] !log wikidata-dev wikidata-dev-3: installed SpamBlacklist extension [12:41:56] Why does the above not log any more? :´( [12:48:33] log bot is down... [12:48:57] tbh it's not worth of restart, it will crash anyway in 5 min [12:53:07] ok [13:17:43] Damianz: hi, you want to fix the logbot :)0 [13:18:02] !log is I am broken, and I can't log anything, please fix me [13:18:09] @infobot-ignore- log [13:18:09] Item log was removed from list [13:18:12] !log is I am broken, and I can't log anything, please fix me [13:18:12] Key was added [13:18:15] !log blah [13:18:16] I am broken, and I can't log anything, please fix me [13:18:22] :D [15:02:25] Hi! [15:03:39] hi Silke_WMDE! [15:06:00] A question concerning Apache rewrite rules: For Wikidata we don't have a whole farm on our test system, but one repo installation and one client installation. We would now like to pretend we had separate wikidata instances for the languages. For the rewrite rules, I'm wondering if I can define in apache "these following things are language codes". This could be macros, variables or something like that. Has anybody used anything [15:06:00] similar and could point me to documentation or give me an example how you did it??? [15:09:04] My mind just exploded [15:13:34] Hello all! [15:13:38] hi dschwen [15:13:56] "Please ask a Nova administrator to create credentials for you." [15:14:01] who would that be? [15:14:03] :-) [15:14:25] I cannot sudo in my instance (wikiminiatlas) [15:14:42] the reason might be that I have no clue what my password is [15:14:56] since all outher authentication is key based [15:15:33] Silke_WMDE: you might want to write labs-l [15:15:48] dschwen: wow, there is really 0 docs on labsconsole about that, huh [15:15:59] dschwen: It's what you login to labsconsole with [15:16:06] do I detect a hint of sarcasm? ;-) [15:16:09] and yeah, we really need to fix docs because they're crap [15:16:20] sumanah: Ok, thx! [15:16:37] Especially now account != shell access, which people apparently find confusing heh [15:16:39] Damianz: how would I find out who the Nova administrators are, via labsconsole? [15:17:11] cloudadmin group [15:17:13] though [15:17:16] you shouldn't see that [15:17:26] keystone is using the ldap backend so if you can login you have credentials [15:17:38] might be the stupid stupid but that drops your domain though [15:18:02] Damianz: so, what would we advise Labs users to do, to figure out who Nova admins are and contact them? [15:19:13] dschwen: right now Labs is highly unstable and only suitable for use by people who don't mind the ground shifting under them every 2 weeks :) [15:19:54] 1sec I'm looking at my dev version to see where that's actually thrown [15:19:57] sumnah, yeah, I know [15:20:22] but since the death of the toolserver was basically just announced I'd rather get myself familiar with labs sooner than later [15:20:25] btw dschwen what IRC client do you use? [15:20:31] xchat [15:20:33] why? [15:20:35] Toolserver is shite [15:20:43] do you know about tab-completion, dschwen? in IRC? [15:20:45] 09/25/2012 - 15:20:45 - User dschwen may have been modified in LDAP or locally, updating key in project(s): bastion,maps [15:20:47] http was down for like 10hours the other day [15:20:56] 09/25/2012 - 15:20:56 - Updating keys for dschwen at /export/keys/dschwen [15:20:58] toolserver works for my projects NOW, labs does not work for my projects NOW [15:21:04] that is all i care about NOW [15:21:20] bashing the services of the TS does not help anyone [15:21:34] The only thing labs doesn't work for me currently is sql replicas, the biggest outage to the bots I run is the TS going down heh [15:22:02] ^ regarding toolserver death/blah, please can I get a project created for ACC? I'd like to start thinking about migration at some point in the future. [15:22:02] I'm with dschwen - bashing what TS provides a lot of users isn't helping anything [15:22:04] I need SQL replicals for almest every project I run [15:22:51] I am looking forward to labs. The infrastructure sounds impressive. Unfortunately it is not ready for me. [15:23:38] ok wtf [15:23:40] * Damianz stab [15:23:52] dschwen: If you logout and back in does it work? [15:23:53] This question probably has been asked over and over again: [15:24:00] That error is when it can't get your ldap details -.- [15:24:08] I run 3 bots, and about 5 tools [15:24:29] should I get 8,6,2 or 1 instances? [15:24:51] 1 or use the 'shared' bots setup (which we need to work on a bit) [15:24:57] (one for bots, one for tools; one per tool, one per bot (a common bot instance) etc.) [15:25:04] We have 2 sql servers, 1 web and like 3 tools [15:25:09] Depends how much recourses you need really [15:25:13] sumanah: That's a stupid error message [15:25:15] Damianz: I can log in perfectly fine [15:25:36] I just sudo'ed after resetting my labsconsole password, which I ^H^H^H my browser forgot [15:25:43] Basically if they are not logged in or it can't get their ldap details (silly sessions bug) then that's the error... which is not really useful, it should redirect to login on not logged in. [15:26:01] Damianz: what specifically is the error message that is dumb? "please see a Nova admin"? [15:26:07] yeah [15:26:24] 'to create credentials' might really mean 'login' [15:26:44] I'll look at doing something more logical with this later hmm, or poke Ryan_Lane to see what he thinks/check I'm reading this right. [15:26:51] dschwen: I am (sort of) a nova admin… let me catch up on the backscroll and see what your problem is. [15:27:20] thanks andrewbogott, it seems my problem was only a perceived one [15:27:35] I'm happily updating and istalling packages on my instance now [15:27:39] Really? Are you able to sudo now? [15:27:46] andrewbogott: Do you think the current use of SpecialNova::noCredentials() is silly considering the context is ! $this->userLDAP->exists() or !$this->getUser()->isLoggedIn()? [15:28:26] I am, sudo works. I just did not know which password to use (and had forgotten my labsconsole password anyhow. just resetted it and it worked immediately) [15:28:38] Damianz: Seems like it should direct you to a login/create account page [15:28:42] yeah [15:28:57] dschwen: OK, great. [15:29:32] sumanah, yes, tabcompletion works fine, didn't know it existed in xchat. thanks! [15:29:39] dschwen: it's great! [15:29:42] Damianz, what action is landing you on that page? I don't see it much. [15:29:52] Interesting there is noCredentials and notLoggedIn yet the later is called for in some places. [15:30:11] "Please ask a Nova administrator to create credentials for you." [15:30:16] ^ I'm just reading the code atm. [15:30:59] Damianz, one sec [15:31:01] Considering keystone is ldap backed that message is bleh, it should always be re-auth [15:31:22] main page, manage projects [15:31:28] https://labsconsole.wikimedia.org/wiki/Special:NovaProject [15:32:24] https://labsconsole.wikimedia.org/wiki/Help:Access#Rights [15:32:26] See, that should say login required... but I think if the silly I'm going to drop your session key bug is in effect it sees you logged in then fails to get your details to auth against nova or such.. hmm [15:32:34] the link "Manage Sudo policies"! [15:32:50] I get "You must be logged in to perform this action. " So y'all must be falling into a boundary case. [15:32:57] cause, you know, of course I'm searching the help pages first [15:33:00] ;-) [15:33:22] We don't just have a manual to tell you to RTFM :) [15:33:39] Damianz: The "I'm going to drop your session key bug" is believed to be fixed… so if you can make it happen Ryan might be interested. [15:33:45] That or my knowledge is out of date. [15:33:48] Really? [15:33:54] I thought it was mostly fixed [15:34:07] Apparently if you login on another browser mw core can only store 1 token so drops the second machine [15:34:16] Or such.. I'll try when I go over to my desktop later. [15:37:26] ssmollett: I don't think I helped much on the bz ticket, his response makes me more confused than before and provides not much info. [15:42:38] some of my tools need webserver access (i.e. a permanent portforward to bastion, or a public IP (not necessary yet I guess))) [15:42:58] the docs just mention a portforward to my localhost [15:43:09] which of course only lets me access the tool [15:43:18] is there a policy on how to do this? [15:43:47] For bots we have http://bots.wmflabs.org/~ accessible in /mnt/public_html/, or you can request a public ip [15:44:01] oh, neat [15:44:12] how do I get on the bots project? [15:45:44] 09/25/2012 - 15:45:44 - User tanvir may have been modified in LDAP or locally, updating key in project(s): bots,bastion [15:45:49] ... [15:45:55] 09/25/2012 - 15:45:55 - Updating keys for tanvir at /export/keys/tanvir [15:46:00] * Damianz hopes tanvir is dschwen's shell name [15:46:10] * Tanvir is Tanvir. [15:46:14] nope, not that I know [15:46:19] oh [15:46:20] hi Tanvir [15:46:23] I was under the impression it was dschwen [15:46:26] Hello Damianz! [15:46:27] * Damianz waits for the *next* cron run [15:46:37] dschwen: You might have access when labs-home-wm catches up [15:46:38] I am trying to login.. dunno what's happeing. [15:46:41] occam's razor, dude! [15:46:51] I did not login there for a long time. [15:47:14] labs-home-wm is a little buggy, going to DIAF soon though [15:47:49] Okay, I am logged in now! xD [15:48:06] well, great for you ;-) [15:48:07] Hi hashy [15:48:23] Can I access to project database from WMFLabs? [15:48:32] mysql? sure [15:48:36] Aye. [15:48:41] How can I do that? [15:48:57] what username do you want? [15:49:03] is tanvir asking about mysql, or the replicated project dbs? [15:49:15] In TS, I just type sql -r and I am not the respective DB. [15:49:58] If you're after the replicas of prod then we don't have them setup yet, if you just want a mysql db then we can sort that [15:50:05] * Damianz thinks we really need to backup these servers also [15:50:34] Damianz, the mysql db would be fine. [15:50:46] 09/25/2012 - 15:50:45 - Created a home directory for dschwen in project(s): bots [15:52:42] Tanvir: I just sent you some details via pm [15:52:47] bots.wmflabs.org/phpmyadmin exists if you require it [15:53:43] * Damianz needs to go make some phone calls, bbiab [15:53:51] * dschwen cannot log into bots yet. Waiting patiently... [15:55:43] 09/25/2012 - 15:55:43 - User dschwen may have been modified in LDAP or locally, updating key in project(s): bots [16:03:38] bots still denies my login attempts "ssh_exchange_identification: Connection closed by remote host" [16:14:36] https://labsconsole.wikimedia.org/wiki/Help:Access#Accessing_services_using_port_forwarding [16:14:42] this does not work for me [16:15:01] trying to connect to my instance from bastion already fails: [16:15:04] telnet 10.4.0.234 80 [16:15:04] Trying 10.4.0.234... [16:15:04] telnet: Unable to connect to remote host: Connection timed out [16:15:28] on my instance however I see the http server running and listening on 0.0.0.0:80 [16:15:35] (not just on localhost!) [16:16:09] is there a firewall setting I need to change? [16:17:06] haha [16:17:14] "Manage security groups" [16:17:39] dschwen: You can probably do a double ssh tunnel. I.e. ssh tunnel localhost:80 to bastion [16:17:40] there is the "No Nova credentials" thiong again [16:18:09] that is an ugly hack on top of an ugly hack ;-) [16:18:43] It was how I managed to get tile-mil tunneled out. But yes, it isn't directly elegant [16:19:25] actually with the ssh proxy command set up I can just tunnel straight to my instance [16:19:44] I was under the impression, you can only change the firewall settings on instance creation and I didn't want to restart everything again. Not sure if that is (still) true [16:22:15] anyhow, I changed the help page accordingly [16:22:50] Damianz: no luck with logging in to bots yet [16:36:59] i'd like to work on getting replication of prod db to labs, is that possible or has it to be done by a more priviledged person? [16:41:26] why can't I ssh to the toolserver from my labs instance? [16:47:42] What exactly are "Nova credentials" on Labs? [16:47:56] giftpflanze: Needs an ops person to do it as you need access to the prod cluster to setup replication, I believe it's in the pipeline and possible has some servers earmarked it just needs time. Larger priorities like killing nfs before it causes issues currently [16:48:30] dschwen: Security groups possibly [16:48:41] Avic: Nova is the openstack compute service that provides vms. [16:48:57] apmon: You can only change the groups that an instance belongs to on instance creation. I believe you can change the meaning of those groups at any time. [16:49:44] * Damianz hates muscles, stupid things have to hurt [16:49:50] Damianz: Ah. So, how does one get Nova credentials? do I need to have/be a member of a project, I assume? [16:49:52] giftpflanze: We're planning on having the replication db live on native hardware (vs. on a VM like most of labs). So doing that most likely requires root on real hardware. [16:50:14] giftpflanze: Nevertheless, we might be able to make use of your assistance :) Ping Ryan_Lane when he shows up. [16:50:19] Avic: Well it's sorta complicated, because it's not really a credential (it uses ldap auth like the wiki). Either login or logout and login and it should be ok [16:50:19] OK, but that still required foresight to create a security group in the first place which I didn't... [16:50:20] and/or send an email to the labs list. [16:50:38] but for the future that is good to know [16:51:06] andrewbogott: We really need real hardware for user dbs too :P Meh I'd love to get a proxy setup for domain -> internal ip in nova as well... too much fun to have [16:51:18] apmon: Yes, the current setup has poor defaults. I've tried to document it, but documentation is a poor substitute for proper default behavior. [16:51:57] Damianz: loging out/back in fixed it. Weird. Thanks for that tidbit, though! [16:51:58] Damianz: Yep, the IP proxy thing is on our list; that's going to be a topic at the OpenStack conference next month I believe. [16:52:13] Avic: Do you use multiple browsers by any chance? [16:52:20] There's an annoying bug in core that causes us issues [16:52:22] Yep [16:52:40] I go between Opera/Chrome a lot [16:52:43] Basically if you login to another you have to logout and back in one the first because it can only store 1 token. [16:52:56] andrewbogott: :o [16:53:38] Damianz: It takes a LOOONG time for features to make it from OpenStack Design to labs, though, since there's a six month release cycle and then another six month adoption lag. [16:53:49] I saw the python implimentation of managing haproxy that looked cool, would be really awesome as a service managed though the api rather than hacked up with puppet/nginx. Then iterations of ssl support with dedicated ips etc would be nice :D [16:54:09] andrewbogott: It's ok, we're in beta [16:54:09] :P [16:54:25] apmon: are you a net admin on the maps project? [16:54:35] could you allow outgoing ssh connections, please? [16:54:39] I think I might be [16:54:53] 6month release cycle for something the size of openstack is rather quick imo [16:55:07] Considering the amount they could break every time, we just hope they're on the ball. [16:55:21] Damianz: no luck loging into bots. "channel 0: open failed: administratively prohibited: open failed" [16:55:44] Hmm [16:56:15] That's an intesting message, are you just trying to ssh in or access http or ? [16:56:24] ssh [16:56:33] with the proxy-command setup [16:56:47] Damianz: Well, historically, we break every single thing in every release. [16:56:54] Ah proxy explains that message, I swear I saw the bot make you a dir [16:56:55] So it's not /that/ impressive :( [16:56:58] oh, hold on! [16:57:01] andrewbogott: Oh there is that [16:57:10] But that's what minor releases are for! [16:57:18] dschwen@bastion1:~$ ssh bots -> Name or service not found [16:57:22] Bit like ubuntu, wait until x.x.03 and everything is fine [16:57:23] am I retarded?! [16:57:33] dschwen: Ah ok, it's bots-1 bots-2 bots-3, probably use 3 [16:57:40] aHA! [16:58:15] Tools labs hopefully will be easier if we actually impliment a schedular and puppetize the tools. [16:58:16] and bots-4 :p [16:58:33] oh lord, puppetize? [16:58:43] yeat, THAT will be a lot easier [16:59:20] giftpflanze: IT'S THE FUTURE [16:59:42] I can get on bots-1 [17:00:04] ah, -3 works too now [17:00:07] Yes, puppetization is I think (one of) the reason why the wikimedia-osm server isn't yet up and running [17:00:35] aude: btw, do you know what the status on that is? [17:00:36] Heh, beta is up - it's mostly not puppetized (getting there though!) [17:00:48] and again no outbound ssh from bots-3 either [17:01:33] What port? [17:01:38] I can get ssh on port 22 fine [17:01:38] 22 [17:01:44] try telneting it [17:01:46] ssh: connect to host login.toolserver.org port 22: Network is unreachable [17:01:49] might be some weird acl [17:02:01] telnet: Unable to connect to remote host: Network is unreachable [17:02:06] hi apmon [17:02:09] andrewbogott: Is the default security group "per project"? I.e. can I change the firewall settings for the default group? [17:02:22] I'm pretty sure it's per project. [17:02:27] * aude has been busy with wikidata but needs to spend more time on puppet [17:02:34] I bet that when you dig into the config page that will be obvious. [17:02:36] hmm [17:03:05] Weird [17:03:25] aude: I could probably get the puppet setup for mod_tile and renderd if that is a help. But couldn't quite figure out how to puppetize the postgresql / postgis setup [17:03:38] i was running into a problem running puppetd -tv [17:03:56] err: Could not retrieve catalog from remote server: getaddrinfo: Name or service not known [17:03:59] warning: Not using cache on failed catalog [17:04:00] Ah, I think I got further than that... ;-) [17:04:01] err: Could not retrieve catalog; skipping run [17:04:10] and then stuck :( [17:04:32] dunno [17:04:34] not sure if it's a problem with the instance or me doing something wrong? [17:04:39] Seems TS specific, I can login to other servers fine. [17:04:52] * Damianz thinks we should move apache st00f to project data soon [17:04:52] I did manage to install postgresql with puppet. But setting up the correct users and running the scripts to initialise the db and the likes wasn't ideponent [17:04:59] I sense a labs vs. toolserver rivalery ;-) [17:05:21] apmon: if i could get past the error, i can probably figure out something and get it into gerrit [17:05:26] That reminds me [17:05:32] petan: you around, by chance? [17:05:47] rivalery? use what works. toolserver 40% of the time, labs 90% of the time :P [17:06:09] anyone know where on https://gerrit.wikimedia.org/r/#/admin/projects/ we could have a project for osm puppet stuff? [17:06:16] like what we'd call our repository? [17:06:21] aude: I think I had that error as well to start with, but can't quite remember how I solved it. [17:06:32] apmon: ah [17:06:51] It might have involved just creating a new instance and deleting the old one [17:06:53] * aude just needs the time to ask and my time usually doesn't correspond with wmf staff [17:06:59] apmon: probably that's it [17:07:10] i had trouble with the instance to start with [17:07:10] I think you can pretty much name it anything reasonable [17:07:31] like operations/openstreetmap ? or operation/maps [17:07:51] not "osm" because that's confusing with openstackmanager :D [17:07:55] Is it related to operations? [17:08:01] Damianz: yes [17:08:07] then sure [17:08:12] ^demon can always say no [17:08:33] the mobile folks want production map tiles so we don't rely on a third-party [17:10:26] Given the situation on the toolserver, it is probably good to move the production tiles for {de,ru,it,es,...}wiki off of toolserver and onto a WMF server as well [17:11:15] Probably a good idea [17:11:43] and as if I understand it correctly WMF actually already has dedicated hardware for it (at least for mobile), it would be good to be able to use that as soon as possible [17:11:49] Though imo labs is only another excuse for not funding TS more, I remember when they bitched over hardware last time. Dab does have a point though [17:12:11] Also WTF someone is drilling into a wall with a hammer action down stairs, my wall is shaking [17:12:25] apmon: agree [17:13:18] although not sure how much labs can be an exact replacement for toolserver in terms of functionality for maps stuff [17:14:13] The wikimedia mobile server is not part of labs though? [17:14:35] no [17:14:53] but we start with labs, as a beta, testing and experimental place [17:15:24] with the toolserver, people always are experimenting with stuff and that's still important [17:15:59] apmon: https://www.mediawiki.org/wiki/Git/New_repositories/Requests [17:16:04] put a request at the bottom [17:16:27] when i get puppet working, i can put stuff in there [17:16:44] And you have to remember current labs isn't really tool labs, there's sorta plans and goals for that and running 'production' stuff while being able to develop on it in roughly the same but different envs. [17:17:16] Damianz: right [17:17:32] aude: You can use the maps-test2 instance if you want. There puppet works [17:17:41] apmon: i can try it and see if it works [17:17:49] i think i just need to recreate my instance [17:17:51] although I have already created a postgresql.pp and osmtileserver.pp in /etc/puppet/manifest [17:17:55] ok [17:19:00] You probably want to comment those out though and start fresh. [17:19:29] ok [17:19:38] i'll just see if i can run puppetd [17:19:47] Oh, the other thing I was trying that didn't work if I remember correctly was getting the custom debian repository going [17:19:53] ah, yes [17:20:40] As that will be necessary to package osm2pgsql, mod_tile and renderd (and possibly others like mapnik) [17:20:47] yes [17:21:11] and we can do mod_tile and renderd to start [17:21:26] but i'd be interested in trying different rendering tools and see how they do [17:21:31] (if time allows) [17:23:00] Its possible. Although given that I am sort of the upstream maintainer of mod_tile and renderd, those are the ones I am most familiar with and can hack and extend if necessary [17:23:36] ah, okay [17:24:24] although depending on needs, mod_tile together with tirex might be the better option than renderd [17:24:31] Sadtimes, openstreetmap says my house is lidl [17:24:35] simple to start is ok [17:24:42] Damianz: fix it! [17:24:55] aude: We'll it's sorta correct, but lidl is downstairs :P [17:25:02] ah, okay :) [17:25:07] convenient [17:25:12] Indeed [17:25:17] hi Ryan_Lane [17:25:31] * Damianz hands Ryan a cookie [17:25:41] Damianz: we use cron to schedule bots on labs? [17:26:05] dschwen: i think so [17:26:05] dschoon: Yes [17:26:06] Damianz: 3D mapping (i.e. multi story buildings) is not well established yet [17:26:19] unfortunately [17:26:19] buh? [17:26:21] We really need a distributed schedular that uses multiple nodes, but that's a lot of work [17:26:35] * aude tried mapping my local airport, such as the baggage level and departures :) [17:26:46] apmon: I totally should be able to walk up my stairs, we just need to send people around with rPis and webcams :P [17:27:18] howdy [17:27:27] I do rather like http://stations.aeracode.org/ [17:28:05] * Damianz <3 webgl [17:29:30] Are those complete? I always had Oxford Circus more complicated in remembrence. It nearly looks sane on that map... ;-) [17:30:00] alright, i should go home but be back in a bit [17:30:25] Some of them aren't, the notes say what's missing roughly. They're from memory/observion rather than drawings though. [17:30:42] Still pretty cool and would be interesting to have the full set, done 100% to the drawings. [17:32:00] Yes, it is cool. I always wondered what they would look like on a 3D map while walking through those tunnels [17:34:00] Damianz is there a mechanism on bots that allows for multi maintainer projects? [17:35:44] Roughly the idea is to use /data/project which is shared betwean all instances for development, the actual running etc is a little harder due to the lack of schedular/control etc, so depending what user it's running under you have to do a bit of messing about (or be really awesome and make an init script). [17:37:20] could you mkdir /mnt/public_html/dschwen please [17:38:17] and could we get git on bots? [17:39:17] dschwen: can you still not log into bots? [17:39:54] I'm on bots, I hav just had my first successful bot run [17:40:50] great [17:41:16] Damianz: the "No credentials for your account" thing was reintroduced with keystone [17:41:20] on new account creations [17:41:36] dschwen: you said you saw that error? [17:41:46] Ryan_Lane: Just so you know, I hate you. [17:41:47] was it after changing your preferences? [17:41:49] logging out and back in on labsconsole made it hgo away for me [17:41:53] Damianz: heh. why's that? [17:41:53] Ok, now that's out, I love you again. [17:41:58] Stupid bugs [17:42:12] the credentials bug? [17:42:22] well, I know why it occurs on account creation [17:42:24] mhm, really I hate mw core devs but they're grumpy :P [17:42:42] and I also know why it happens on preference chanegs [17:42:44] dschoon: Yes, and yes, I might move the dir as well though so let me finish reading this stupid bug report [17:42:46] actually, let me fix that right now [17:42:54] ... [17:42:59] https://bugs.launchpad.net/ubuntu/+source/gnome-terminal/+bug/1055766 < This is why I hate ubuntu a little [17:43:00] Why what? [17:43:18] I think maybe you are talking to dschwen [17:43:20] heh [17:43:22] Ok, does anyone know why sshing to login.toolserver.org gives a network is unreachable? Other hosts work. [17:43:35] dschoon: Oh damn you starting your nick with dsc :P Yeah I meant dschwen [17:43:41] dschwen: from labs? [17:43:45] yes [17:43:48] I think it's firewall on their side [17:43:51] hehe [17:43:55] from my wikiminiatlas instance in the maps project [17:43:59] I can ssh out to other servers fine and can ping/traceroute it find [17:44:04] Just TS it fails on. [17:44:05] well, it's possible that we block ssh access into it [17:44:08] yeah yeah, i know [17:44:18] the evil toolserver? ;-) [17:44:26] we block ssh access to our production networks [17:44:34] it's possible that TS is also accidentally blocked [17:45:01] ok, it would be convinient to have access so that I can copy my stuff over easily [17:45:08] Let's blame lesslie, she isn't here :P [17:45:27] Yes, damn you, leslie [17:45:30] who is she? [17:45:35] heh [17:46:22] http://meta.wikimedia.org/wiki/User:LeslieCarr < [17:46:41] Still want her hair :( [17:47:21] Change on 12mediawiki a page Wikimedia Labs/Toolserver features wanted in Tool Labs was modified, changed by Platonides link https://www.mediawiki.org/w/index.php?diff=587101 edit summary: [17:48:08] Platonides++ [17:49:09] oh, right, yeah, her amount of hair would be great on my head [17:49:44] Though I'd probably be in favour of abstracting that hmm, I guess if we used gluster to store the data then scaling out to multiple web nodes later would be easy. [17:49:51] or were you referring to her actual hair, to make a little hair-doll and keep it under your pillow... [17:50:01] ...that went creepy fast [17:50:05] o.0 [17:50:13] * Damianz checks his voodoo dolls are still there [17:50:40]