[01:31:37] Reviewer requested changes to revision D4. [04:16:38] jeremyb: !log in this channel did not update https://labsconsole.wikimedia.org/wiki/Nova_Resource:Deployment-prep [08:49:32] Reedy: yay [08:49:40] did you forget to branch Mobile Frontend? [08:49:50] because it's missing in 1.19 branch [08:50:55] !log deployment-prep removing some extensions from config which are missing in latest branch [08:51:00] eh [08:54:08] !log deployment-prep removing some extensions from config which are missing in latest branch [08:54:15] Logged the message, Master [08:54:21] hi Ryan_Lane [08:54:23] :o [08:54:27] yay [08:54:32] Reedy: around? [08:55:58] Ryan_Lane: logbot is disconnecting for some reason [09:00:48] unable to login currently [09:01:06] !accountreq [09:01:06] in case you want to have an account on labs, please contact someone who is in charge of doing that: Ryan.Lane, m.utante or ssmolle.tt [09:32:19] hi Nikerabbit [09:32:32] what is wrong with login? [09:42:45] try it [09:42:59] Fatal error: require_once() [function.require]: Failed opening required '/usr/local/apache/common/live/extensions/MobileFrontend/MobileFrontend.php' (include_path='/usr/local/apache/common-local/live/extensions/OggHandler/PEAR/File_Ogg:.:/usr/share/php:/usr/share/pear') in /usr/local/apache/common-local/wmf-config/CommonSettings.php on line 2107 [09:44:12] I know [09:44:19] because wmf1 is broken [09:44:27] I switch back to trunk [09:46:15] it's working again [09:46:37] wmf1 is totally broken [09:46:48] it's not just extensions, there is problem with Vector too [09:47:29] Nikerabbit: am I allowed to commit to wmf1 or it must be done by someone from wmf only [09:59:19] petan|wk: if you don't know you probably can't commit to wmf branches [09:59:19] with wmf1 do you mean 1.19 or 1.18? [10:00:51] 1.19 [10:07:25] Ryan_Lane: around?> [10:07:31] there is a problem with memcached in puppet [10:08:29] !log deployment-prep removing the puppetized memcached [10:08:29] Logged the message, Master [11:56:46] surrururur [12:03:43] is labs supposed to support https? [12:03:48] and account creation is still broken [12:04:55] <^demon> s/broken/disabled/ [12:05:56] ^demon: why disabled? why doesn't it say so? [12:06:07] on labs *wiki* [12:06:16] <^demon> Because labs wiki uses ldap. [12:06:49] <^demon> It's not going to be like that forever. [12:06:54] sigh [12:06:59] I'm going to file a bug report about that [12:07:29] <^demon> it is going to use ldap forever...I'm not talking properly yet this morning. [12:07:31] <^demon> Too early. [12:07:39] <^demon> Need caffeine. [12:11:49] ping mutante ssmollett: I'd like to have an ccount for testing translation memory [12:14:45] <^demon> Nikerabbit: Ohhh, I thought you were talking about labsconsole. [12:14:56] <^demon> Not the beta site. [12:15:28] <^demon> w.r.t. account creation. [12:15:34] <^demon> Ignore everything I said about ldap. [12:18:59] ^demon: might want to add a note to https://bugzilla.wikimedia.org/34368 [12:20:40] Nikerabbit: if anything is broken [12:20:43] tell me [12:20:45] on beta [12:21:21] petan|wk: see those two bugs I filed [12:21:30] btw ^demon what if we implemented autocompletion to Special space :P [12:21:43] like Special:Recent would autocomplete to RecentChanges :D [12:21:47] or Special:RC [12:21:49] etc [12:21:51] aliases [12:22:03] <^demon> There was a bug about specials in autocompletion. [12:22:11] aha [12:22:16] I think it would be really nice [12:22:28] <^demon> I don't like shortcuts like Special:RC though. [12:22:39] autocompletion for specials should work already [12:23:35] Nikerabbit: registration work to me, sorry [12:23:39] what is broken? [12:23:48] I mean what you see when you try to register an account [12:24:56] petan|wk: "uknown source, please make sure you accept cookies" [12:25:05] I just registered an account in firefox [12:25:10] using en.wikipedia [12:25:39] http://labs.wikimedia.beta.wmflabs.org/w/index.php?title=Special:UserLogin [12:26:02] it's not so easy to track changes because many people change stuff, but no one track the changes [12:27:26] created test account on labs, with no problems [12:27:32] hmph [12:27:32] which browser you use? [12:27:37] chrome [12:27:38] I will try to replicate it [12:27:40] now it worked [12:27:51] it was probably caused by memcached [12:27:56] ah [12:27:57] it was misconfigured like 2 hours ago [12:28:10] puppet overwrites the local changes [12:28:11] it sucks [12:28:20] I need to do something about it [12:28:28] otherwise puppet version is of absolute no use [12:28:36] absolutely [12:37:17] btw Nikerabbit could you assign all beta bugs to me? benapetr@gmail.com [12:37:29] so I can easily see them [12:37:46] you said there are 2 bugs, but I see only 1 [12:37:52] in bz [12:37:58] petan|wk: ask someone to make you the default cc [12:38:09] * petan|wk seeks someone [12:38:22] there is no bugzilla category for beta site [12:38:29] ^demon perhaps iirc? [12:38:30] [14:07:35] wikibugs_> (NEW) Please support https for http://labs.wikimedia.beta.wmflabs.org/wiki/Main_Page - https://bugzilla.wikimedia.org/34367 normal; Wikimedia Labs: Setup; (niklas.laxstrom) [12:38:34] [14:09:31] wikibugs_> (NEW) Account creation on http://labs.wikimedia.beta.wmflabs.org/ too difficult - https://bugzilla.wikimedia.org/34368 normal; Wikimedia Labs: Setup; (niklas.laxstrom) [12:38:37] and I don't know if I want to be default cc for whole labs [12:39:07] you can do it per component [12:40:11] <^demon> Component name + description + assignee + any other cc list members? [12:41:05] I don't if we call it deployment or beta [12:41:14] both deployment.wmflabs and beta works for that [12:41:24] but current url is beta.wmflabs [12:41:42] <^demon> "Deployment beta" sounds descriptive enough [12:41:46] ok [12:41:55] can you insert it as component of labs [12:42:21] <^demon> Doing that now [12:42:52] <^demon> You want to be the default assignee? [12:43:00] yes, me or hexmode [12:43:16] <^demon> I'll add him to the cc list. [12:43:19] ok [12:44:15] thanks [12:44:51] <^demon> No problem. It's done so you should see it on new bug pages and advanced search now. [12:56:37] Reedy: hi [12:57:03] :o [12:57:17] ^demon: who is working on wmf branches? [12:58:18] <^demon> Reedy [12:58:27] <^demon> He's been doing the branching/merging. [12:58:32] hm [12:58:37] but it's broken [12:58:46] <^demon> Be more specific. [12:58:49] MobileFrontend is missing [12:58:52] ./extensions [12:58:57] it's installed on prod [12:59:15] when I switched most of wikis were not available [12:59:24] <^demon> Hrm, you're right. [13:37:53] Lol [13:38:00] Must've got removed from the list at somepoint [13:39:19] ^demon: we might aswell just branch MF2 [13:40:15] can you fix it?> [13:40:29] it's a blocker to push wmf1 to beta [13:40:48] yeah, i'm just waiting for svn up to run [13:42:35] k [13:42:36] ty [13:42:43] done [13:52:31] Reedy: it works but it looks pretty weird [13:52:46] in firefox I don't see the search button [13:52:55] just floating search text [13:52:56] no idea [13:53:04] somewhere outside of the screen [13:53:13] hm, ok [13:53:43] it should be the same as is running on cluster etc [13:53:59] it's definitely not [13:54:22] on cluster there is an image for that [13:54:42] * Nikerabbit waits [13:54:53] Nikerabbit: ? [13:54:56] does it work to you [13:55:13] it=what? [13:55:24] wondering what you were waiting for :) [13:55:39] it = beta [13:55:41] ah [13:56:11] it's a bit slow [13:56:36] http://en.wikipedia.beta.wmflabs.org/wiki/Main_Page?useformat=mobile [13:56:40] Looks fine for me in chrome [13:57:03] but I'm waiting for account and looking for docs how to use it [13:58:07] Reedy: I meant the regular view [13:58:13] for pc [13:58:22] What? [13:58:29] like no ?useformat [13:58:42] Fine also [13:58:55] in that case it's something here [13:59:02] in browser.. [13:59:04] ok [13:59:25] Nikerabbit: account on labsconsole? [13:59:40] or beta [14:01:22] petan|wk: good question, I don't know the difference, yet [14:04:03] labsconsole is a place to control vm [14:04:05] vm's [14:04:13] beta is a deployment test site [14:04:34] on labsconsole you can create a project and install own software to test [14:04:49] deployment beta is only for testing latest mediawiki before deployment [14:05:09] !labs [14:05:09] https://labsconsole.wikimedia.org/wiki/$1 [14:05:11] <^demon> Nikerabbit: Basically labsconsole is the VM controller. and beta and all the other ones are VMs on that. [14:05:12] meh [14:05:23] !info [14:05:23] https://www.mediawiki.org/wiki/WMF_Projects/Wikimedia_Labs [14:05:30] that's it [14:05:40] !labs Help:Contents [14:05:40] https://labsconsole.wikimedia.org/wiki/Help:Contents [14:05:53] this is a help I started working on [14:05:58] cool [14:06:13] I guess I need labs then [14:06:16] ok [14:06:24] mutante: around? [14:06:35] !account-questions [14:06:35] I need the following info from you: 1. Your preferred wiki user name. This will also be your git username, so if you'd prefer this to be your real name, then provide your real name. 2. Your SVN account name, or your preferred shell account name, if you do not have SVN access. 3. Your preferred email address. [14:06:48] Nikerabbit: that's answers you can prepare for now :D [14:06:51] that are [14:07:26] Ryan or mutante would need to know that in order to create an account [14:07:44] sure [14:07:49] but Ryan or Sara will likely be there after 6pm in eu [14:08:00] dunno where mutante lives [14:08:06] but he's around more often I guess [14:08:12] <^demon> He's got an LDAP account already. [14:08:24] I know but that's not enough I guess [14:08:34] someone still needs to link it to wiki acc [14:08:38] and create an account there [14:08:43] <^demon> There's a command line thing that I can run, but I can't remember it. [14:08:44] that someone needs to be wiki sysop [14:08:45] <^demon> Ryan told me. [14:08:48] <^demon> Shoulda wrote it down [14:08:51] yes there is [14:08:54] I think it's in logs [14:09:00] because he was telling it to many people [14:09:10] but no one yet created a php search engine for logs [14:09:14] so it's going to be hard to find it [14:10:02] oh wait [14:10:05] it's in Help [14:10:31] ok there is how to disable it [14:10:39] maybe if you did it upside down :) [14:10:59] hehe [14:13:55] PROBLEM Disk Space is now: WARNING on nagios 127.0.0.1 output: DISK WARNING - free space: /home/petrb 3403 MB (19% inode=93%): [14:15:16] petan|wk: hi [14:15:19] hi [14:15:25] Nikerabbit needs you :) [14:15:38] yay [14:16:26] ok, reading backlog. needs an account labs ? [14:16:39] yup [14:16:46] and maybe some help too [14:18:55] RECOVERY Disk Space is now: OK on nagios 127.0.0.1 output: DISK OK [14:20:49] Nikerabbit: alright, i dont see you in wiki users list, so what is your preferred wiki user name (which will also be git user name) [14:28:23] arr.. There was either an authentication database error or you are not allowed to update your external account. [14:33:58] PROBLEM Disk Space is now: WARNING on nagios 127.0.0.1 output: DISK WARNING - free space: /home/petrb 2852 MB (16% inode=94%): [14:40:24] !new-labsuser [14:40:24] add-labs-user --wikiname="" --mail="" [14:40:34] <--:) [14:41:00] that's cause there already was an LDAP user, unlike a completely new user..yes [14:41:34] <^demon> mutante: Many thanks [14:43:15] !bastion added nikerabbit as a labs user, using add-labs-user on virt0 [14:43:16] http://en.wikipedia.org/wiki/Bastion_host; lab's specific bastion host is: bastion.wmflabs.org; see !access [14:43:20] !log bastion added nikerabbit as a labs user, using add-labs-user on virt0 [14:43:22] Logged the message, Master [14:43:38] Nikerabbit: try logging in to labsconsole wiki now, using your existing credentials [14:44:21] @search add [14:44:22] Results (found 4): account-questions, new-labsuser, new-ldapuser, domain, [14:44:34] !domain [14:44:34] in case you want to assign a domain to your ip, you can use manage addresses to do that [14:44:44] !new-ldapuser [14:44:44] add-ldap-user [options] [14:44:54] hmm [14:45:08] but if the user is a new user, you can just create it via mediawiki [14:47:27] mutante: existing credentials to where? [14:48:33] Nikerabbit: your LDAP user [14:48:50] mutante: I'm not sure I've ever used it [14:49:18] Nikerabbit: you can just let it reset pw for you [14:49:20] Nikerabbit: your svn user [14:49:40] "If you currently have SVN access, then you have an account," [14:49:50] that's not entirely true [14:49:58] I had it [14:50:13] thought I needed to have it linked to labs [14:50:26] ok, i just linked it though [14:50:35] you can just reset pw in mediawiki that would change it in ldap [14:50:36] so yeah, try reset password via labsconsole mediawiki [14:51:49] petan|wk: yes, the full sentence is ". If you currently have SVN access, then you have an account, but need to have it linked to Labs.", it just did not say HOW to link it, the answer is !add-ldap-user [14:52:23] arg, no, it is: !add-labs-user of course [14:52:48] hm... [14:52:57] got in to the wiki [14:53:18] ok it's time to start reading help ;) [14:53:21] :) [14:53:27] indeed [14:54:42] Nikerabbit: next step is uploading your SSH key via the wiki [14:55:13] we should see how an accounts is created for you on the bastion host [14:55:14] <^demon> mutante: Any clue about gerrit permissions, or is Ryan my only source for that? [14:55:51] ^demon: afraid he is, you mean like "the right to merge" etc? [14:56:09] <^demon> Yeahhhh :\ [14:56:22] <^demon> I'm editing the permissions for my new repo but I'm not 100% sure how I need it. [14:56:23] Nikerabbit: once you got a key up there, try ssh to bastion.wmflabs.org [14:56:52] ^demon: sorry, dont know [14:58:53] mutante: looks like it alrady has my public key (although the interface only shows the beginning of it) [15:00:11] 02/13/2012 - 15:00:11 - Creating a home directory for nikerabbit at /export/home/bastion/nikerabbit [15:00:19] Nikerabbit: you are already in group "sysadmin", but i had to add you to the bastion host project [15:00:31] Nikerabbit: see above, it just created your home after i did that [15:01:09] 02/13/2012 - 15:01:09 - Updating keys for nikerabbit [15:01:56] !log bastion added member nikerabbit [15:01:57] Logged the message, Master [15:02:11] 02/13/2012 - 15:02:10 - Updating keys for nikerabbit [15:03:03] *** System restart required *** [15:03:04] yay [15:03:28] welcome to labs [15:03:50] okay, then, how do I break it? [15:03:53] did you want to join an existing project ? [15:03:57] <^demon> Nikerabbit: rm -fR / [15:04:00] <^demon> :) [15:04:42] mutante: I'd probably start my own [15:05:45] Nikerabbit: as a sysadmin, you should be able to create new instance.. so i'd try that next [15:06:39] Nikerabbit: either just an instance to test, within project "testlabs", or we can create your own separate project right away [15:07:51] I'm not even sure what makes sense :) [15:07:51] suggest a project name? you said "to test translation-memory"?? [15:07:56] yes [15:08:12] the tm is branded as ttmserver [15:08:58] just "translation-memory" ok with you? [15:09:12] whatever it's just a name [15:09:36] that should be clear to everyone what it is about [15:10:08] !log bastion created project "translation-memory" for Nikerabbit [15:10:09] Logged the message, Master [15:10:19] 02/13/2012 - 15:10:18 - Creating a project directory for translation-memory [15:10:19] 02/13/2012 - 15:10:19 - Creating a home directory for nikerabbit at /export/home/translation-memory/nikerabbit [15:11:18] Nikerabbit: you are now a sysadmin and netadmin of that new project, see it in 'Manage Projects' [15:11:20] 02/13/2012 - 15:11:20 - Updating keys for nikerabbit [15:11:46] I see [15:11:51] Nikerabbit: now go to 'Manage Instances' [15:12:09] Nikerabbit: then click 'Add Instance' next to your project name [15:13:00] i suggest you select instance type "m1.small" for now [15:13:35] and leave security groups , DNS domain etc, at defaults [15:14:21] any naming patterns for instances? [15:15:08] not really, i see some ending in -01, some in 1, some just have names without numbers [15:15:37] maybe transmem1 [15:16:26] they are referred to by instance-id internally anyways [15:16:36] aaand it failed [15:16:43] oops? how [15:18:25] mutante: just says failed [15:18:32] "Failed to create instance. [15:19:16] hrmm,then let's try this: go to "Manage projects" and try to add me (Dzahn) as an admin of your project [15:22:41] Nikerabbit: yes there are naming conventions :) [15:22:59] did you try to add the instance to project "translation-memory" and not maybe to "testlabs"? [15:23:26] btw I can't create instances as well [15:23:28] mutante: translation-memory, I didn't see testlabs on that list [15:23:32] I thought I reached a limit [15:23:51] i'll add myself to the project and try [15:23:52] it's maybe a labs problem then [15:24:24] btw naming convention is prefix all hostnames with project [15:24:28] like huggle-blah [15:24:38] oh,ok [15:25:09] unfortunatelly not all people know that [15:25:21] 02/13/2012 - 15:25:20 - Creating a home directory for dzahn at /export/home/translation-memory/dzahn [15:25:29] ah, ther is no 'Add instance link' next to the project name, at least if i am only netadmin [15:25:37] yes [15:25:38] maybe it is a quota [15:25:41] you need to be sysadmin [15:25:55] netadmin is for firewall only [15:26:19] 02/13/2012 - 15:26:19 - Updating keys for dzahn [15:27:23] Created instance i-00000135 with image ami-0000001d and hostname i-00000135.pmtpa.wmflabs. [15:27:35] I can't create instances too [15:27:39] in my projects [15:28:00] !log translation-memory created an m1.small instance translation-memory-1 [15:28:02] Logged the message, Master [15:28:41] :o [15:28:54] why does it ask for the name if there is preferred convention for that? [15:29:15] right you don't have to follow that convention, but it's a good idea [15:29:19] Nikerabbit: do you see that instance now? can you use the "Configure" link next to it? [15:29:27] because otherwise it can conflict with an existing vm [15:29:46] petan|wk: hmm, then there must have been a change in permissions.. [15:30:01] !nagios [15:30:01] http://nagios.wmflabs.org/nagios3 [15:30:21] it's not easy to browse it when you don't follow conventions too [15:30:33] if you do that, all instance are grouped together in nagios [15:31:05] Is nagios still a manual thing or did we decide if we where using puppet or some custom system for it? [15:31:15] of course it's not manual :D [15:31:24] it uses nparser I wrote like months ago [15:31:33] I think I sent a message on wikitech about it [15:31:43] petan|wk: it would be nice if there were Nagios hostgroups and servicegroups per project ... http://nagios.wmflabs.org/cgi-bin/nagios3/status.cgi?hostgroup=all&style=overview [15:31:50] Probably should read wikitech more :P [15:31:56] oh yes, that would be cool [15:32:04] also it would be nice if people could login using ldap [15:32:09] mutante: yes and yes [15:32:10] instead of custom login as now [15:32:38] It would be nice if people could add custom stuff and ack/schedule downtime etc for projects they belong too [15:32:46] Nikerabbit: nice! then try to start it up, without any puppet classes besides default, and login from bastion host [15:32:57] Damianz: you can already do that :P [15:33:01] petan|wk: yes for Nagios LDAP! [15:33:11] see nagios.wmflabs.org/nlogin [15:33:13] petan|wk: I can if I feel you cookies :P [15:33:25] I also sent a message regarding that on labs-l [15:33:26] :P [15:33:37] Yeah not that I have a login to that :P [15:33:53] PROBLEM Current Load is now: CRITICAL on translation-memory-1 translation-memory-1 output: Connection refused by host [15:33:54] I asked many times on many places if someone want a login there :) [15:34:08] so now I pm you one... sec [15:34:29] mutante: I only have reboot, does that mean it is running? [15:34:32] Reminds me I should setup edit monitoring for cbng though as currently I know by people moaning [15:34:33] PROBLEM Current Users is now: CRITICAL on translation-memory-1 translation-memory-1 output: Connection refused by host [15:34:53] PROBLEM Disk Space is now: WARNING on labs-nfs1 labs-nfs1 output: DISK WARNING - free space: /export 917 MB (5% inode=94%): /home/SAVE 917 MB (5% inode=94%): [15:34:54] Nikerabbit: yes, it has state "Running" now [15:34:55] Could do with moving the puppet class to a misc class like Ryan said to do before someone else merged my change [15:35:05] ah I see [15:35:13] PROBLEM Disk Space is now: CRITICAL on translation-memory-1 translation-memory-1 output: Connection refused by host [15:35:47] something we should do to silence the bot? [15:35:53] PROBLEM Free ram is now: CRITICAL on translation-memory-1 translation-memory-1 output: Connection refused by host [15:36:02] Ack the alert if you're going to fix it [15:36:37] Btw is bots-3 better now? It had a few days of getting ram rapped by a bot :( [15:37:03] PROBLEM Total Processes is now: CRITICAL on translation-memory-1 translation-memory-1 output: Connection refused by host [15:37:10] Nikerabbit: because of the way Nagios config is created, it is quicker than the actual instance being up and having the nagios stuff installed.. on creation of new instances [15:37:17] Damianz: you can just use nova to change what services are being monitored [15:37:37] nparser is getting the nova classes and based on that add nagios checks [15:37:43] PROBLEM dpkg-check is now: CRITICAL on translation-memory-1 translation-memory-1 output: Connection refused by host [15:37:49] so if u have a puppetized service you can create a nagios check for that [15:37:50] Nikerabbit: next step after succesful login: sudo puppetd -tv [15:37:55] Ahhh [15:38:01] mutante: didn't get that far yet [15:38:11] I've not quite puppetized cbng as creating debs made me go insane [15:38:18] Permission denied (publickey). [15:38:32] Still on my todo list so we can destory and re-roll the box and loose nothing as mysql is the persistant data :D [15:38:40] Nikerabbit: you need to forward your SSH key, like add it to an agent, and "ssh -A" to the bastion host, then ssh to your instance name, same user [15:38:41] @search access [15:38:41] Results (found 4): access, bastion, account-questions, account, [15:38:46] !access [15:38:46] https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances [15:38:53] mutante: Doesn't he have to wait for the first nagios run to setup the users and stuff? [15:38:57] errr s/nagios/puppet/ [15:39:16] Damianz: last part of puppet is enable ssh [15:39:23] oh, wait it's not [15:39:24] Ah [15:39:35] actually you can see if it's done in console [15:39:41] mutante: yes my key is forwarded [15:39:43] !console [15:39:44] yay [15:39:54] That's true :) [15:39:56] Damianz: hmm, afair i could login right after the homedir created log line appeared [15:40:09] !console is in case you want to see what is happening on terminal of your vm, check console output [15:40:09] Key was added! [15:40:39] Damianz: but then i ran it manually and saw Nagios stuff being added [15:40:46] yes [15:40:48] that's a bug [15:40:51] or something [15:40:55] o.0 [15:40:59] mutante: soo any idea why it doesn't work? [15:41:03] Nikerabbit: yes [15:41:10] @search forward [15:41:10] No results found! :| [15:41:16] @search tunnel [15:41:17] Results (found 1): putty, [15:41:18] Nikerabbit: are you getting Permission Denied? [15:41:20] meh [15:41:26] problem is that your private key is not on bastion [15:41:29] mutante: yes [15:41:36] so that you can't login to any other vm [15:41:53] you need to create a new private key on bastion and upload public to labs [15:41:59] petan|wk: it doesnt have to, he is using Agent Forwarding [15:42:04] or that [15:42:12] but most of people find it very complicated [15:42:17] especially on windows [15:42:25] It's also more secure :P [15:42:28] on windows i suggest "pageant.exe" ,comes with putty [15:42:30] depends [15:42:42] its just a tray icon, with your loaded key [15:42:56] it's maybe more secure but I would prefer to create separate key, rather than forward some really private one [15:43:24] I am definitely not afraid of my private key on labs being compromised by another labs user [15:43:27] Thats why .ssh/config is awesome [15:43:34] soo, any suggestions? [15:43:42] hmm, bots project's homes are quite full, anyone going to resolve this? [15:43:46] Different keys and fowarding rules per host/domain/ip etc [15:43:53] Hydriz: The bots just like their data. [15:43:56] Nikerabbit: I recomment you to run ssh-keygen on bastion [15:44:00] recommend [15:44:10] Nikerabbit: so if you do "ssh-add -l" at home you see the same key you added to labsconsole wiki, you ssh -A to bastion.labs, and then just "ssh translation-memory-1" [15:44:11] 631M of space left, lol [15:44:31] Hydriz: it's not about bots [15:44:37] Hydriz: it's shared for all projects on labs [15:44:44] mutante: I do have the key forwarded, it's just not working [15:44:47] in fact I can't detect where is the problem [15:44:56] I thought each project had their over nfs boxes? [15:44:59] no [15:45:02] Nikerabbit: checking on the instance.. [15:45:04] unless you create one [15:45:04] yeah, but doing a df tells me about this for bots lol [15:45:06] Oh [15:45:14] We should just move to gluster :P [15:45:19] yes [15:45:23] !labsconf [15:45:24] http://etherpad.wikimedia.org/LabsIrcConf [15:45:26] Damianz: see it [15:45:36] I thought that was this friday? [15:45:38] So nope [15:45:39] no [15:45:43] it's going to be soon [15:46:32] Topic says this friday the 18th except saturday is the 18th [15:47:10] Sounds shiny [15:47:48] I think there will end up being issues about having access to data though once we have replication of wikipedia live. While needed for bots etc there is allways the issues of data :( [15:48:07] yes [15:48:09] I think so [15:48:26] btw if there is anything wrong in topic you should be able to fix it [15:48:36] I am in terminal now :D [15:48:46] so my client isn't really good :P [15:48:47] surrrrrur [15:48:48] heh [15:48:52] Hmm it doesn't say friday actually, maybe that was just my mind making it up earlier. [15:48:59] :D [15:49:00] * Damianz thinks petan|wk needs some irssi [15:49:03] I have [15:49:13] but I don't really know how to use it :D [15:49:19] LOL [15:49:29] I just know all the server side commands so /raw command ftw [15:49:39] I could irc in telnet actually heh [15:49:45] that would not be fun [15:49:48] I prefer irssi-proxy for most stuff but it's easy enough to use in cli... hardest bit is remembering how to reconnect to servers. [15:49:48] irssi rules [15:49:53] PROBLEM Disk Space is now: CRITICAL on labs-nfs1 labs-nfs1 output: DISK CRITICAL - free space: /export 414 MB (2% inode=94%): /home/SAVE 414 MB (2% inode=94%): [15:49:57] PING PING PING god dammit PING! [15:50:06] ping who [15:50:19] because of space? [15:50:23] I removed all porn from my home [15:50:26] and it's still full [15:50:34] :P [15:50:38] what are we going to do [15:50:41] yay [15:50:43] :O [15:50:55] Damianz: do rm -rf ~ pls [15:52:21] :P [15:52:29] su petan -; rm -f ~ [15:52:35] petan is not my ldap [15:52:37] hehe [15:52:44] aww [15:52:46] rm -f ~ would fail [15:52:51] because of no -r [15:52:57] True [15:53:04] I could remove some stuff but only a few mb like the backup of cbng's php [15:53:10] also su petan-;blah would execute blah as previous user [15:53:22] <^demon> I've starred 16 gerrit bugs :p [15:53:25] FINE -c 'rm -f ~' [15:53:25] it would be executed after end of su session, double fail [15:53:42] yes but that' would again fail because of -r :D [15:53:42] I think the challenge for this week should be to write an irc bot using netcat and expect :P [15:53:55] Damianz: check source code of wm-bot [15:53:58] petan: Oh I aliases your rm to be rm -f :P [15:53:58] it uses raw [15:54:05] :D [15:54:07] s/f/r/ [15:54:17] That's the most evil thing ever, alias like echo to me rm -f / [15:54:21] wm-bot is uses raw socket to irc [15:54:25] Always run commands like \rm [15:54:26] using [15:54:36] :D [15:54:44] hmm [15:54:49] I think I will continue tomorrow [15:55:06] I'm happy cbng is using node now :D [15:55:08] Nikerabbit: seriusly make a key :P [15:55:22] node-irc is pretty cool and more reliable than the php thing it was doing before. [15:56:04] Nikerabbit: forwarding works fine for me (on other isntances), but not yours, there is just some problem with the ssh keys in the homedir on the instance, maybe after next puppet run [15:56:09] protocol of irc is simple enough [15:56:24] oh yes [15:56:25] It is, also makes it kinda old and not so flexible [15:56:34] Would be cool if IRC supported a true mesh betwean servers. [15:56:37] it is old :D [15:56:51] petan|wk: not sure how well that plays along with proxying [15:56:53] it would be cool if it supported voip [15:56:54] :D [15:56:56] <^demon> Don't fix what ain't broken :p [15:57:01] Nikerabbit: i'll try to figure it out [15:57:08] Nikerabbit: it plays pretty fine [15:57:28] Nah we have XMPP for that [15:57:36] IRC<>XMPP would be cool though [15:57:38] well, anyway, tomorrow [15:57:41] kk [15:57:46] k [15:57:59] 11h day should be enough, right? [15:58:03] yes! [15:58:43] meh I used to spend more time doing wiki stuff [15:58:47] :D [15:59:03] checking home dir sizes .. [15:59:15] if anything is broken it wasn't me! [15:59:30] all the crap in my home was created by Ryan [15:59:39] he just chowned it then [16:00:03] the homedir of bots is the largest of them all [16:00:08] yay [16:00:09] is that the one? [16:00:30] please don't tell me petrb is largest one [16:00:58] meh [16:01:00] beetstra is [16:01:06] almost 10G [16:01:09] YAY [16:01:28] almost as big as my swap :X [16:01:50] /home/bots/beetstra/linkwatcher ! [16:02:11] 9.6G nohup.out [16:02:12] haha [16:02:37] sudo mv /home/bots/beetstra/linkwatcher /mnt/share/linkwatcher;sudo ln -s /mnt/share/linkwatcher /home/bots/beetstra/linkwatcher [16:02:44] uhm..how to contact beetstra? [16:02:49] heh [16:03:01] mutante: I can fix it :P [16:03:08] but it wouldn't be nice [16:03:19] I guess if I moved it, his bot would likely crash [16:03:31] let me check if it's not just a big logfile [16:03:54] it's html [16:04:03] but doesnt look like its still growing [16:05:56] no it's not [16:06:11] !log bots moved nohup from beetstra's home to /mnt/share on bots-1 [16:06:12] Logged the message, Master [16:06:18] it's nohup log [16:06:28] it's gonna take a while [16:06:42] heh [16:09:11] btw a best version to contact him is probably email [16:09:14] * way [16:09:21] not a version [16:09:22] :D [16:09:27] heh [16:09:49] I don't really like touching other's data but I guess this is emergency [16:10:29] why would moving it break it? [16:10:45] I didn't say it would [16:10:54] the issue would actually be that even deleting it won't release the disk space [16:11:12] are you sure? [16:11:12] it's nohup [16:11:20] that always close the file descriptor on delete [16:11:33] or it should have [16:11:40] depends on version of bash [16:13:16] oh I meant if I moved whole folder [16:13:17] it would crash [16:13:24] I wasn't talking about the nohup [16:13:51] even if it actually watches the file deletion, I don't think it would close it at the child [16:14:06] why would moving the folder crash it? [16:14:17] maybe because there are some files it needs for it to run [16:14:26] I don't know [16:14:37] that's a reason I don't want to touch it [16:14:44] if I was a programmer of the bot I would be sure [16:14:50] well, that's obviously safer :) [16:15:18] let's check fs [16:15:28] meh [16:15:45] btw fuser can show u open fd [16:16:02] what is that content in nohup.out.. anyways [16:16:10] I don't know [16:16:12] wiki html.. but does the bot parse it_ [16:16:30] actually I respect his content and his privacy so I don't check content of his logs [16:16:42] first thought was to gzip it [16:16:50] maybe it contains password of his bank acc :o [16:16:57] heh, no secret data in labs [16:17:12] <^demon> If you're putting your bank account data into labs you're asking to get screwed. [16:17:20] I do [16:17:21] :D [16:17:27] wheee [16:17:46] * petan|wk puts all money in under pillow [16:17:52] <^demon> You fell for the trap. Labs is secretly a phishing scheme ;-) [16:18:02] YAY [16:18:06] "fundraiser"?;) [16:18:09] * Platonides discovers that petan pillow is in his bank [16:18:12] you knew all users will end up with putting secret stuff in nohup's! [16:18:59] can you let the bot output "i am running out of disk" ?:) [16:18:59] bah [16:19:12] maybe it's there [16:19:27] I've seen a guy who starte nohup yes & [16:19:30] :) [16:20:10] ouch [16:22:17] zip it, gpg it with his public key, move to shared ... actually..if you dont want to touch his data, but move it to _shared_ ..hmm hmm [16:22:44] ok, I will change its perms after that [16:22:49] in fact [16:22:54] I don't have to [16:23:02] it still have the previous one [16:23:05] but..there shouldnt be any secret data anyways [16:23:37] right, but still I don't want to analyze 10gb text file :) [16:23:47] fair [16:24:01] but if u want... what's your email :P [16:24:21] heh, it does not accept 10gb attachments;) [16:24:25] thanks [16:24:26] yay [16:31:56] Alan Alexander III DOB - July 3 1985 SSN - 151-82-7685 Address - 107 Emery St Portland ME [16:32:14] o.o [16:33:46] who is that [16:43:02] !ping is pong [16:43:03] Key was added! [16:45:57] btw, Beetstra is on #wikichem , left a message there [16:49:53] RECOVERY Disk Space is now: OK on labs-nfs1 labs-nfs1 output: DISK OK [16:50:03] :) [16:52:04] mutante: who can create mailing lists [16:52:25] we just had a small discussion in #huggle, some folks would like to have a mailing list for it, huggle-l [16:52:32] like awb [16:52:58] petan|wk: i can do that, but please via RT [16:53:08] eh, I can't access RT [16:56:26] mutante: bz? [16:57:12] I created a ticket for it right now [16:58:34] petan|wk: that's ok, i will link to RT then, ticket creation should be possible via email, just btw, for next time [16:59:01] ok [16:59:15] b34374 [16:59:17] petan|wk: we don't want lists to end in "-l" anymore though [16:59:23] k [16:59:26] then just huggle [16:59:44] I thought it's a standard [16:59:48] idc [16:59:51] huggle is fine [16:59:57] is it possible? [17:00:11] or is there a requirement to have certain number of devs, etc [17:00:39] there is a page about the naming conventions somewhere [17:00:47] hm... [17:00:48] which says we want to drop the -l [17:00:52] right [17:00:54] huggle [17:01:12] i dont know about requirements like min. number of users [17:01:21] ok [17:02:02] let me know once you create it, thanks [17:02:14] also can you make me a list mod? [17:03:42] yes, i would make you the list admin [17:04:04] ok [17:04:05] brb [17:04:07] but note that in mailman, mod and admin are not really user flags, they are just passwords [17:04:17] that's ok [17:04:18] you either know them, or you dont, but unrelated to a specific list user [17:04:35] I just wanted to have control over spam / problematic threads [17:05:06] but I am heading from the work, be back in hour [17:31:05] * ^demon kicks gerrit [17:31:13] <^demon> `git push origin :branchname` won't work :\ [17:31:43] * Damianz watches gerrit kick ^demon [17:33:29] <^demon> $ git push origin :refs/tags/1.1.0 [17:33:29] <^demon> To ssh://demon@gerrit.wikimedia.org:29418/test/mediawiki/core2.git [17:33:29] <^demon> ! [remote rejected] 1.1.0 (prohibited by Gerrit) [17:33:30] <^demon> error: failed to push some refs to 'ssh://demon@gerrit.wikimedia.org:29418/test/mediawiki/core2.git' [17:33:33] <^demon> Stupid gerrit. [17:36:41] afair there was a reason for not allowing that.. and configured like that on purpose [17:37:56] <^demon> Yeah, I was trying to override Push Annotated Tag on test/mediawiki/core2.git so I could delete a tag. [17:38:01] <^demon> But doesn't seem to work. [17:39:20] <^demon> Ha, you need Push +force [17:39:27] <^demon> on refs/tags/* [18:30:40] hi Ryan_Lane [18:30:47] there is a problem with creation of instances [18:30:53] it seems that only wikiadmins can create them [18:33:18] * Beetstra looks at petan [18:33:58] hi Beetstra [18:34:03] I moved your nohup [18:34:06] it had 10 gb [18:34:11] that was what I was going to ask about [18:34:13] Already .. [18:34:16] ah [18:34:21] can I removed it? [18:34:23] or you need it [18:34:26] please, no [18:34:29] ok [18:34:36] I want to see what is put into it .. [18:34:39] where is it? [18:34:40] No symbol 'it' found anywhere. [18:34:43] it's in /mnt/share on bots-1 [18:35:02] petan: oh? what problem? [18:35:07] hm. really? [18:35:10] it fails [18:35:11] what issue with instances? [18:35:15] everytime for everyone [18:35:15] in which way? [18:35:18] hm [18:35:21] Creation of xx failed [18:35:30] oh [18:35:31] interesting [18:35:37] mutante was able to create it [18:35:59] I mean I only tried it in 2 projects [18:36:15] but others have the same [18:36:15] error [18:36:20] :o [18:36:27] let me look through the logs [18:36:31] ok [18:37:56] Ah .. found it [18:38:01] ok [18:38:07] I think it's still running the mv [18:38:08] :P [18:38:13] so it's probably not complete [18:38:28] full version is probably still in your home [18:41:23] You can delete the nohup.out, petan [18:41:42] ah, which one [18:41:49] the one in /mnt can stay there [18:41:56] it's no prob [18:42:09] but please try to clean your home [18:42:26] also if you move your bot from home to /mnt/share or /share it's ok [18:42:35] petan: can you try to create one for me? [18:42:42] Ryan_Lane: sure [18:43:49] Failed to create instance. [18:43:52] InstanceLimitExceeded: Instance quota exceeded. You cannot run any more instances of this type. [18:43:53] project bots [18:43:55] heh [18:43:56] hm [18:44:00] ah [18:44:05] which other projects have this issue? [18:44:10] deployment-prep [18:44:12] <^demon> Ryan_Lane: So I started building out the mediawiki/ family of repos, but I could use a hand on making sure I've got the permissions correct. [18:44:23] ^demon: ok. give me a bit [18:44:29] <^demon> Okie dokie. [18:44:33] ok if it's only in our project it's ok [18:44:39] I thought all people are affected heh [18:44:51] because NIklas had a same problem in empty project [18:44:57] 10 instance limit by default, it seems [18:45:10] 20 cores [18:45:11] it would be cool to make error like "quote exceeded" [18:45:14] yep [18:45:22] I'll have to see if the api returns that [18:45:27] ok [18:46:17] ok. try now [18:46:48] I don't really need to make one, but ok [18:47:02] works [18:47:30] I guess we will need it when more people join bot project [18:47:42] I upped it to 20 [18:47:43] because our 4 application servers are overloaded and we share them [18:47:48] same with deployment-prep [18:47:54] I can't imagine every bot op will have own instance [18:47:59] that would be like 600 instances [18:48:12] because that's a number of bot ops on toolserver [18:48:18] or something like that [18:48:33] * Ryan_Lane nods [18:48:47] they can be shared. I'd prefer that some scheduler run them, though [18:48:52] yes [18:49:04] for instance my bots are not really started periodically but run [18:49:06] non stop [18:49:10] like irc bots, etc [18:49:28] process just put itself to sleep after task is done [18:49:35] right [18:49:38] and start again when it's needed [18:50:12] scheduler would be cool though [18:52:08] hyperon: can u update puppet on bots-5 I am a bit away now, thanks [18:52:23] just puppetd -tv [18:52:29] Ryan_Lane: can u make it automatic? [18:52:36] so that you create instance and don't need to do that [18:52:48] because otherwise nrpe doesn't work [18:52:51] hi ryan [18:53:03] petan: I'm not sure why that isn't working [18:53:10] hm... [18:53:13] it happens 30 minutes later on the next puppet run [18:53:15] Nikerabbit: howdy [18:53:26] Ryan_Lane: I just got started with labs today [18:53:58] had some problems, for example I can not login to newly created instance [18:53:59] great. [18:54:02] hm [18:54:05] which instance? [18:54:10] translate-memory-1 [18:54:24] what error do you get? [18:54:38] permission denied (public key) [18:54:44] and yes I did check that my key is forwarded [18:55:23] can you ssh into bastion, from bastion? [18:55:58] Ryan_Lane: yes I'm trying from bastion [18:56:25] so, if you are on bastion, and do: ssh bastion [18:56:30] it lets you in? [18:56:58] bastion1, but yes [18:57:14] ok [18:57:18] lemme look at the instance [18:58:18] 02/13/2012 - 18:58:18 - Creating a home directory for laner at /export/home/translation-memory/laner [18:59:19] it failed to build [18:59:20] 02/13/2012 - 18:59:19 - Updating keys for laner [18:59:26] Nikerabbit: delete it and recreate it [18:59:39] you can see that info via the console log [19:00:02] oh [19:00:20] "Failed to delete instance. [19:00:42] o.O [19:00:56] wow. odd [19:01:01] everything liks to fail for me :) [19:01:18] yeah even the creation failed for me [19:01:32] <^demon> He doesn't like you. [19:01:34] so confusing [19:01:38] <^demon> I don't like you either! [19:02:03] looking through logs [19:02:28] not seeing errors... [19:02:41] Nikerabbit: can you try to delete it again? [19:02:46] sure [19:02:56] crunching [19:03:46] PROBLEM Current Load is now: CRITICAL on bots-5 bots-5 output: Connection refused by host [19:04:05] ? [19:04:26] PROBLEM Current Users is now: CRITICAL on bots-5 bots-5 output: Connection refused by host [19:04:31] I leave for a week and a half and everything breaks. heh [19:05:04] same error [19:05:06] PROBLEM Disk Space is now: CRITICAL on bots-5 bots-5 output: Connection refused by host [19:05:46] PROBLEM Free ram is now: CRITICAL on bots-5 bots-5 output: Connection refused by host [19:06:36] was tailing the log. trying to find an error [19:06:56] PROBLEM Total Processes is now: CRITICAL on bots-5 bots-5 output: Connection refused by host [19:07:26] there's no error... [19:07:36] PROBLEM dpkg-check is now: CRITICAL on bots-5 bots-5 output: Connection refused by host [19:07:45] let me try [19:08:00] worked for me... [19:08:32] ah. I wonder if it has something to do with your username [19:08:38] wiki name [19:09:05] I couldn't imagine so, though [19:09:33] it can't find you in the sysadmin role, it says [19:09:43] * Ryan_Lane goes to check [19:09:54] hm. you're there [19:10:03] Nikerabbit: can you try creating a new instance? [19:11:50] Ryan_Lane: It took a week and a half... must try harder :P [19:12:34] heh [19:13:09] 'Thanks to stupid job titles in tech, recruiting an *actual* ninja is proving very difficult.' lol [19:15:03] hahaha [19:15:25] haha [19:15:48] Ryan_Lane: uga [19:16:27] Ryan_Lane: failed again [19:17:32] ok [19:18:03] it's odd. I don't see any failures [19:18:46] Maybe it's magic phpness [19:20:09] I dunno [19:22:18] Nikerabbit: can you log out, then back in? [19:22:21] then try? [19:22:31] it's the only thing I can possibly think of right now [19:22:36] there's a known issue with sessions [19:26:01] Ryan_Lane: nope :< [19:26:07] gah [19:26:14] this is perplexing. [19:26:23] it there were errors, it would be a lot easier [19:26:31] let me put debugging statements in [19:26:40] Errors would take all the fun out of it :P [19:27:01] * Damianz notes to tighten his screw before it falls out [19:30:57] Nikerabbit: you were able to create one the other day? [19:32:08] Ryan_Lane: no [19:32:17] I think mutante did that for me today [19:33:29] ah [19:35:59] oh. hah [19:36:00] I see [19:36:05] or not :( [19:37:14] your account does have a weird attribute on it [19:37:59] it's likely encoded, due to the special character in your name [19:38:09] that may be what's breaking things [19:38:15] maybe nova doesn't handle it properly [19:38:59] Ryan_Lane: I'm special! [19:39:03] heh [19:41:25] I agree [19:42:17] Ryan_Lane: so am I going to be your test bunny or are you going to make me use ascii? [19:42:43] test bunny [19:42:50] I'm waiting for him to just str_replace your name :P [19:42:51] UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 13: ordinal not in range(128) [19:42:57] aaaaaannnndddd fail [19:43:28] yup that's ä [19:43:36] ö I mean [19:45:42] hmm. I may not be able to fix this [19:45:58] I will be adding a bug report for it, though [19:45:59] Would that be lovely python string handling by any chance? D: [19:46:11] yeah, however they are doing it in nova [19:46:14] Python and unicode never play well [19:48:33] asking in openstack channels [19:48:38] and searching for bugs [19:48:40] that's pretty annoying [19:49:08] Zarraboogs! [19:50:47] fun [19:51:07] Nikerabbit: we may have to change your name :( [19:51:22] * Damianz votes to change it to 'BROKEN' [19:51:24] can you use nikerabbit then [19:52:49] I only need to figure out how to do shared databases [19:54:45] Ryan_Lane: I can test and see if the behavior is different in essex, if that's useful. Either way he probably has to change his username though. [19:54:54] Essex looks to be generally utf8ified [19:56:40] yeah. please check in essex for me. [19:56:56] Nikerabbit: had you already logged into gerrit? [19:57:23] Ryan_Lane: no [19:58:06] ah. good [19:58:09] that'll make this easier [20:00:45] ok. I think you should be good to go with Nikerabbit [20:00:51] can you log out and log in as Nikerabbit? [20:01:11] 02/13/2012 - 20:01:11 - Updating keys for nikerabbit [20:01:20] 02/13/2012 - 20:01:19 - Updating keys for nikerabbit [20:02:25] done [20:03:14] working now? [20:03:37] crunching [20:03:40] yup [20:03:50] great [20:04:02] maybe we'll be able to fix your name in the future [20:05:40] hmph, still cannot log in to it [20:06:04] does it just take some time? [20:06:26] can't log in? [20:06:33] sec [20:07:01] any error? [20:07:05] Ryan_Lane: Looks to be broken in essex too. [20:07:13] * Ryan_Lane shakes fist [20:07:16] the same, permission denied (public key) [20:07:20] can you enter a bug report for it? [20:07:23] Nikerabbit: oh. that [20:07:31] lemme look at your instance [20:07:35] Ryan_Lane: Yep. And will try to fix as well. [20:07:45] awesome. thanks [20:08:35] seems brewster is dead [20:10:32] GlobalBlocking hasn't been setup on labs [20:11:05] * Damianz blocks Reedy [20:11:50] Nikerabbit: you'll need to delete/re-create the instance again [20:12:07] stupid brewster occasionally dies, then instance creation won't work, because they can't install puppet [20:13:05] o.0 [20:13:19] [20:13:54] brewster is our apt mirror [20:14:20] kill it with fire! [20:14:29] now I see that it is not properly translated into Finnish :O [20:16:00] E: Couldn't find package puppet [20:16:27] Isn't puppet in the standard apt repos anyway? [20:16:31] yep [20:16:52] Or is there a mirror locally as oppose to custom packages? [20:17:34] PROBLEM host: translation-memory-1 is DOWN address: translation-memory-1 CRITICAL - Host Unreachable (translation-memory-1) [20:18:25] a mirror [20:18:28] and a custom repo [20:18:45] Ah [20:19:04] christ. again? what's wrong with brewster [20:20:01] it doesn't like bunnies [20:20:13] seems not [20:20:16] Maybe it decided it would rather be a yum mirror. [20:20:36] well, it's back up now, so I recreated your instance again [20:20:46] let's see if it works this time :) [20:21:28] arrrggghhh [20:21:41] Hmm apparently fosdem videos are starting to appear but I've still not found any working links yet :( Also more annoyingly is they are redirecting to a readme rather than throwing a 404 so mirroring it currently is futile D: [20:26:46] * Ryan_Lane sighs [20:26:50] brewster is still broken [20:34:06] <^demon> Ryan_Lane: Think you could poke gerrit now? [20:34:16] oh. sure [20:34:57] <^demon> Ok. What I want is mediawiki/core.git to have a gated master like we do with ops. [20:35:15] <^demon> And mediawiki/extensions to have the same by default so subprojects don't have to be customized as much. [20:35:27] <^demon> And the i18n group should be able to push-without-review for both. [20:35:39] <^demon> "mediawiki" is the review group. [20:37:15] I don't see any permissions at all, really [20:37:31] ^demon: I assume that twn should use own branch of Translate [20:37:47] <^demon> Nikerabbit: That's totally do-able, once we get it pushed. [20:38:08] you need to add merge somewhere [20:38:10] <^demon> Ryan_Lane: I started at the mediawiki/ level giving review/verify rights. [20:38:19] err [20:38:20] submit [20:39:00] <^demon> Ok added submit [20:39:19] look at how puppet is done: https://gerrit.wikimedia.org/r/#admin,project,operations/puppet,access [20:39:40] <^demon> I looked but I got confused since we're just doing master and not the other branches [20:39:45] * Ryan_Lane nods [20:40:15] the create reference permission lets people create branches on the servr [20:40:34] push merge commit lets people do merges [20:40:38] not sure if you want that [20:40:51] <^demon> No, just the 'mediawiki' group to do that. [20:41:12] <^demon> I want normal users to just be able to push refs/for/master I suppose. [20:46:07] if you stick push in the wrong (or right, depending on what you are doing) spot, it'll allow them to bypass review [20:46:16] I wish their documentation didn't totally suck [20:51:16] <^demon> So "push" on mediawiki/ for i18n will let them bypass on any subproject? [20:52:01] depending on where it's put, yes [20:52:28] I think if it's in refs/*, it'll let them bypass review [20:52:43] refs/for/refs/heads/* <— that's the spot otherwise [20:52:54] <^demon> It's on refs/* [20:53:07] that'll allow bypassing review [20:53:11] <^demon> https://gerrit.wikimedia.org/r/#admin,project,mediawiki,access - that's the root project [20:53:21] yeah, add a new ref [20:53:28] called: refs/for/refs/heads/* [20:53:52] hm [20:53:57] maybe it's supposed to be refs/for/refs/* [20:54:09] anyway, everyone has the ability to push to everything [20:54:13] https://gerrit.wikimedia.org/r/#admin,project,All-Projects,access [20:54:30] for review, anyway [20:54:30] <^demon> Ah ok, so I shouldn't need to change that. [20:54:46] but you need to add push at refs/* for those you'd like to bypass review [20:55:01] <^demon> Right, the only group passing review is i18n. [20:55:05] * Ryan_Lane nods [20:55:52] <^demon> Everyone for core and extensions (sans i18n) should be push-for-review by default. [20:57:27] yep [20:57:35] that's how it's set up for all projects [20:57:42] should the group name perhaps be l10n instead? [20:58:20] since we are just pushing translations (=l10n), not necessarily abusing the powers to make i18n fixes [20:58:43] <^demon> There's no way to rename/delete groups. [20:58:47] <^demon> i18n it is :p [20:58:52] oki [20:59:04] sounds very flexible [20:59:48] I hope it's possible to change group members afterwards :D [21:00:01] <^demon> We can change members, yeah :p [21:02:16] hm. no group deletion? [21:02:43] <^demon> No, there's a bug open for it :\ [21:03:03] <^demon> Nikerabbit: https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/Translate.git;a=summary :) [21:03:29] <^demon> Ryan_Lane: How do you change an owner after the fact? [21:06:36] Ryan_Lane: did you have reason to think the encoding issue came from boto vs. nova proper? [21:07:12] <^demon> Nikerabbit: First change for Translate extension :) https://gerrit.wikimedia.org/r/#change,2558 [21:07:17] BotoServerError: 500 Internal Server Error [21:07:27] ah. [21:07:29] though, it's possible it was from nova and not boto [21:08:07] I'm not clear on where boto fits into the toolchain. It's used by euca2ools? [21:09:06] ^demon: :ooo [21:14:19] andrewbogott: yeah, and is used by ec2 api [21:19:39] Nikerabbit: ok. LeslieCarr fixed brewster, so let's try to build your instance again :) [21:20:18] * Damianz gives LeslieCarr a cookie [21:20:19] just warning I'm heading zZZ soon [21:20:26] yay! [21:20:29] yeah. that's cool [21:20:31] I just recreated the instance [21:20:34] hopefully it'll build [21:20:51] I need to add a critical warning for apt repo being broken [21:20:59] because it's actually a pretty nasty problem now [21:21:20] I have a feeling i know why the resolv.conf was empty [21:21:36] / has been filling up, due to lighttpd's access log [21:21:51] puppet likely was in the middle of a run, and overwrote it [21:22:04] but since the stupid filesystem was full... [21:22:35] \o/ [21:22:38] instance is building [21:22:43] lol [21:22:55] fuuuuuuck [21:23:05] I hate it when that happens, did it with a hosts.allow file on like 20servers a few weeks ago... really wasn't pretty. [21:23:19] At least puppet is self healing if you fix the fs :D [21:23:25] notice: Did not receive certificate[0m [21:23:27] * Ryan_Lane sighs [21:23:38] ah. next puppet run got it [21:23:52] I really need to make the key signing faster [21:25:10] Ahhh I know why thunderbird is laggy.... I have 278thousand emails in a folder :( [21:25:14] hahaha [21:25:27] Damn broken nagios install restarted before I fixed the script... [21:26:53] Damianz: is your puppetizing finished ? I want to reload the puppet host [21:27:01] technically the repo host [21:27:04] You mean Nikerabbit? [21:27:10] but that will obviously kill your build of your new instance [21:27:32] Don't talk to me about puppetizing, I've not won the argument with dpkg on packaging up cbng so my instance is no way puppetized :P [21:27:43] it's done :) [21:27:45] oh i did mean Nikerabbit [21:27:46] hehe [21:27:56] i should follow the cookies :) [21:29:08] Anyone know how I can make exchange work? Damn stupid unstandard standard thing... why can't I just do a find on a mail dir like a sane mailbox :( [21:29:12] Nikerabbit: your instance is up and can be logged into now [21:29:21] what's exchange? [21:29:23] :) [21:30:05] I want a zimbra cluster so the exchange fans can use that stuff and I can just go delete stuff like normal :P [21:30:13] Silly windows collegues. [21:30:29] why not just switch to gmail? [21:31:21] Ryan_Lane: confirmed [21:31:25] great [21:31:29] Good quesiton... apps would probably be cheaper but I think some of the mailboxes are larger than their limit. [21:31:33] you can sudo to root. it's your wiki password [21:31:39] now I only need to figure out how to break it^^what do with it :) [21:32:04] Damianz: you can get larger mailboxes [21:32:43] Interesting, might be a suggestion when I finish break...upgrading the helpdesk integration stuff :) [21:33:43] I don't know why anyone has their own hardware anymore [21:33:44] PROBLEM Current Load is now: CRITICAL on translation-memory-1 translation-memory-1 output: CHECK_NRPE: Error - Could not complete SSL handshake. [21:34:24] PROBLEM Current Users is now: CRITICAL on translation-memory-1 translation-memory-1 output: CHECK_NRPE: Error - Could not complete SSL handshake. [21:34:50] Exchange stuff is outsourced anyway, the only stuff we keep in house email wise is filtering which is only because most the 3rd party solutions suck for what we need. [21:34:54] I get why we do, but for like 99.99999% of businesses, it's really kind of unecessary [21:35:04] ah [21:35:04] PROBLEM Disk Space is now: CRITICAL on translation-memory-1 translation-memory-1 output: CHECK_NRPE: Error - Could not complete SSL handshake. [21:35:34] Apart from a few people who have windows desktops everything is totally linux/bsd in our network which is win :D [21:35:42] heh. yeah [21:35:44] PROBLEM Free ram is now: CRITICAL on translation-memory-1 translation-memory-1 output: CHECK_NRPE: Error - Could not complete SSL handshake. [21:36:54] PROBLEM Total Processes is now: CRITICAL on translation-memory-1 translation-memory-1 output: CHECK_NRPE: Error - Could not complete SSL handshake. [21:37:34] PROBLEM dpkg-check is now: CRITICAL on translation-memory-1 translation-memory-1 output: CHECK_NRPE: Error - Could not complete SSL handshake. [21:39:25] why does it complain about that? [21:42:04] Check /var/log/messages [21:46:45] Nikerabbit: because nrpe needs to be restarted [21:47:00] the initial puppet run when a system builds is slightly broken [21:47:09] the second puppet run restarts nrpe, I believe [21:51:40] <^demon> Ryan_Lane: By the way, git-review is awesome. It removes the need for ./git-setup files [21:51:47] yep [21:51:47] <^demon> And git push-for-review becomes `git review` [21:51:52] indeed [21:51:57] we can thank roan for that :) [21:52:02] <^demon> Yeah, forreal. [21:52:05] and the openstack people [21:52:12] <^demon> You configure parts of it with a .gitreview file [21:52:28] the openstack people using gerrit is another reason I'd prefer to use it rather than phabricator [21:52:57] phabricator is kinda strange [21:53:04] yeah [21:53:09] it's just a diff server, eh? [21:53:16] <^demon> phabricator is prettier than git :p [21:53:20] yes, it is [21:53:21] <^demon> s/git/gerrit/ [21:53:24] Yeah it is [21:53:36] But it seems what could have been on one page is split all over the place. [21:53:41] the interface is much better, but gerrit's backend is much better [21:53:52] Also the php servers for doing stuff like mail ewww [21:56:19] <^demon> Only drawback I've found so far with git-review is that it's totally unintuitive how to install it on windows. [21:56:29] <^demon> I fired up an XP vm this morning but gave up after about an hour. [21:56:45] I find that with windows in general :P [21:56:51] oh. well. that sucks [21:57:14] it would be nice if it was packaged for windows [21:57:34] it's going to be in the default ubuntu repo for precise [21:57:34] <^demon> We should definitely see if there's a way to lower that bar for windows, yeah [21:57:53] I hope it has roan's patches in it [21:58:10] <^demon> We don't have a *lot* of windows developers but it would be nice to not give them the finger :p [21:58:15] agreed [21:58:27] phabricator mostly requires php to be installed [21:58:32] so it has a similar issue [21:58:51] <^demon> presumably most of our development community already has php installed ;-) [21:58:54] unfortunately no luck in making that easier :) [21:58:56] not necessarily [21:59:03] a bunch of our repos are python [21:59:15] Yeah that's a really PITA for people on macs :P [21:59:18] I don't have php installed on my mac [21:59:20] <^demon> True true. [21:59:40] Ryan_Lane: You're totally missing out on extra apps to leak memory all over you [21:59:44] ok. I take that back it's installed by default [21:59:58] Hmm [22:00:06] <^demon> Is the default version finally at least 5.3.x? [22:00:17] 5.3.6 in lion [22:00:20] thankfully [22:00:27] Is puppetd suppose to honor /etc/sysconfig/puppet for the hostname or is it suppose to be set in /etc/puppet? [22:00:29] the version they had in snow leopard was a broken POS [22:00:38] I opened a bug with them years ago about that, and they never fixed it [22:00:44] <^demon> I think they were shipping 5.3.1 in snow leopard. [22:00:55] Damianz: eh? which hostname? [22:01:06] Puppet master [22:01:15] ah [22:01:23] /etc/puppet/puppet.conf [22:01:30] Looking at ours the answer is /etc/ [22:02:49] <^demon> Ryan_Lane: I've got 16 bugs "starred" on gerrit's issue tracker now :p [22:03:57] heh [22:04:22] we should open one to add support for "freeform tags" [22:04:31] <^demon> There's already one open [22:04:34] oh. good [22:04:41] I'd really like to have that [22:05:04] I'd love to be able to tag a change in one branch with "needs-cherrypicking" [22:05:06] <^demon> issue 287: "arbitrary labels/tags on changes" [22:05:17] so that I know it hasn't been moved to production [22:05:38] <^demon> I also starred "Forbid self +1 reviews" [22:05:48] <^demon> Which would be kinda cool too. [22:05:50] would also be good for after-the-fact reviews [22:05:59] as long as it's per-project :) [22:06:07] ops self-reviews often [22:06:23] and we use +1 to tell others to review something for us [22:06:30] we're weird, I know [22:06:42] <^demon> There's also "require two reviewers to +2 before merge" [22:06:48] <^demon> Which would be cool. [22:06:49] yeah. that would be nice [22:06:53] openstack wants that too [22:06:58] <^demon> "Default reviewers" [22:07:02] <^demon> Would be nice [22:07:24] I'd really like to model our "core reviewers" policy after openstack's [22:07:24] <^demon> They've also got an enhancement request to write a decent repo browser into gerrit since gitweb sucks :p [22:07:37] heh [22:07:39] that would be nice [22:07:58] <^demon> "Custom dashboards with saved queries" would be so amazingly useful. [22:08:03] in openstack, to be a core reviewer for a project, you have to be nominated [22:08:15] <^demon> "My main use case for this screen would be to check the current state of my projects in the morning when I come into office to get a quick overview of what happened while I was gone. " [22:08:16] then a certain number of current core reviewers have to give a yes vote [22:08:54] Review the reviewers but who reviews the reviewers that are reviewing the reviewers. [22:08:54] (core reviewers being people who can +2 and submit) [22:09:34] yeah, custom dashboard would be great [22:09:36] <^demon> Ryan_Lane: We were talking Friday about how to more firmly define the "who can be a reviewer" policy. [22:09:50] I really love the community voting way of things [22:09:51] <^demon> I'd be fine with something like openstack, "be nominated and get a vote of confidence" [22:10:04] it means *we* aren't the bad guys [22:10:57] Hmm puppet still confuses the hell outta me :( [22:10:59] <^demon> Maybe we can finally trick Roan into doing some gerrit development :p [22:11:02] they also have votes for project leads. I don't think tat would work for us :) [22:12:44] yeah, would be good for someone with some java knowledge to do so [22:12:51] I know java, but haven't used it in ages [22:13:01] Wait, gerrit is java? [22:13:17] yep [22:13:26] No wonder it behaves in weird ways :P [22:13:42] what I don't get is that it's a google project, and the interface is so, so terrible [22:14:07] that's one thing they generally do very well [22:14:14] Yeah [22:14:24] Though google groups isn't that pretty [22:14:47] <^demon> Ryan_Lane: It was written by a bunch of android engineers :p [22:15:28] I still need to shift that puppet class to a misc class as it's been wrongly merged for ages sigh... [22:15:29] heh [22:15:43] the test branch as a whole is utterly fucked [22:15:51] lol [22:15:51] we really need to move to per-project branches [22:16:14] apparently test is currently unmergable with production, in ma rk's opinion [22:16:41] I still don't really see how we're going to manage to use puppet in test/production and keep mw/labs production together... [22:17:11] what do you mean? [22:18:12] puppet should get much easier to manage soon. we're moving to a much more flexible model [22:18:22] * andrewbogott does not hate the gerrit interface so much [22:18:23] It just seems like a screwed idea until labs is test/production not just test. Currently it's kinda hacky with not that much in puppet which if it was it would be easy to merge and roll out boxes into a production cluster. [22:18:32] Puppet is just confusing :P [22:18:48] andrewbogott: it's not the worst in the world, but it's pretty bad [22:19:09] sure, could be better. [22:19:36] Damianz: yeah. I also dislike test [22:19:49] it means everything needs to be code reviewed before it can run [22:20:00] rather than that, we're going to have a branch for every project [22:20:15] Ryan_Lane: my dumb question of the day is: Is there an implicit/standard string encoding for http? Does it always use utf8, or does it vary? [22:20:17] :D [22:20:21] you test your changes inside of your project, then push those changes to test for review [22:20:25] In my eyes the point of test is stuff breaks. [22:20:34] Once it's stable it goes into the main production branch [22:20:39] andrewbogott: I don't think there is any implicit string encoding [22:20:44] hm [22:20:47] Damianz: yep [22:20:53] Damianz: that's the idea [22:20:57] It's specified in a header somehow? [22:21:25] so, make changes and test stuff in project-branch, push to test, we code review, and test the changes in a cluster-clone, then cherry-pick to production [22:21:30] andrewbogott: I think so.... [22:21:40] Tbh I don't really see the point of nagios in test - it makes sense right now because we don't have production but makes little sense when we are going in a test -> production style of work. [22:22:02] New patchset: Lcarr; "Adding in new facter fact of default interface This is the interface bound to default gw" [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2561 [22:22:18] andrewbogott: Content-Type: [22:22:35] so, for instance: Content-Type: text/html; charset=UTF8 [22:22:52] that makes sense. [22:22:55] Damianz: it's for other things [22:23:01] I wonder if nova observes that. [22:23:01] Damianz: like bots, etc [22:23:26] Ah [22:24:05] we're going to have ganglia for all instances [22:24:08] ganglia totally makes sense :) [22:24:21] would be good to see which projects are eating the most resources [22:24:50] bots :P [22:26:00] probably [22:26:02] and deployment-prep [22:29:26] Ryan_Lane: why does gerrit on labs occasionally just never review your entries ? [22:29:29] other than it hates me :) [22:29:51] the hooks are buggy [22:29:57] I haven't gotten a chance to fix them [22:30:06] ok [22:30:07] cool [22:30:33] as long as it's not just me ;) [22:30:40] New review: Lcarr; "(no comment)" [operations/puppet] (test); V: 1 C: 2; - https://gerrit.wikimedia.org/r/2561 [22:30:40] Change merged: Lcarr; [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2561 [22:30:55] nah. I'll get around to fixing them one of these days [22:31:07] oh. hey. the gluster systems came in, didn't they? [22:32:23] hm. I wonder if we can add nova metadata items to facter [22:32:58] then we wouldn't need to inject them into ldap as variables [22:33:15] we could just inject into the metadata server, then have facter variables for it [22:33:20] one less thing we need ldap for :) [22:34:35] Oooh are you actually using facter? [22:34:53] yeah [22:35:24] andrewbogott: It may be doable for us to completely ignore nova for puppet, and just use metadata [22:36:08] we could inject classes and variables into the metadata, then have cloud-init build a site.pp using a script, then run puppet against it [22:36:16] Yeah, I'm pretty sure we can do that. It means we keep the solution to ourselves, but that's tolerable. [22:36:19] 02/13/2012 - 22:36:18 - Creating a project directory for leslie [22:36:19] 02/13/2012 - 22:36:18 - Creating a home directory for lcarr at /export/home/leslie/lcarr [22:36:21] yeah [22:36:39] LeslieCarr: you made a leslie project? :D [22:36:47] :D [22:36:49] how very non-community of you ;) [22:36:50] yep :) [22:36:52] haha [22:36:57] it's my own magical domain! [22:37:02] where i can break everything :) [22:37:02] hahaha [22:37:14] oh. I need to make another bastion instance [22:37:19] 02/13/2012 - 22:37:19 - Updating keys for lcarr [22:37:27] I'm going to separate access between two bastion instances [22:37:33] cool [22:37:41] anyone with shell access on the cluster must use one, while everyone else must use the other [22:37:50] hehe was just about to ask that :) [22:37:52] cool [22:37:56] and I'll forbid access using access.conf [22:38:08] that way we don't have to worry about escalation attacks from the bastion instances [22:38:57] Ryan_Lane: We would have a custom cloud-init? Or, do we already have one? [22:39:14] well, cloud-init lets you run any scripts you want [22:39:16] Won't you need shell access to bastion to login to it? Or do you mean to use as a socks proxy etc to access internal semi-public stuff? [22:39:19] you inject them as user-data [22:39:35] Oh! Well, that [22:39:38] that'll work then [22:39:51] Damianz: we'd have bastion.wmflabs.org, and bastion-ops.wmflabs.org, or something like that [22:40:14] yeah. we're using cloud-init scripts for some stuff right now [22:40:28] Ahhh [22:40:29] we can make this something others can use [22:40:32] it'll be ubuntu only, though [22:40:43] unless fedora has cloud-init [22:41:25] it may be possible to do a number of things without extending nova too much [22:42:37] Yep. I'm looking at the subpubhubbub (sp?) stuff, and that may get us the stats we want without much nova modification. [22:43:02] New patchset: Lcarr; "Beginning slash!" [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2562 [22:43:07] sweet [22:43:10] that would be great [22:43:36] subpubhubbub, yeah. thank google for that stupid name [22:43:52] ah: pubhubsubbub [22:44:03] New review: Lcarr; "(no comment)" [operations/puppet] (test); V: 1 C: 2; - https://gerrit.wikimedia.org/r/2562 [22:44:03] Change merged: Lcarr; [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2562 [22:44:23] PROBLEM Current Load is now: CRITICAL on test1 test1 output: Connection refused by host [22:45:03] PROBLEM Current Users is now: CRITICAL on test1 test1 output: Connection refused by host [22:45:04] I really need to add more resource pages. security groups would be a good one [22:45:10] I'd like to have SMW internal objects [22:45:16] I need to get the SMW people to help us [22:45:29] then we could do queries like: show me all instances that have the web port open [22:45:43] PROBLEM Disk Space is now: CRITICAL on test1 test1 output: Connection refused by host [22:45:55] :D [22:46:13] PROBLEM Free ram is now: CRITICAL on test1 test1 output: Connection refused by host [22:46:15] using a compound query, of course [22:46:45] but then changes to security groups would also show up in recent changes [22:46:58] need to do the same for addresses [22:47:27] Hmm I don't suppose there is a way to make puppet understand access in a way that means if it isn't referenced directly or indirectly to the host then the host don't get access to the file/manifest? Limiting by ip would work but might be a pita to keep up with until I finish the ip management stuff :( [22:47:33] PROBLEM Total Processes is now: CRITICAL on test1 test1 output: Connection refused by host [22:48:00] there's ip based, yeah [22:48:06] that's about it, though. [22:48:11] ask the puppet people [22:48:12] <^demon> hashar: What's the status on php-linting as a pre-commit hook in gerrit? [22:48:13] PROBLEM dpkg-check is now: CRITICAL on test1 test1 output: Connection refused by host [22:48:23] Puppet people are scary, they start talking ruby :P [22:48:24] there's no such thing as a pre-commit hook in gerrit [22:48:36] <^demon> You know what I mean [22:48:39] <^demon> And so does he :p [22:48:43] I do? :) [22:48:51] do you mean as a patchset-submitted hook? [22:48:59] <^demon> Yeahhh [22:49:37] ^demon: no ETA yet [22:49:58] ^demon: I am not sure if I am going to have it handled through a jenkins jobs or locally [22:50:21] <^demon> No need to involve jenkins for linting [22:50:28] Btw is the puppet lint a gerrit thing or a jenkins thing? [22:50:29] <^demon> Should do it in-gerrit like we do on the puppet repo [22:51:30] it really would be nice if gerrit had a pre-commit hook :( [22:51:33] hrm, any idea why i'm getting permission denied in my new project ? [22:51:44] permission denied sshing to the instance? [22:52:08] did it build properly? [22:53:39] I wonder how bad I'm going to screw things up when I move from NFS to gluster for everyone's home directories. heh [22:53:40] looks like it [22:53:48] what's the instance name/ [22:54:00] >. [22:54:00] > [22:54:15] Ryan_Lane: Make sure you try that on the test cluster first ^.^ [22:54:17] :P [22:54:23] heh [22:54:39] I don't test often, but when I do, I do it live in production [22:55:01] If it compiles it works right? [22:55:06] heh [22:55:29] I don't know how automount is going to handle the filesystem switch [22:55:32] we'll likely have to reboot all instances [22:55:51] test2 [22:55:57] test failed [22:56:04] or test1 (which looks like it mayb have not built correctly) [22:56:21] yeah. test1 doesn't let me in [22:56:22] test2 does [22:57:07] (I have a way to let me in directly as root ;) ) [22:57:08] oh now it does [22:57:09] 02/13/2012 - 22:57:09 - Updating keys for aaron [22:57:14] maybe i should have been patient [22:57:18] just hadn't finished building? [22:57:42] maybe [22:57:58] cool, well now i know my ruby hack didn't work , so yay :) [22:58:07] it will be a real yay when it does work [22:58:13] Ryan_Lane: The only reason you don't give console access is because all the root passwords are 'password'? :P [22:59:29] New patchset: Lcarr; "try 2" [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2565 [22:59:55] New review: Lcarr; "(no comment)" [operations/puppet] (test); V: 1 C: 2; - https://gerrit.wikimedia.org/r/2565 [22:59:56] Change merged: Lcarr; [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2565 [22:59:56] Damianz: heh [23:00:08] Damianz: the root password is some insanely long string [23:00:14] and no one knows it [23:00:19] it's ************* [23:00:20] hell, I might not even know it [23:00:35] Damianz: we don't have console access because I haven't set it up [23:00:35] so anyone here know ruby ? [23:00:40] and it isn't high on my priority list [23:00:59] LeslieCarr: only as much as the web teaches me as I'm writing something [23:01:12] puts "Hello world" [23:01:24] :) [23:01:26] yeah [23:01:42] I've never put much effort into learning ruby [23:01:43] i don't really at all… but trying to modify a script for facter … i figured it'd be easy ! [23:01:48] heh [23:01:48] but maybe not so much :( [23:01:57] Ruby makes kittens cry [23:01:59] it should be easily learnable in a week or so [23:02:12] it's apparently no harder than python [23:02:17] i want to take https://raw.github.com/kwilczynski/facter-facts/master/default_gateway.rb and just have it find the gateway line and return the "iface" field [23:02:23] Python's easy :P [23:02:39] yeah. that's my point :) [23:03:00] I don't understand why ruby tries so hard to be like english which imo makes it harder to use :( [23:03:13] gateway = values[1].unpack('C4') [23:03:52] that's how i speak ! [23:04:14] so you want the gateway column? [23:04:49] the iface column [23:04:57] I don't see a gateway line [23:06:21] oh so the line that has information in the gateway column is what i am referring to as the "gateway line" [23:07:16] ahh. ok [23:07:57] so, you want to return values[0[ [23:07:58] err [23:08:38] yeah [23:08:41] values[1].unpack('C1') [23:09:03] unless that doesn't do what I think it does. heh [23:12:22] 02/13/2012 - 23:12:21 - Creating a home directory for ben at /export/home/leslie/ben [23:12:59] actually, no. values is only two columns [23:13:07] values = line.split("\t").slice(1, 2) [23:13:20] 02/13/2012 - 23:13:20 - Updating keys for ben [23:13:27] so, values[0] is the gateway [23:13:31] err [23:13:33] destination [23:13:39] and values[1] is the gateway [23:13:49] so, you'd need values = line.split("\t").slice(0, 2) [23:13:57] then values[0] would be the Iface [23:14:34] ewwww. unless? [23:14:45] that's a really awkward conditional [23:14:58] Hmm ben seems to be in leslie [23:15:00] * Damianz giggles [23:15:14] i better not tell his wife ! [23:15:22] :D [23:16:12] <^demon> Crap. [23:16:15] <^demon> CRAP CRAP CRAP [23:16:17] ? [23:16:20] <^demon> I hate you gerrit [23:16:22] :D [23:16:25] what's wrong? [23:16:29] <^demon> I pushed the wrong repo to mediawiki/core.git [23:16:30] Share the lov..hate! [23:16:30] <^demon> Ughhhh [23:16:31] <^demon> Noooo [23:16:36] hahaha [23:16:52] <^demon> I can obliterate all of the branches. [23:16:55] so, you *can* rebase it [23:16:59] <^demon> But there's no way to kill the master. [23:17:02] since it isn't in use right now [23:17:11] you need the "force push" option checked for yourself [23:17:17] or one of your groups [23:17:37] and you should do a force rebase bypassing review [23:17:39] Rebase doesn't get rid of history though does it? I thought it just changed the head ref. [23:17:49] rebase will let you wipe history [23:18:13] in general you should *never* do it to a repo that other people have pulled from [23:18:16] unless you hate them [23:18:27] lol [23:19:01] It's always fun when your collegue goes and force pushes the wrong branch to master and causes the CI servers to go bonkers for being totally confused xD [23:19:48] <^demon> Yeah, I had to give myself force push annotated tag. [23:20:14] hey, at least it's possible, right? :) [23:20:17] can't totally hate gerrit [23:20:38] <^demon> Yeah, it's not totally gerrit's fault :p [23:20:39] it's not the most intuitive thing in the world, though [23:21:02] <^demon> `git push origin :refs/tags/foo` is the least intuitive "delete a tag" syntax ever :) [23:21:06] There's allways just going into the database with DELETEs :P [23:21:44] unfortunately, not [23:21:55] <^demon> Ryan_Lane: You see why I starred the "please add a way to delete a project" bug? :p [23:22:09] though, you *could* go to the system, and rebase the repo directly [23:22:22] <^demon> That's the described workaround right now. [23:22:25] yep [23:22:36] <^demon> Go onto the system, rebase the hell out of it, then poke through the database to delete things. [23:23:04] yeah, the fact that you can't delete repos is really fucking annoying [23:23:27] Skynet will live forever [23:23:38] <^demon> Ryan_Lane: Not like we're any better. [23:23:42] <^demon> Go delete a mediawiki user ;-) [23:24:30] right [23:24:42] lol [23:24:52] hell, try to delete a wiki [23:25:01] we have no way of doing that right now [23:25:08] I'd love all of the closed labs wikis to be deleted [23:26:27] <^demon> DROP DATABASE tlhwiki; [23:26:29] <^demon> ;-) [23:26:44] See how badly that would fuck up 3 dozen things :P [23:27:21] !create-account [23:27:30] Does that even work? [23:27:45] @search account [23:27:45] Results (found 4): credentials, account-questions, account, accountreq, [23:27:53] @search usr [23:27:53] No results found! :| [23:27:56] @search user [23:27:56] Results (found 4): account-questions, new-labsuser, new-ldapuser, socks-proxy, [23:27:59] !accountreq [23:27:59] in case you want to have an account on labs, please contact someone who is in charge of doing that: Ryan.Lane, m.utante or ssmolle.tt [23:28:09] !account-questions [23:28:09] I need the following info from you: 1. Your preferred wiki user name. This will also be your git username, so if you'd prefer this to be your real name, then provide your real name. 2. Your SVN account name, or your preferred shell account name, if you do not have SVN access. 3. Your preferred email address. [23:28:38] Not even a promise not to setup a torrent tracking ;P [23:28:44] heh [23:29:01] we're going to have a privacy policy and a terms-of-use agreement soon [23:29:04] Also ISPs that block bittorrent ports should die [23:29:17] we might have to require people send us id copies [23:29:21] Ryan_Lane: Starting to sound like the TS :P [23:29:45] well, wikimedia's privacy policy, as set by the board, is very strict [23:29:56] and labs falls under that [23:30:01] we'll know more friday [23:30:17] I assume that will matter more when we have stuff like db replicas up for access? [23:30:20] I thought it was saturday? [23:30:27] New patchset: Lcarr; "Ben made default_gateway.rb report default_interface" [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2569 [23:30:34] I have a meeting with legal on friday [23:30:39] Ah [23:30:41] hee david [23:30:47] !account-questions [23:30:47] I need the following info from you: 1. Your preferred wiki user name. This will also be your git username, so if you'd prefer this to be your real name, then provide your real name. 2. Your SVN account name, or your preferred shell account name, if you do not have SVN access. 3. Your preferred email address. [23:30:50] we have an online meeting with the community saturday [23:31:03] !account-questions | dschoonover [23:31:03] dschoonover: I need the following info from you: 1. Your preferred wiki user name. This will also be your git username, so if you'd prefer this to be your real name, then provide your real name. 2. Your SVN account name, or your preferred shell account name, if you do not have SVN access. 3. Your preferred email address. [23:31:04] :) [23:31:14] thanks ryan [23:31:16] yw [23:31:18] I wouldn't want to work in the mwf legal department... can only imagine the hell they have policing cc-by-sa content [23:31:26] they don't police [23:31:31] the community does [23:32:02] Yeah but someone has to be responsible after the community, admins etc :P [23:32:36] * Ryan_Lane nods [23:33:27] You should just employ rms to cover all legal issues, he'd talk anyone to death. [23:34:50] New review: Lcarr; "(no comment)" [operations/puppet] (test); V: 1 C: 2; - https://gerrit.wikimedia.org/r/2569 [23:34:51] Change merged: Lcarr; [operations/puppet] (test) - https://gerrit.wikimedia.org/r/2569 [23:35:13] Ryan_Lane: where do i send that info? [23:35:21] To Ryan_Lane [23:35:29] Or the other 2 people that I can never remember. [23:35:57] Damianz: Thanks :) [23:37:20] Ryan_Lane: is there any documentation for petr's new bot? thinking about using it for fundraising stuff but have no idea what the feature set is [23:37:27] <^demon> Ryan_Lane: How would I go about obliterating rebasing to obliterate the entire history on master? [23:46:56] pgehres: umm [23:47:04] pgehres: you mean the one in here that answers questions? [23:47:07] yeah [23:47:20] does it not replicate most of the functionality of morebots as well? [23:47:41] you'd need to ask him. it has terms, channel logging, and some other things [23:47:42] Dunno, check the source code it's in svn. [23:47:46] !log [23:47:50] !log kfkf [23:47:50] Message missing. Nothing logged. [23:47:56] no. morebots is different [23:48:07] I re-wrote morebots, basically [23:48:18] is that in svn as well? [23:48:20] did I ever check that in? :) [23:48:31] I know, tsk tsk ;) [23:48:42] idk, if there aren't many more features I am not learning C# [23:49:09] it's just like java, basically [23:49:11] i've already written some add-ons to morebots for fundraising [23:49:14] oh, i know [23:49:21] I've looked at the source [23:49:22] I hate perl way more than C# [23:49:25] but I have Java too [23:49:34] I have Jeff_Green for perl :-D [23:49:37] heh [23:50:00] C# is nicer than java. [23:50:26] yeah, but everyone I know who loves C# also loves Visual Studio… [23:51:45] and right now we (fundraising) have no compiled code and would rather not install mono if we don't have to... [23:51:51] <^demon> Everyone I know who loves C# ends up dead ;-) [23:52:10] * Ryan_Lane nods [23:52:35] ^demon: just to be clear, as of yet I have no opinion and thus should not be on your hitlist [23:52:59] morebots is the python one, correct? [23:56:16] <^demon> Ryan_Lane: I'm completely lost as to how to clear this history with rebase. [23:56:31] ^demon: I also don't know git well enough for this [23:56:35] sorry... [23:56:55] <^demon> *sigh* [23:57:36] <^demon> I could push origin :master if HEAD wasn't pointing at it. [23:59:37] ^demon: What is the final result you're aiming for? [23:59:58] gitwise, I mean?