[00:00:35] Ryan_Lane: Nah just ProfilerSimpleText [00:00:39] petan: ughhhhh [00:00:40] ah ok [00:00:46] we can switch over to the UDP one later [00:00:54] that's thankfully puppetized, from the daemon perspective [00:00:55] Ryan_Lane, when are you getting virt-1? [00:01:01] Reedy: eh? [00:01:05] * johnduhart runs git reset --soft [00:01:05] ah [00:01:06] heh [00:01:07] :p [00:02:27] Reedy: I'm using virt0 as the controller [00:02:33] and all nodes past that are compute [00:15:17] * Ryan_Lane groans [00:15:24] this is going to be a PITA [00:15:44] * Ryan_Lane checks whois [00:16:03] Name Server:VIRT1.WIKIMEDIA.ORG [00:16:03] Name Server:LABSCONSOLE.WIKIMEDIA.ORG [00:16:05] * Ryan_Lane sighs [00:16:25] fail [00:16:43] well, thankfully the secondary is acceptable [00:32:52] I'm switching the controller.... [00:33:06] there's a possibility during this process I'll fuck up, and we'll have some downtime [00:34:15] no matter what we're going to have some labsconsole downtime [00:34:24] at least from the perspective of modifications [00:48:17] Ryan_Lane: sorry i may have missed a reply yesterday [00:48:25] is there any news about my Labs account? [00:49:12] you didn't provide me with all info [00:49:42] !account-questions | philinje [00:49:42] philinje: 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. [00:50:59] 1. philinje 2. philinje 3. pchang@wikimedia.org [00:54:38] !initial-login | philinje [00:54:38] philinje: https://labsconsole.wikimedia.org/wiki/Access#Initial_log_in [01:17:53] dear opendj, you just scared the hell out of me: Initializing schema on server virt1.wikimedia.org:4444 with the contents of server virt0.wikimedia.org:4444 ..... Done. [01:17:59] I didn't see "schema" [01:18:07] I thought it was wiping out all the data. heh [01:18:33] well, ldap server is done [01:26:28] Ryan_Lane1: thanks [01:26:33] yw [02:23:22] mysql done [03:04:55] Ryan_Lane: wy no labs? [03:05:02] whats' broken? [03:05:36] hexmode: ?? [03:05:49] Ryan_Lane: I can get to google but not deployment [03:06:05] hm [03:06:07] crap [03:06:10] is dns broken? [03:06:27] "ssh: Could not resolve hostname bastion.wmflabs.org: Name or service not known" [03:06:48] pdns on virt1 isn't responding.... [03:07:20] hmm. wtf [03:08:06] ah. I know [03:08:24] one quick fix. [03:08:31] ok. working now [03:09:32] well. it should be anyway [03:10:07] yeah. it is [03:12:36] ok. did the permanent fix too [03:13:00] and to think I thought I was going to do this with no downtime! [06:17:54] Ryan_Lane: mysql is broke? [06:18:00] * Ryan_Lane groans [06:18:15] I moved all services from virt1 onto virt0 [06:18:32] i saw :) [06:18:46] seems something is wrong with apache or mysql [06:18:51] but that was 3 hours ago. and mysql brokenness was discussed [06:18:59] Ryan_Lane: i think it's missing grant [06:19:03] it was [06:19:22] i mean what i'm getting now is missing grant [06:19:23] it's taking ages to load right now [06:19:26] or wrong password [06:19:35] I'm not anymore. I just added it [06:19:38] but it's timing out now [06:19:52] memcached not running or wrong port/address? [06:20:10] (we had that problem with deployment-prep i think) [06:20:13] did you just connect? [06:20:21] that was DNS a little while ago [06:20:35] for a short period of time. [06:20:56] hmm. It's loading, but it's taking a really long time... [06:21:11] agreed, long time [06:21:53] wtf. [06:23:30] doesn't look like this recovered per my channel buffer ----> 13 02:21:16 <+nagios-wm> PROBLEM - Memcached on virt0 is CRITICAL: Connection refused [06:25:17] I can connect to it on localhost [06:25:27] and that's how mediawiki is accessing it [06:26:14] spence is slow... [06:27:25] found it [06:27:41] seems memcached used a different port on virt1 [06:27:50] mediawiki was improperly configured [06:28:26] this one is using the cluster's port, the other was using the default for some reason. [06:28:27] sooo fast [06:28:41] heh. yeah. way better now [06:29:15] there's some issue with the zones now too [06:29:25] when I added a new scheduler it added a new zone [06:29:32] scheduler? [06:29:37] the correct zone, but a new one [06:29:44] nova-scheduler service [06:30:06] it chooses which compute node is going to run an instance [06:30:18] what's this? word missing? ----> the correct zone, but a new one [06:30:23] zones are different clusters of compute-nodes [06:30:40] now we have nova and pmtpa [06:30:54] what's the diff? [06:31:15] it was only nova before [06:31:28] though I did have the zone listed in the config as pmtpa for all of the nodes [06:31:48] the scheduler we are using is a single zone controller, so it's odd [06:31:53] it's a bug, likely [06:32:10] now to see if an instance will create in both zones and still work [06:32:16] what did it add? [06:32:35] pmtpa [06:32:49] which is what I wanted all along :D [06:32:55] the sidebar's not right on labsconsole [06:33:24] huh, fixed [06:33:34] I'm not sure what the deal with that is. I changed the version of vector to a tagged version for 1.18 [06:41:04] nacht [06:49:07] ugh, for some reason the floating ip for bastion got disassociated [07:29:37] Change on 12mediawiki a page Wikimedia Labs was modified, changed by Ryan lane link https://www.mediawiki.org/w/index.php?diff=484775 edit summary: /* Proposals */ [07:35:41] Change on 12mediawiki a page Wikimedia Labs was modified, changed by Ryan lane link https://www.mediawiki.org/w/index.php?diff=484776 edit summary: /* Proposals */ [07:57:07] 01/13/2012 - 07:57:07 - Updating keys for beetstra [07:57:09] 01/13/2012 - 07:57:09 - Updating keys for beetstra [07:58:07] 01/13/2012 - 07:58:07 - Updating keys for beetstra [07:58:09] 01/13/2012 - 07:58:09 - Updating keys for beetstra [07:59:07] 01/13/2012 - 07:59:06 - Updating keys for beetstra [07:59:09] 01/13/2012 - 07:59:08 - Updating keys for beetstra [08:00:14] :-) [08:35:31] Is there someone from the bots-group around? [08:36:01] I need some help about passwords used by a bot [08:42:22] !log bots Installed XLinkBot on bots-2, running properly [08:42:23] Logged the message, Master [08:47:32] PROBLEM host: deployment-test is DOWN address: deployment-test check_ping: Invalid hostname/address - deployment-test [09:18:12] PROBLEM host: deployment-test is DOWN address: deployment-test check_ping: Invalid hostname/address - deployment-test [09:21:44] !logs [09:27:33] !sal [09:27:33] https://labsconsole.wikimedia.org/wiki/Server_Admin_Log see it and you will know all you need [09:28:20] ok. time for bed. way too tired to do anything correctly [11:51:34] !logs [11:52:24] Can someone help setting up mysql for my perl bots on bots-2 ?? [11:52:56] I can't sudo sucessfully on deployment-wmsearch [11:53:06] any suggestions [11:58:24] Beetstra: SQL shouldn't be installed directly on that server [11:58:56] that is fine, so I have to connect to an 'external' server [11:59:27] that is, i.s.o. localhost, something else in the connect-statement [11:59:45] But I would need to set up tables somewhere, and then to connect to them ... don't know how [12:10:59] :o [12:11:02] Beetstra: hey [12:11:12] Hi! [12:11:24] Got the first bots running, perl is running fine there!! [12:11:28] I know [12:11:32] :-) [12:11:35] how many mb you need for your db [12:11:41] hehehehe [12:11:53] hm? [12:12:05] At first, I will start from scratch, just testing, so maybe a Gb to start with [12:12:25] ok I will create you an account on bots-sql1 with create [12:12:27] In the end, when I transfer the existing tables, it will be in the order of 100 Gb already [12:12:30] petan|w: I have permission issues on mwsearch [12:12:44] OrenBo: like? [12:12:53] cant su [12:12:53] Beetstra: want a service account too? [12:12:55] for bot [12:12:59] OrenBo: use sudo [12:13:00] no su [12:13:02] cant make dirs [12:13:05] with sudo [12:13:06] OrenBo: where [12:13:10] sec [12:13:16] OrenBo: because there is no root squash [12:13:19] petan|w .. what do I do with a service account? [12:13:25] Beetstra: for boy [12:13:28] * bot [12:13:41] Beetstra: so bot uses own acc for mysql [12:14:01] ehm .. the bots access each others tables ... [12:14:03] OrenBo: tell me which directory you need to create [12:14:17] But that is a good idea, as long as they can get to each others tables [12:14:25] sure [12:14:43] Beetstra: I will create a root account for you and standart acc [12:14:44] ok? [12:14:48] petan: botnames - linkwatcher, coibot, monitorbot [12:15:00] you want a maria or mysql? [12:15:02] OK, sounds good [12:15:07] oren@deployment-wmsearch:/usr/local/search/ls2$ sudo mkdir bin [12:15:08] mkdir: cannot create directory `bin': Permission denied [12:15:09] the bots are used to mysql [12:15:54] OrenBo: again :) [12:15:57] OrenBo: fixed [12:16:13] Beetstra: maria is same as mysql just a better version [12:16:32] OK, use the better one [12:16:34] :-D [12:16:41] any idea why ? [12:16:49] Enough troubles already with the old MySQL on the box they are running on atm [12:16:50] Beetstra: ok see bots.wmflabs.org/phpmyadmin [12:17:13] OrenBo: yes permissions were wrong [12:17:18] OrenBo: fixed [12:17:26] petan: not found .. [12:17:37] ah, wait [12:18:04] Beetstra: http://bots.wmflabs.org/phpmyadmin/ [12:18:23] OrenBo: anything else? [12:18:25] petan|w is there a jdk or just a jre on mwsearch ? [12:18:33] I don't know [12:18:36] checking [12:18:43] I get can't find tools.jar [12:20:15] petan .. error #1045 Cannot log in to the MySQL server [12:20:27] I presume it uses my (lowercase) username and standard password? [12:20:35] * Beetstra sees PM [12:20:55] no there is no ldap yet [12:21:05] we work on that with Ryan [12:21:29] I was looking at the path you supplied - there was no localsetting.php there [12:21:48] no there is wmf-config/CommonSettings.oho [12:21:49] php [12:21:54] we don't have LS.php on prod [12:21:55] in /usr/local/apache/conf [12:22:24] usr/local/apache/common/wmf-config [12:22:39] why not, how does it work? [12:22:41] before you make any change there let me know [12:22:47] OrenBo: all wikis use same config [12:22:52] check that file [12:22:58] ok [12:23:01] we have almost 1000 wikis on production [12:23:13] using a separate LS.php would be really hard [12:24:57] OrenBo: let me know if you needed me [12:25:04] ok [12:25:09] I am in work right now so probably won't be around quick [12:25:36] I'm trying evrything again [12:25:42] should take about a minute [12:25:46] OrenBo: use etherpad [12:25:47] ok [12:26:58] I stll have issue with tools.jar not found in jdk [12:27:04] ok [12:27:27] JavaHome point to a jre [12:35:40] OrenBo: I installed tomcat [12:35:46] that's probably what you were missing [12:37:09] why tomacat ? [12:38:02] Beetstra: installed [12:38:10] OrenBo: it was on search or not? [12:38:31] OrenBo: what libs are you missing, I thought you want all stuff you had on search [12:39:02] I got libs [12:39:06] ok [12:39:09] this jar is part of the jdk [12:39:10] so it works? [12:39:13] no [12:39:19] hm... I can't find jdk in repos [12:39:24] only jdk emacs [12:39:27] :O [12:39:31] :-( [12:39:52] jeremyb: ping [12:40:01] I remember we fixed it on search-test [12:40:30] !sal [12:40:30] https://labsconsole.wikimedia.org/wiki/Server_Admin_Log see it and you will know all you need [12:41:13] how do I get the value of the JAVA_HOME environment var [12:41:28] OrenBo: there is no log [12:41:34] regarding jeremy's setup [12:42:08] env | grep 'java_home' [12:42:56] * JAVA_HOME [12:43:00] it's case sensitive [12:43:53] env | grep 'JAVA_HOME=' | sed 's/HOME=//' [12:43:59] if you onl need the value [12:44:25] OrenBo: I installed some other try now [12:45:25] OrenBo: hm? [12:45:42] sec [12:45:43] what all tools you need to set up :o [12:46:11] good [12:46:50] thanks [12:47:27] ok [12:47:57] can you get junit 3.x into lob [12:48:00] can you get junit 3.x into lib [12:48:18] hm? [12:49:04] sure [12:49:11] installed [13:01:17] I think the btw jdk is not good :-) I run stuff on wrong machine [13:01:45] i.e. still not tools.jar in jdk [13:04:22] sudo apt-get install openjdk-6-jdk [13:04:28] but it fials with [13:05:06] Could not get lock /var/lib/dpkg/lock - open (11: Resource temporarily unavailable) [13:05:10] E: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it? [13:05:46] installing [13:05:47] done [13:07:06] ok [13:07:34] log that stuff [13:07:50] !log deployment-prep installed jdk on search [13:07:51] Logged the message, Master [13:08:23] ps test now build but 142 test fail [13:08:31] hm? [13:09:01] OrenBo: /usr/local/search/indexes [13:09:06] ?? [13:09:06] can I scp a file from my local machine [13:09:06] what is it [13:09:09] yes [13:09:24] !socks-proxy [13:09:25] see https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances ssh @bastion.wmflabs.org -D [13:10:02] hmm [13:10:05] complex [13:10:11] heh [13:10:23] OrenBo: there is /mnt/index [13:10:25] perhaps you can put the junit3 lib also into search-test instance [13:10:33] ok [13:12:09] petan|w: Can you please not blindly commit stuff https://github.com/johnduhart/deploymentprep-conf/commit/e04f979b5240e4260ba81e0aed4d20ac1b95eea6 [13:12:12] !log search installed junit 3.8 [13:12:12] Logged the message, Master [13:12:47] johnduhart: that wasn't from me [13:12:55] no idea who inserted it there [13:12:59] I probably commit -a [13:13:03] petan|w: I did, you committed it. [13:13:07] ah [13:13:13] petan|w: You need to check what's committed before you commit it [13:13:17] how [13:13:22] I don't know git [13:13:27] petan|w: You never added the file before you committed [13:13:28] git check? [13:13:33] * johnduhart sighs [13:13:36] I probably changed some [13:13:39] ok we have a problem [13:13:45] petan|w: Go read some git docs [13:13:53] if only I have a time [13:13:57] petan|w: http://progit.org/book/ [13:14:01] I know [13:14:04] git status is your friend [13:14:13] OrenBo: yes? [13:15:22] it realy want to find a LocalSettings.php [13:15:40] ok there is a LocalSettings but it doesn't include much [13:15:45] contain [13:15:50] it include CommonSettings [13:15:59] where is it [13:16:01] OrenBo: on prod there is no LS.php [13:16:16] OrenBo: /usr/local/apache/comm*/live [13:16:54] Talk to me about configuration [13:17:09] sec [13:19:57] johnduhart: 404 is working now [13:19:58] is CommonsSetting the same a LocalSettings just a different name [13:20:12] it's a replacement for that [13:20:15] or is it a different file format [13:20:23] it's a configuration using wgCong [13:20:25] Conf [13:20:27] for all wikis [13:20:35] check the source code [13:20:37] SiteConfiguration* [13:20:38] thanks [13:20:53] let me explain the issue [13:20:57] petan|w: Is it done how it is in production? [13:21:05] johnduhart: I don't have access on prod [13:23:52] it looks like it [13:23:56] but I don't know [13:24:39] actually I doubt there is a live folder on prod [13:24:45] just like /usr/local/apache [13:24:52] johnduhart: ^ [13:24:57] so probably it's not same [13:31:51] re [13:32:00] I had some buisneess to take care [13:33:27] search-2 looks for LocalSettings - its hard coded [13:33:58] I can add locations and tell it to look for CommonSettings [13:34:24] but the code for parsing the config is not in search - it is in MWdumper [13:34:30] another project [13:35:33] OrenBo: ask notpeter [13:35:39] production use Common [13:35:41] not Local [13:35:58] OrenBo: there is only one Local Settings and it's same for all wikis [13:36:23] which vars you want to get from it [13:41:31] btw OrenBo I don't want search to use db access which is being used by wiki [13:41:40] I would rather create a new acc [13:41:58] petan I don't follow you [13:42:09] you need to use sql? [13:42:17] I don't know [13:42:18] I don't want search accidentally damage db [13:42:26] so I wanted to create a new account for it [13:42:26] hmm [13:42:34] no need for it to be able to write to db [13:42:39] wiki tables [13:42:57] I am looking through the spagetti code [13:43:01] ok [13:43:54] I also don't have access porduction [13:44:14] I have no idea if it uses dbase access [13:44:19] hm [13:44:25] I don't know what it needs LS for [13:44:25] I'm 95% sure that it does [13:44:34] righjt [13:44:56] I am pretty sure I know what it uses it for [13:45:07] hm? [13:45:12] LS that is [13:45:35] give me a minute to open the source files [13:45:42] ok [13:47:25] hexmode: you still wait with announcement or not? why is that section what needs to be done missing in ep :o [13:48:33] johnduhart: LQT is working [13:48:34] ? [13:48:55] what about uw [13:49:15] ah, that is [13:57:37] the configuration is quite a hack - it has a java method that executes a bash script to call a php script [14:01:51] can I get a map to the maintenance folder? [14:07:27] hm... [14:07:33] map? [14:12:48] Hello all [14:13:11] Anyone awake? [14:17:31] make a NFS share to the mainenece folder [14:24:29] I added a problem report: Uncaught TypeError: Object function (selector,context){return new jQuery.fn.init(selector,context,rootjQuery);} has no method 'cookie' [14:24:44] I'm trying to code a new config class to woek with a common config [14:24:51] I'm trying to code a new config class to work with a common config [14:25:17] what instance is the common config on [14:25:20] ? [14:25:49] enwiki [14:27:59] Edokter: hm... [14:28:09] OrenBo: everywhere [14:28:17] OrenBo: /usr/local/apache/common [14:28:26] Edokter: thanks for report [14:39:19] to paraphrase - which deployment host has mediawiki installation + common.conf [14:39:45] I need to do some checks [14:40:04] deployment-web ? [14:43:42] what should I be looking for the maintence dir with eval.php [14:48:23] hi all, who can help Andre Engels to give him the rights to upload his public ssh key to https://labsconsole.wikimedia.org/wiki/Special:NovaKey [14:48:42] or is this something that only Ryan can help with/ [14:49:51] !logs created databases and accounts for coibot, xlinkbot, monitorbot on bots-sql3 for bots on bots-2 [14:51:40] drdee_: generally he should be able to do that [14:51:53] OrenBo: every instance has access to mw config files [14:51:58] I sent you path [14:52:04] OrenBo: do not use -ewb [14:52:07] web [14:52:17] petan: okay, nobody else can help right now? [14:52:28] drdee_: he can't login to console now? [14:52:39] he is logged in to labsconsole [14:52:45] but he cannot upload his ssh key [14:52:53] drdee_: what error he get? [14:52:59] !logout [14:53:02] hm... [14:53:06] @search log [14:53:06] Results (found 7): morebots, labs-morebots, credentials, logging, terminology, newgrp, initial-login, [14:53:13] !credentials [14:53:13] when you see No Nova credentials found for your account just relog to wiki and should be ok [14:53:50] 1 sec [14:54:17] hi andre [14:54:20] hi [14:54:27] !credentials [14:54:27] when you see No Nova credentials found for your account just relog to wiki and should be ok [14:54:41] try logging out and then back in [14:54:56] Thanks, that worked [14:54:57] OrenBo: petan|w: log of what? [14:55:26] But I really need to take a time-out now, sorry [14:55:35] petan: how high is this bug on the priority list, it has made me trip a few times already [14:55:44] !logs bots created databases and accounts for coibot, xlinkbot, monitorbot on bots-sql3 for bots on bots-2 [14:55:54] my mistake, I think [14:57:21] !mw-bot [14:58:46] petan|w: is there a way to search just in keywords not in the values? (for wm-bot) docs seem to say no [14:59:11] @whoami [14:59:11] You are trusted identified by name .*@wikimedia/.* [15:00:40] anyway, everything i did was logged. see https://labsconsole.wikimedia.org/w/index.php?title=Nova_Resource%3ASearch%2FSAL&action=historysubmit&diff=1269&oldid=1264 [15:00:47] jeremyb: no [15:00:54] jeremyb: you can't search only keys [15:01:00] k [15:01:25] jeremyb: there is a list of keys on web [15:01:35] drdee_: I don't know... [15:02:01] is there a space where we file labs bugs? [15:02:06] just bugzilla [15:02:07] ? [15:02:14] or RT? [15:02:14] there is a section for that [15:02:16] bz [15:02:23] tgx [15:02:28] yw [15:02:38] basically Ryan is a person who should fix it [15:02:52] check [15:07:06] !learn http is http://boingboing.net/2011/12/14/http-status-cats-by-girliemac.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+boingboing/iBag+(Boing+Boing)&utm_content=Google+Reader [15:07:29] !http is http://boingboing.net/2011/12/14/http-status-cats-by-girliemac.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+boingboing/iBag+(Boing+Boing)&utm_content=Google+Reader [15:07:30] You are not autorized to perform this, sorry [15:07:54] !http is http://boingboing.net/2011/12/14/http-status-cats-by-girliemac.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+boingboing/iBag+(Boing+Boing)&utm_content=Google+Reader [15:07:55] You are not autorized to perform this, sorry [15:08:01] :o [15:08:04] @trusted [15:08:04] I trust: petan!.*@wikimedia/Petrb (2admin), .*@wikimedia/.* (2trusted), .*@mediawiki/.* (2trusted), .*@wikimedia/Ryan-lane (2admin), .*@wikipedia/.* (2trusted), [15:08:04] :o [15:08:07] I thought it allowed meee [15:08:14] Ah not identified. [15:08:27] yay [15:08:29] :D [15:08:34] it should [15:08:35] @whoami [15:08:35] You are trusted identified by name .*@wikipedia/.* [15:08:38] :) [15:09:25] Also the en.wikipedia's caching servers are nazis. [15:09:32] why [15:11:12] Keep getting really confused, make edit -> go back to page gah no edit -> go to edit, hmm it's there -> wait a few min -> go back to page -> all there! The caching servers are stealing my content :( [15:11:33] petan|w: you may like to add *@nightshade.toolserver.org as well [15:11:40] to the list of trusted [15:12:03] Danny_B|backup: ask Damian [15:12:09] he can do it [15:12:23] @help trustadd Damianz ^ [15:12:23] Unknown command type @commands for a list of all commands I know [15:12:28] @help trustadd [15:12:28] Info for trustadd: Make an entry to the access list, example @trustadd regex admin [15:12:44] Lol [15:13:02] @trust add .*@nightshade.toolserver.org [15:13:02] petan|w: the list above says you and ryan are admins [15:13:07] @trust add regex .*@nightshade.toolserver.org [15:13:11] Yeah she ignores me [15:13:16] Damianz: no space [15:13:17] * Damianz refers back to petan [15:13:18] Damianz: todether! [15:13:21] yup [15:13:22] :P [15:13:27] she? [15:13:27] :D [15:13:30] heh [15:13:38] bottie [15:13:39] :D [15:13:41] All bots are a she ;) You tell them what to do then they ignore you [15:13:54] req: can I get admin on enwiki deployment? [15:14:00] Edokter: yes [15:14:10] Danny_B|backup: I replicated cs wiki [15:14:12] @trustadd regex .*@nightshade.toolserver.org [15:14:12] Unknown user level! [15:14:20] Ah that works better [15:14:31] petan|w: cool [15:14:33] Though @trust would make more sense imo [15:15:10] * Damianz gives wm-bot a cookie and goes back to figuring out how to move a 4year old nagios install without breaking it into hell. [15:15:40] Damianz: you did not add it ;-) [15:15:57] Edokter: got it [15:15:58] @trustadd .*@nightshade.toolserver.org trusted [15:16:04] Damianz: ^^ [15:16:08] OMG [15:16:08] :D [15:16:09] without space [15:16:13] ah [15:16:15] that was you [15:16:18] petan|w: i am not admin [15:16:23] I thought it was damian :D [15:16:24] petan|w: Thank you! [15:16:30] so i didn't want to trigger error msg [15:16:39] Danny_B|backup: you don't need to be admin trusted is enough [15:16:45] @whoami [15:16:45] You are unknown to me :) [15:16:47] nor trusted [15:16:58] I mean to add someone there [15:17:16] @whoami [15:17:16] You are unknown to me :) [15:17:18] see? [15:17:28] I thought it was damian who did it [15:17:39] didn't know it was you giving him a tip how to do that :D [15:17:56] It would have been me but wm-bot doesn't think I have access to do that [15:18:01] I can just do alias stuff apparently [15:18:03] you do [15:18:10] Damianz: @trustadd regex trusted [15:18:15] @trustadd regex .*@nightshade.toolserver.org [15:18:16] Unknown user level! [15:18:16] try it [15:18:19] LEVEL [15:18:19] ^ [15:18:20] :D [15:18:34] @trustadd regex trusted .*@nightshade.toolserver.org [15:18:34] Successfuly added regex [15:18:35] ahh [15:18:38] Fail [15:18:39] omg [15:18:40] :D [15:18:41] Your help really sucks [15:18:42] :P [15:18:47] LOL [15:18:54] I should read the docs though I guess... [15:19:13] @trusted [15:19:13] I trust: petan!.*@wikimedia/Petrb (2admin), .*@wikimedia/.* (2trusted), .*@mediawiki/.* (2trusted), .*@wikimedia/Ryan-lane (2admin), .*@wikipedia/.* (2trusted), regex (2trusted), [15:19:30] @trustdel regex [15:19:30] User was deleted from access list [15:19:39] @trustadd trusted .*@nightshade.toolserver.org [15:19:40] Unknown user level! [15:19:50] @trustadd .*@nightshade.toolserver.org trusted [15:19:50] Successfuly added .*@nightshade.toolserver.org [15:19:52] @trustdel regex [15:19:53] User not found, sorry [15:19:57] @trusted [15:19:57] I trust: petan!.*@wikimedia/Petrb (2admin), .*@wikimedia/.* (2trusted), .*@mediawiki/.* (2trusted), .*@wikimedia/Ryan-lane (2admin), .*@wikipedia/.* (2trusted), .*@nightshade.toolserver.org (2trusted), [15:20:00] fml [15:20:03] Try now! [15:20:55] heh, may i test too:) [15:20:57] @trustadd *@wikimedia/Mutante admin [15:20:57] Permission denied [15:21:01] ok:) [15:21:17] * johnduhart is here [15:21:20] hey [15:21:29] petan|w: Where's OrenBo on search? [15:21:34] I have no clue [15:21:58] Danny_B|backup: you should be on list but it was really not easy :D [15:22:15] we almost blown up the bot [15:22:21] :-P [15:22:27] it's all logged too heh [15:22:28] :D [15:22:48] Damianz: seriously I was close to posting to bash.org [15:22:50] it's not only me, there is a bunch of users connected via ts [15:22:51] :D [15:23:02] I know but they don't use -labs :P [15:23:10] yet [15:23:52] johnduhart: researching coding for a new configuration class [15:24:40] OrenBo: What's wrong with what's in production right now? [15:25:17] johnduhart: : it does not work [15:25:44] OrenBo: How can that be? Search worked on wikimedia last time I checked [15:25:45] it only does single configuration setup [15:27:32] Q: Does anyone know if wikidiff2.cpp has been recompiled on the labs version? [15:27:47] Edokter: it's possible it wasn't :o [15:28:06] OrenBo: did you consult it with notpeter? [15:28:07] Edokter: No, I'm using what's in our apt repo. Same as in production [15:28:10] That explains the old the old issues :) [15:28:37] Edokter: Have you consulted with engineering about updating the package? [15:28:42] I have various fixes patched, but they don't show [15:28:49] johnduhart: it would make sense to test it here [15:28:52] before packaging it [15:28:59] it's deployment test [15:29:00] petan|w: Not right now [15:29:05] okokokok [15:29:06] nope [15:29:28] but I'm pretty sure the configuration was made by hand [15:29:31] OrenBo: I think he could know it? [15:29:42] OrenBo: yes in that case he would know it [15:29:50] sure would [15:30:00] but I don't know how to do it [15:31:05] :P [15:31:08] do you want to have to make a new configuration every time ? [15:31:25] I don't know how config for search look [15:31:30] where can I find engineering? [15:31:38] Edokter: huh? [15:31:43] Next to hair and beauty. [15:31:44] Edokter: File a bug report [15:31:46] OrenBo: isn't it on noc [15:31:47] I'm trying to work with common.conf [15:32:18] yep it is [15:32:26] that version doesn't work? [15:32:28] let me see it [15:32:31] but that one is no good for us [15:32:47] OrenBo: why [15:33:38] good question - like Ryan_Peter says copy paste don't always work [15:34:11] there are some diffs but otherwise it looks useable [15:34:23] some paths need to be updated only [15:36:23] ok OrenBo what you don't know how to set up? [15:36:34] maybe I would help you [15:36:39] let me open it [15:36:42] ok [15:36:44] ep [15:46:00] ok see line #3 {file:///home/wikipedia/common/pmtpa.dblist} : (single,true,20,1000) (prefix) (spell,10,3) [15:46:18] we need that file [15:46:25] all.dblist [15:46:34] try /usr/local/apache/common/all* [15:46:39] that's it [15:46:42] all.dblist [15:46:55] OrenBo: it's on search instance too [15:47:06] i'm skeptical [15:47:13] huh [15:47:31] you see some of the databes should not be in it [15:47:46] where you see that [15:47:49] :o [15:48:01] the next 7 lines [15:48:40] line 3 says how to index those dbs [15:48:51] line 3 where [15:48:54] single index [15:49:08] maybe you should create a config on fs that we can update [15:49:09] generate spell checking [15:49:21] there is a template [16:06:14] re [16:06:23] my name was said. what's up? [16:07:02] how is http://noc.wikimedia.org/conf/lsearch-global-2.1.conf generated in production ? [16:07:17] same for pmtpa.dblist [16:07:45] statically [16:07:52] it lives on the bastion host [16:08:00] and then gets pushed out via various sync scripts [16:08:13] although, in the case of search, it's read on search boxes via nfs [16:08:18] petan|w: are the database names same as in production [16:08:20] notpeter: And pmtpa is the same as all.dblist though [16:08:21] but yeah, it's just deited by hand [16:08:23] OrenBo: Yes. [16:08:39] ok [16:08:46] johnduhart: yes. we should just switch to all.dblist [16:08:51] although [16:09:02] the format is not fully described in the docs [16:09:10] I'm also in a "if it ain't broke, don't fix it" mindset [16:09:21] heh [16:09:42] we'll in labs we take a more gang ho attitude [16:09:56] definitely :) [16:10:01] OrenBo: anything else you need? [16:10:02] we like to crash test everything [16:10:36] just to confirm the database names are the same as in production [16:10:43] yes [16:12:07] ok let's try to replace mwsearch global config with noc [16:12:32] ok [16:13:45] btw johnduhart if Ryan showed up here I think we should consider moving sql to new instance once Oren start indexers load of sql server could be really high and slow down cluster even more [16:14:17] I mean, we can do it now or never :D [16:14:35] Talk to ryan about that. [16:14:43] I know, just wanted to tell you [16:15:53] Sure [16:18:10] 01/13/2012 - 16:18:09 - Updating keys for a_engels [16:18:13] 01/13/2012 - 16:18:13 - Updating keys for a_engels [16:22:09] 01/13/2012 - 16:22:09 - Updating keys for a_engels [16:22:13] 01/13/2012 - 16:22:12 - Updating keys for a_engels [16:23:05] also I think this global config will require 3 search instances [16:23:56] or 4 [16:24:11] Do we really need 4 instances? [16:24:19] nope [16:24:22] just one [16:24:37] I just don't know how to change the congiguration [16:24:41] yet [16:25:21] also OAI section looks no good [16:26:37] will deploymet support dns names for each wiki? [16:26:58] OrenBo: What do you mean? [16:27:38] look at the lsearch-global.conf [16:27:46] OAI section [16:28:07] Okay [16:28:12] those entries need to point at the deploinment hostnames [16:28:26] OrenBo: Okay, change them [16:28:35] to what ? [16:28:47] To the current hostnames [16:28:59] en.wikipedia.deployment.wmflabs.org/wiki/Main_Page [16:30:00] that one is not in the list... [16:31:32] what is simple english [16:34:03] simple.wikipedia.deplo.... [16:36:05] are you doing a full import an any wikis? [16:36:11] simple [16:36:21] i mean all pages [16:36:25] yes [16:36:26] simple has [16:36:30] full db [16:36:33] what about othes [16:36:38] only mw space [16:40:23] can someone install cite extention on labconsole ? [16:40:40] OrenBo: um, what for? [16:40:58] ask Ryan... [16:41:18] OrenBo: it is better to use mediawiki.org for documentation [16:41:23] of labs [16:41:37] labsconsole should be only for docs of interface of wiki [16:41:53] ok [16:42:11] problem is I can't focus [16:42:44] um? [16:43:01] I'm working with 20 web pages and can't summrise anything [16:43:09] ah [16:43:41] petan|w: Uh, really? [16:43:53] johnduhart: what? [16:44:00] labsconsole should be only for docs of interface of wiki [16:44:14] I got the source, docs on lucene_search, irc logs, instance names etherpad ... [16:44:27] Ryan told me that docs and proposals should go to mediawiki [16:44:30] seems like Cite would be a reasonable ext to install on labsconsole [16:44:38] footnotes need it, right? [16:44:48] I guess you are right [16:44:48] docs need footnotes [16:44:49] yep [16:45:10] johnduhart: actually I don't know... [16:45:15] I'm thinking to document you instance you need to use the labs wiki [16:45:34] to document how labs works you need to use medaiwiki [16:45:47] I think Ryan was like, labsconsole is restricted so it's not good for general documentation, that's why I put all to mw:Wikimedia Labs [16:46:21] but have a look at https://labsconsole.wikimedia.org/wiki/User:Oren [16:46:32] OrenBo: to document how labs work use labsconsole [16:46:34] do you think that should go in mw [16:47:20] I think it should because it describe something people without access to labs are affiliated with [16:47:35] it's documentation of mwsearch setup or not? [16:47:44] just on labs [16:47:48] ah [16:47:54] wait for Ryan [16:47:56] I don't know [16:48:07] anything in user space is probably ok [16:50:47] what's the path to rsync [16:52:11] did you install rsync? [16:54:01] to see the path to the rsync binary, type: which rsync [16:54:31] thanks [16:55:35] sudo apt-get install rsync [16:55:37] OrenBo: ^ [16:55:40] I need to go [16:55:44] ok [16:55:52] anything else, use EP, ok? [16:55:54] I've updated local congif [16:55:54] I will install it [16:55:59] which one [16:56:09] lsearch.conf [16:56:12] ok [16:56:21] and some of the global [16:56:36] I'm missing one setting [16:56:40] at least [17:00:12] do you think we could try to push indexes using murder rather than rsync ? [17:02:11] need to make a security rule for port 8321 [17:03:58] ok I need nfs share of a common/php/languages/messages [17:04:03] or a copy from svn [17:04:10] and a path to it [17:04:56] and the password for the OAI.password for user lsearch2 ;-) [17:05:22] that should wrap up the local settings [17:22:38] OrenBo: Write that stuff down on the pad please [17:22:51] already done [17:24:05] * Beetstra breaks his head .. why does bot1 have access to a database, and bot2 (using the same account name and password) not .. [17:27:48] Beetstra: MAX_USER_CONNECTIONS ? [17:28:08] set to 0 .. [17:28:12] which would be unlimited [17:28:25] hmm.. [17:28:38] and bot1 has already 3-4 connections into his database, but bot2 can't .. [17:28:49] maybe they should use different accounts anyways? independent from this [17:29:18] * Beetstra tries [17:29:20] and just to test [17:30:44] MAX_CONNECTIONS_PER_HOUR ? [17:34:06] They are all set to 0 .. [17:36:04] when you say you use the same account.. does it mean there is no IP address in the GRANT? [17:37:18] not sure what you mean [17:37:36] bot1 logs into its own database with 'bot1' and password, stores data [17:37:54] show grants for 'user'@'host'; [17:38:12] is this all on the same host? bot1, bot2 and the db ? [17:38:47] bot1 and bot2 are running on 'bots-2' [17:39:05] the db are on bots-sql3 [17:39:36] so you allowed 'someuser'@'localhost' ? ok, then nevermind [17:39:56] i thought bot1 and bot2 might be connecting from different places [17:40:09] no [17:40:19] arg, not localhost, got it [17:40:33] but they are connecting _from_ the same host [17:41:21] yes [17:41:49] weird. the code they use to connect is identical? [17:42:37] you get an actual permission denied? [17:43:09] no, one bot does not get any data out of the tables, and does not write to tables [17:43:38] I changed it now, so bot1 logs into both database as bot1, and bot2 into both databases as bot2 .. [17:43:49] And have given them the correct rights to do so [17:53:48] and it works ? nice [17:59:49] Nyan! [18:01:03] Better 40 days late than never. :-) (I need to check my mw.org talkpage more often) :-) [18:13:44] Hm. https://labsconsole.wikimedia.org/wiki/Access has everything /except/ info on how to get an account in the first place. :-) [18:15:08] !Ryan [18:15:08] man of the all answers ever [18:15:10] Coren: ^ [18:15:14] here we go [18:15:25] Coren: you need to ask him [18:15:34] !search access [18:15:46] @regsearch \. [18:15:46] Results (found 60): puppet, instance, morebots, git, nagios, bot, labs-home-wm, labs-nagios-wm, labs-morebots, gerrit-wm, wiki, labs, extension, wm-bot, projects, putty, gerrit, change, wikitech, revision, monitor, alert, unicorn, help, bz, os-change, instancelist, instance-json, amend, queue, socks-proxy, sal, info, security, ask, sudo, access, $realm, keys, $site, bug, pageant, blueprint-dns, bots, rt, pxe, group, pathconflict, terminology, etherpad, nova-resource, pastebin, newgrp, osm-bug, bastion, initial-login, account-questions, SAL, manage-projects, rights, [18:15:54] @regsearch 2\. [18:15:54] Results (found 1): account-questions, [18:16:00] !account-questions [18:16:00] 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. [18:16:04] mutante .. not sure if it works, but found a more pressing bug to solve first (ut8-character encoding problem) .. I moved the bots back to the old box for now .. don't have time [18:17:11] * Coren nodsnods. [18:17:29] Coren: do you already know the answers to the account questions? [18:19:16] mutante: They're not so hard that I'd need time to think about it, though I don't see a list of usernames in use. [18:19:36] But yeah, "Coren", "marc" and "marc@uberbox.org" respectively. :-) [18:19:57] Special:ListUsers [18:20:22] That's for the wiki usernames. :-) [18:21:23] do you want "real name" (optional)? [18:21:56] No reason not to; it's not like it's a secret. "Marc A. Pelletier" [18:25:27] Coren: A randomly generated password for Coren has been sent to marc@uberbox.org. [18:27:22] Coren: should work (at least the wiki login) [18:27:54] johnduhart: why you removed button from banner [18:30:34] lo [18:30:40] I have been exhausted [18:30:44] hy [18:30:49] catching up [18:30:51] hi [18:31:13] petan: so did you guys get uw working? [18:31:19] * hexmode goes to check [18:31:21] it seems so [18:32:36] \o/ ... just tried and no more endless spinner [18:33:45] pt wikibooks has lqt... checking [18:34:22] another reason to have all prod wikis [18:35:28] search isn't happening, though, right? OrenBo? [18:36:03] http://labs.wikimedia.deployment.wmflabs.org/w/index.php?title=Problem_reports&action=edit§ion=8 [18:36:11] without edit [18:36:12] :D [18:40:22] Suddenly, I am cool! [18:40:26] ? [18:42:29] petan: how were you importing dumps? [18:42:43] Coren: so works for you ?:) [18:42:47] hexmode: mwimport [18:43:30] petan: where is mwimport? [18:43:34] 'bout to login to bastion to test. [18:43:36] svn [18:43:42] and where were you running it from? [18:43:46] mediawiki.org/wiki/wmimport [18:43:50] dbdump [18:43:57] mediawiki.org/wiki/MWImport [18:44:45] -test, -web, -nfs-memc ? [18:45:04] Ah, hm, I'm not in the Bastion project. [18:46:21] https://www.mediawiki.org/wiki/Mwdumper [18:46:54] petan: you have that already on the cluster? [18:46:55] mutante: Yep. Wiki login works. [18:47:14] petan: can you run an import for me? [18:48:27] Coren: cool [18:49:44] yes [18:51:12] hexmode: hm? [18:51:19] tell me which wiki [18:52:21] petan: ptwikibooks ... dump is in my home dir [18:52:36] doing [18:53:10] hexmode: ssh deployment-dbdump [18:53:18] I need to automount it [19:02:31] hexmode: will be done in hour or so [19:02:52] don't logout :D [19:03:21] :) [19:03:33] actually I think it would be done earlier [19:03:44] 50 000 revisions so far [19:03:58] * Coren grumbles at IPv6 routing. [19:04:03] k, I'm gonna chance down Helder.wiki and see if he can test wikibooks on deploy [19:04:17] if Ryan gave us a better ram it would be a matter of few minutes [19:04:24] right [19:05:13] So, Ryan's the one to ask for project access, then? [19:05:34] no [19:05:35] PROBLEM Free ram is now: WARNING on deployment-sql deployment-sql output: Warning: 19% free memory [19:05:44] Coren: which project [19:06:12] petan: Well, Bastion for one. Also bots I expect since my objective is to move CSBot there. :-) [19:06:47] right [19:06:55] mutante: give him access to bastion :) [19:07:18] done [19:07:19] bots [19:07:29] Coren: you are in bots [19:07:34] ask me if you needed any help [19:07:43] !log bots added coren to project [19:07:44] Logged the message, Master [19:07:46] That was fast. Thanks. :-) [19:08:07] 01/13/2012 - 19:08:07 - Creating a home directory for marc at /export/home/bots/marc [19:09:07] 01/13/2012 - 19:09:07 - Updating keys for marc [19:09:46] !log bots added Beetstra to project [19:09:48] Logged the message, Master [19:12:34] !log bots created new vm for bots [19:12:35] Logged the message, Master [19:12:40] Coren: let me set it up [19:12:48] I think you should move your bot to that one [19:13:02] lucid ubuntu 2gb ram, ok? [19:13:11] 01/13/2012 - 19:13:09 - Creating a home directory for marc at /export/home/bastion/marc [19:13:20] mutante: :) [19:13:21] Coren: i just added you to Bastion [19:13:38] !ssh [19:13:51] !access | Coren [19:13:51] Coren: https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances [19:13:53] petan: ooh.. did something happen twice? [19:14:09] 01/13/2012 - 19:14:09 - Updating keys for marc [19:14:19] mutante: ] 01/13/2012 - 19:13:09 - Creating a home directory for marc at /export/home/bastion/marc [19:14:26] few sec before you :P [19:14:32] bot told us [19:14:52] so when you added to bots project, [19:15:06] it automatically added a home dir on bastion already? [19:15:13] no it didn't, it was you, I just responded to bot [19:15:14] oh, no got it:) [19:15:19] then you posted message to chan [19:15:27] i was confused for a minute..nevermind [19:15:47] heh [19:15:47] RECOVERY Free ram is now: OK on deployment-sql deployment-sql output: OK: 20% free memory [19:16:29] !log bastion added new member Coren [19:16:30] Logged the message, Master [19:17:05] alright, i am in Euro timezone, so out for weekend now for real:) [19:17:09] hexmode: done [19:17:14] mutante: same [19:17:16] :) [19:17:26] petan: :) [19:17:38] but now it needs to convert db to 19 [19:17:42] working on that [19:17:46] ok:) cu guys [19:17:50] bb [19:27:31] !log bots Bots writing to Wiki disabled due to utf8 problems in perl [19:27:31] Logged the message, Master [19:31:04] hi Ryan_Lane [19:31:18] howdy [19:31:19] we added coren to labs (mu tante did) [19:31:25] cool [19:31:25] for bots [19:31:29] sounds goo [19:31:32] *good [19:31:37] migration kind of sucked last night [19:31:42] heh I've seen [19:31:59] wasn't too happy to be up till 1:30am [19:32:04] Beetstra: i'm also working on utf8 in perl issues right now (this very minute) [19:32:11] that's usual time for me [19:32:12] :D [19:32:22] most of the problems were my fault. one was a bug in nova, another was a bug in php aws-sdk [19:32:47] I am trying to upgrade the installed perl modules within the self-written module .. it is not a problem on the box where I am from with the same code [19:32:54] But see the last handful of edits of XLinkBot on en.wikipedia [19:33:24] many of the modules are not up to date .... [19:33:29] heh. well, I had to wake up at 8, so I wasn't too amused to have to stay up that late :) [19:33:37] ah, right [19:33:37] anyway, it's done now [19:33:42] just shut down virt1 [19:33:59] I'm going to make sure it keeps working, then I'm going to move it to the compute cluster [19:34:00] right, when you expect sql server to arrive :o [19:34:15] or gluster [19:34:18] :P [19:34:19] well, we have hardware in for sql right now [19:34:24] cool [19:34:32] but have basically no time to set it up [19:34:37] hm, ok [19:34:42] we would use puppet to manage it? [19:34:53] maybe I could write some config over weekend [19:35:16] althought I am unsure how it's going to work, didn't try to configure mariadb using puppet yet [19:36:52] PROBLEM host: nova-dev2 is DOWN address: nova-dev2 check_ping: Invalid hostname/address - nova-dev2 [19:37:01] jeremyb, on which botserver are you? [19:37:03] PROBLEM host: labs-ocg1 is DOWN address: labs-ocg1 check_ping: Invalid hostname/address - labs-ocg1 [19:37:04] PROBLEM host: labs-cp2 is DOWN address: labs-cp2 check_ping: Invalid hostname/address - labs-cp2 [19:37:09] Beetstra: i'm on linode atm [19:37:18] Beetstra: seen perl -C ? [19:37:38] no [19:38:10] what does that do? [19:38:26] perldoc perlrun [19:38:34] i've been using perl -CSDA [19:38:37] doesn't seem to help [19:38:38] Ryan_Lane: is everything ok? [19:38:46] * Coren grumble, grumble [19:38:47] I can't connect to labs from outside [19:38:59] there is a newer perl version installed on mediawiki-labs comparing to what I had [19:39:06] PROBLEM host: canonical-bridge is DOWN address: canonical-bridge check_ping: Invalid hostname/address - canonical-bridge [19:39:13] Ryan_Lane: is the sql something that notbinasher can set up? [19:39:17] PROBLEM host: labs-mc1 is DOWN address: labs-mc1 check_ping: Invalid hostname/address - labs-mc1 [19:39:27] PROBLEM host: master is DOWN address: master check_ping: Invalid hostname/address - master [19:39:27] PROBLEM host: labs-mw2 is DOWN address: labs-mw2 check_ping: Invalid hostname/address - labs-mw2 [19:39:29] YAY [19:39:43] PROBLEM host: labs-mc2 is DOWN address: labs-mc2 check_ping: Invalid hostname/address - labs-mc2 [19:39:44] PROBLEM host: labs-mw1 is DOWN address: labs-mw1 check_ping: Invalid hostname/address - labs-mw1 [19:39:49] Ryan_Lane: what's going on [19:39:56] DNS must have borke [19:40:00] broke* [19:40:18] !nagios [19:40:19] http://nagios.wmflabs.org/nagios3 [19:40:19] it's not just dns [19:40:19] it's everything [19:40:26] !ping [19:40:28] hey :) is there anything broken with labs? I can not connect on jenkins and IIRC it uses labs LDAP as an authentication backeend [19:40:31] @help [19:40:31] Type @commands for list of commands. This bot is running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.1.4 source code licensed under GPL and located in wikimedia svn [19:40:33] bah [19:40:43] PROBLEM host: labs-lvs1 is DOWN address: labs-lvs1 check_ping: Invalid hostname/address - labs-lvs1 [19:40:44] PROBLEM host: pad1 is DOWN address: pad1 check_ping: Invalid hostname/address - pad1 [19:41:01] Ryan_Lane: 15 vm's crashed [19:41:17] PROBLEM host: bastion1 is DOWN address: bastion1 check_ping: Invalid hostname/address - bastion1 [19:41:17] PROBLEM host: ganglia-master is DOWN address: ganglia-master check_ping: Invalid hostname/address - ganglia-master [19:41:17] guess i have my answer :-) [19:41:18] this is why I turned off virt1, to test it [19:42:02] gimme a sec [19:42:05] it's likely due to dns [19:42:10] ok [19:42:15] the instances themselves are probably find [19:42:18] *fine [19:42:34] PROBLEM host: labs-cp1 is DOWN address: labs-cp1 check_ping: Invalid hostname/address - labs-cp1 [19:43:40] dns is *supposed* to have two damn dns servers listed [19:43:42] this is annoying [19:43:53] PROBLEM host: wep is DOWN address: wep check_ping: Invalid hostname/address - wep [19:44:14] hm [19:44:18] jeremyb - perl and utf8 wikipages have always been annoying, but why does this install behave different from an older install [19:44:21] strangly enough, dns works [19:44:27] jeremyb, let me know if you find anything, it is bedtime here [19:44:28] why the hell would virt1 going down kill instances? [19:45:03] DNS don't work here [19:45:06] petan: how much longer will this be offline? [19:45:11] hexmode: no idea [19:45:13] Ryan knows [19:45:15] I am getting NXDOMAIN sadness on bastion. [19:45:17] ah [19:45:21] Coren: wait [19:45:23] outage [19:45:23] nxdomain? seriously? [19:45:27] So if DNS is up, it's not visible from outside. :-) [19:45:34] RECOVERY host: nova-dev2 is UP address: nova-dev2 PING OK - Packet loss = 0%, RTA = 0.58 ms [19:45:34] Ryan_Lane: same [19:45:35] definitely shouldn't be getting nxdomain [19:45:46] ok, virt1 is back up [19:45:51] still getting nxdomain? [19:45:53] RECOVERY host: labs-mc1 is UP address: labs-mc1 PING OK - Packet loss = 0%, RTA = 0.32 ms [19:45:54] RECOVERY host: labs-lvs1 is UP address: labs-lvs1 PING OK - Packet loss = 0%, RTA = 7.63 ms [19:45:54] RECOVERY host: pad1 is UP address: pad1 PING OK - Packet loss = 0%, RTA = 7.48 ms [19:45:59] hexmode: up [19:46:01] It's back now [19:46:03] RECOVERY host: labs-mw1 is UP address: labs-mw1 PING OK - Packet loss = 0%, RTA = 0.49 ms [19:46:03] RECOVERY host: master is UP address: master PING OK - Packet loss = 0%, RTA = 0.62 ms [19:46:04] RECOVERY host: bastion1 is UP address: bastion1 PING OK - Packet loss = 0%, RTA = 3.01 ms [19:46:04] RECOVERY host: ganglia-master is UP address: ganglia-master PING OK - Packet loss = 0%, RTA = 1.52 ms [19:46:06] :) [19:46:08] Maybe negative cache [19:46:08] fucking DNS [19:46:08] 01/13/2012 - 19:46:07 - Updating keys for marc [19:46:10] 01/13/2012 - 19:46:09 - Updating keys for marc [19:46:34] bastion likes me now [19:46:38] works now [19:46:44] RECOVERY host: labs-cp1 is UP address: labs-cp1 PING OK - Packet loss = 0%, RTA = 0.63 ms [19:46:44] RECOVERY host: labs-cp2 is UP address: labs-cp2 PING OK - Packet loss = 0%, RTA = 0.53 ms [19:46:53] RECOVERY host: wep is UP address: wep PING OK - Packet loss = 0%, RTA = 0.45 ms [19:46:54] RECOVERY host: canonical-bridge is UP address: canonical-bridge PING OK - Packet loss = 0%, RTA = 0.44 ms [19:47:13] RECOVERY host: labs-ocg1 is UP address: labs-ocg1 PING OK - Packet loss = 0%, RTA = 0.39 ms [19:47:26] Ooo. Lucid goodness. [19:47:28] well, that sucks [19:47:33] RECOVERY host: labs-mc2 is UP address: labs-mc2 PING OK - Packet loss = 0%, RTA = 2.60 ms [19:47:45] I guess I need to get it changed upstream to virt0 [19:47:50] I have a fucking secondary [19:47:54] * Ryan_Lane grumbles [19:48:03] RECOVERY host: labs-mw2 is UP address: labs-mw2 PING OK - Packet loss = 0%, RTA = 0.42 ms [19:48:42] hexmode: was it full dump? [19:48:45] wait. did I set it properly in our primary dns? [19:48:49] Ryan_Lane: want me to slave for your domains? [19:49:07] nah. we're setting up another slave in eqiad [19:49:08] soon [19:49:10] petan: I think so, but it doesn't look like it [19:49:15] yes I see [19:49:34] will check [19:50:02] kk. I have colo infrastructure in Ohio with darn near four sigmas of reliability if you ever need. [19:50:03] http://pt.wikibooks.deployment.wmflabs.org/wiki/Especial:Estat%C3%ADsticas [19:51:15] hexmode: checking [19:51:33] PROBLEM Disk Space is now: CRITICAL on cantlogin cantlogin output: DISK CRITICAL - free space: / 0 MB (0% inode=19%): [19:51:36] what's a sigma? is that like a 9? :) [19:51:50] well, we can't really host DNS outside of us :) [19:51:57] also, it's reliant on LDAP [19:52:17] the mode of powerdns we are using doesn't allow zone transfers [19:52:31] would just zone xfer... oh, nevermind. [19:52:35] :-) [19:52:43] will there be an opendj slave/replica/whatever you call em in eqiad? [19:57:49] jeremyb: standard deviation. That'd translate to 99.38% uptime for a continuous service. [20:03:29] Ryan_Lane: Powerdns can replicate using mysql [20:03:38] yes, but we aren't using mysql [20:03:40] we're using ldap [20:03:46] it can replicate using ldap too [20:04:30] and we can do fractional replication, but there's also information on those entries we shouldn't be exporting [20:04:35] to third parties [20:20:10] shut [20:20:16] *shit [20:20:16] I see the problem now [20:20:25] the damn recursor was set to the wrong server [20:20:31] for wmflabs [20:21:49] ah. crap. the instances are also still set to hit vir1 [20:21:51] virt1 [20:22:05] problems all around! [20:24:02] !account-questions | ezachte [20:24:02] ezachte: 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. [20:26:47] New patchset: Ryan Lane; "Making virt0 the new controller. Moving all nova config to point to it." [operations/puppet] (test) - https://gerrit.wikimedia.org/r/1912 [20:30:03] New review: Ryan Lane; "(no comment)" [operations/puppet] (test); V: 0 C: 2; - https://gerrit.wikimedia.org/r/1912 [20:30:04] Change merged: Ryan Lane; [operations/puppet] (test) - https://gerrit.wikimedia.org/r/1912 [20:34:23] way too many freaking moving parts [20:39:11] ezachte on irc, that's a first ;-) [20:39:59] ok. recursor is changed [20:40:04] now to try to kill the pdns server on virt1 [20:40:34] outage imminent? [20:41:43] seems to still be working [20:42:59] hmm. virt2 is still hitting 131 [20:46:33] PROBLEM Disk Space is now: UNKNOWN on cantlogin cantlogin output: Invalid host name cantlogin [20:49:53] PROBLEM host: labs-mc1 is DOWN address: labs-mc1 check_ping: Invalid hostname/address - labs-mc1 [20:50:13] PROBLEM host: master is DOWN address: master check_ping: Invalid hostname/address - master [20:50:14] PROBLEM host: nova-dev2 is DOWN address: nova-dev2 check_ping: Invalid hostname/address - nova-dev2 [20:50:23] PROBLEM host: labs-mw1 is DOWN address: labs-mw1 check_ping: Invalid hostname/address - labs-mw1 [20:50:53] PROBLEM host: pad1 is DOWN address: pad1 check_ping: Invalid hostname/address - pad1 [20:50:55] * Ryan_Lane sighs [20:51:19] puppet isn't running properly on recursor1 [20:51:33] RECOVERY host: nova-dev2 is UP address: nova-dev2 PING OK - Packet loss = 0%, RTA = 1.06 ms [20:51:34] crap [20:51:43] RECOVERY host: master is UP address: master PING OK - Packet loss = 0%, RTA = 0.60 ms [20:51:54] RECOVERY host: labs-mc1 is UP address: labs-mc1 PING OK - Packet loss = 0%, RTA = 0.90 ms [20:52:03] RECOVERY host: labs-mw1 is UP address: labs-mw1 PING OK - Packet loss = 0%, RTA = 0.44 ms [20:52:23] RECOVERY host: pad1 is UP address: pad1 PING OK - Packet loss = 0%, RTA = 0.67 ms [20:52:33] there we go. [20:57:19] PROBLEM Disk Space is now: WARNING on puppet-lucid puppet-lucid output: DISK WARNING - free space: / 40 MB (3% inode=35%): [20:58:38] PROBLEM dpkg-check is now: CRITICAL on puppet-lucid puppet-lucid output: DPKG CRITICAL dpkg reports broken packages [21:17:47] testing taking down virt1's pdns process again [21:22:18] PROBLEM host: asher1 is DOWN address: asher1 check_ping: Invalid hostname/address - asher1 [21:23:58] PROBLEM host: bots-3 is DOWN address: bots-3 check_ping: Invalid hostname/address - bots-3 [21:24:18] Ryan_Lane: again [21:24:18] PROBLEM host: mobile-enwp is DOWN address: mobile-enwp check_ping: Invalid hostname/address - mobile-enwp [21:24:43] RAWR [21:24:52] coming back [21:24:57] petan: can we redirect deployment.wmflab to beta.wmflabs ? [21:25:06] hexmode: yes [21:25:09] RECOVERY host: bots-3 is UP address: bots-3 PING OK - Packet loss = 0%, RTA = 0.62 ms [21:25:15] robla just pointed out deployment is overloaded [21:25:18] and confusing [21:25:19] hexmode: huh? [21:25:22] overloaded? [21:25:42] Ryan_Lane: can you rename the dns? [21:25:42] petan: meaning isn't clear [21:25:49] some of the performance problems may be due to the stupid dns stuff I'm causing [21:26:10] funny enough, only some of the instances had problems this time [21:26:22] seems I may just need to wait another hour or two [21:26:23] Ryan_Lane: can we keep deployment as a cname and make the primary one beta? [21:26:35] I don't know what you mean [21:26:43] we don't use cnames. they are all A records [21:26:57] hexmode: wouldn't work [21:27:06] if we rename it to beta it will be beta [21:27:08] RECOVERY host: mobile-enwp is UP address: mobile-enwp PING OK - Packet loss = 0%, RTA = 0.56 ms [21:27:11] I want to have deployment redirected to beta [21:27:16] why? [21:27:17] that's possible too [21:27:22] np [21:27:28] RECOVERY host: asher1 is UP address: asher1 PING OK - Packet loss = 0%, RTA = 1.57 ms [21:27:31] Because the name is out there already [21:27:47] I would need to use 2 hostnames is it ok Ryan? [21:27:57] *.*.deployment and *.*.beta [21:28:00] what are you guys trying to do? [21:28:04] I have no ide [21:28:06] * idea [21:28:07] well, it's just *.beta [21:28:11] and *.deployment [21:28:15] hexmode's request [21:28:30] this all is his idea I am just making it work :) [21:28:40] hexmode: what are you trying to do? [21:28:42] Ryan_Lane: I'm trying to make the main name beta and phase out deployment [21:28:42] he takes the responsiblity hehe [21:28:47] oh [21:29:00] in DNS [21:29:01] gimme a sec, then [21:29:07] Ryan_Lane: can you keep both [21:29:08] what is in labs doesn't matter [21:29:11] so I can redirect it [21:29:14] yeah, I'm gonna keep both [21:29:27] well, if you want the old address to work, it needs to stay in dns [21:29:32] just to be clear [21:29:33] PROBLEM Free ram is now: WARNING on deployment-sql deployment-sql output: Warning: 18% free memory [21:29:37] ok [21:29:44] btw Ryan_Lane we are out of ram on sql :o [21:29:50] heh [21:29:51] but I think it can wait [21:29:57] deployment.wmflabs.org redirects to beta.wmflabs.org ? [21:29:58] :) [21:30:04] hexmode: np [21:30:07] k [21:31:23] done [21:33:21] done [21:33:22] :D [21:33:34] !log deployment-prep moved deployment to beta.wmf... [21:33:36] Logged the message, Master [21:34:30] !log deployment-prep assigning new dns [21:34:31] Logged the message, Master [21:34:33] RECOVERY Free ram is now: OK on deployment-sql deployment-sql output: OK: 20% free memory [21:34:46] ah you did it [21:35:02] oh lol [21:35:57] petan: ugh: "unknown mysql server host beta-sql" [21:35:58] works [21:36:03] fixed [21:36:03] k [21:36:16] just cookies are cleared now :D [21:36:21] Ryan_Lane: petan tyvm [21:36:27] yw [21:38:08] ?? [21:38:11] fixing [21:38:42] k, so search. I guess I missed OrenBo earlier so I don't know what is going on there [21:39:00] but I'm glad uploadwizard is working [21:39:05] done [21:39:10] :o [21:39:37] petan: do you know anything about search? [21:40:04] little [21:40:24] we can either use prod version or switch to mw classic slow for now [21:40:37] I can't set up whole search system alone [21:40:46] I would need help of someone who understand it better [21:40:57] I could switch to slow one [21:41:18] production is using indexer, mw is using sql search [21:41:57] hexmode: everything is fine, pt is broken will fix in 2 hours ok? [21:42:02] kk [21:42:06] need to enjoy friday :> [21:42:45] OrenBo go help hex :P [21:42:47] ok. I just changed virt1 to be a cname of vir0 [21:42:50] he needs search heh [21:42:53] virt0, that is [21:43:06] so, until the NS records update, I should be able to bring pdns down on virt1 [21:43:10] in about an hour [21:43:10] yay [21:43:19] since the ttl was set to an hour [21:43:30] now its 60 seconds [21:43:38] incase I need to quickly change it back in an hour :D [21:47:24] heh. crap [21:47:29] I broke puppet on all instances [21:50:06] OrenBochman: OrenBo ?? [21:50:28] re [21:50:53] wassup [21:51:02] OrenBochman: should I go ahead and announce w/o search? [21:51:14] ok [21:51:24] sounds like it isn't going to happen, or am I missing someting? [21:51:37] sorry I wasn't around earlier [21:51:56] I might get it indexing tommorow [21:52:08] I put up a couple of task in the etherpad [21:52:17] k, looking [21:52:24] if they are ok I'll try again later [21:53:14] Ryan_Lane: I have no idea how to redirect it in apache... [21:53:26] deployment to beta [21:53:39] johnduhart: around? [21:53:47] it's a rewrite rule [21:53:50] based on hostname [21:53:52] I know but it's too hard [21:54:04] *.xx.domain to *.gsg.domain [21:54:09] I don't know how to do that [21:54:17] I can do simple redirect but no idea how to do this [21:54:28] need to get two $ [21:54:34] one for domain and one for content [21:56:16] !log bastion installed tofrodos and python-paramiko on bastion1 [21:56:17] Logged the message, Master [21:59:17] hexmode: I put in evrything except about further modifing global config which is trail and error [21:59:36] k [22:00:07] petan: you're trying to figure out apache redirects [22:00:07] I got to go [22:00:21] see you tommorow [22:00:22] petan: I can probably give you some mod_rewrite help [22:00:28] OrenBochman: thanks! [22:00:59] btw we can make a second anouncement on search when it ready [22:01:25] I'm revamping the extention's documentation on this type of use case [22:01:39] OrenBochman: right. And I think having you to help us figure it out is a big bonus. [22:09:42] petan: hi [22:10:09] * johnduhart gets a new battery for his mouse... [22:17:12] how is http://commons.wikimedia.beta.wmflabs.org/ updated ? [22:17:22] is there some kind of cronjob that svn update it? [22:18:12] The code? [22:18:20] The code for all the sites is updated by hand [22:18:29] svn up [22:20:51] k k [22:20:54] thx john [22:21:26] kudos on replicated the cluster [22:22:23] Thank you, I appreciate it. [22:29:53] PROBLEM Disk Space is now: CRITICAL on cantlogin cantlogin output: DISK CRITICAL - free space: / 0 MB (0% inode=19%): [22:37:43] PROBLEM Free ram is now: WARNING on deployment-sql deployment-sql output: Warning: 19% free memory [22:40:55] I'm force running puppet everywhere [23:02:43] RECOVERY Free ram is now: OK on deployment-sql deployment-sql output: OK: 20% free memory [23:03:51] NS records are accurate now [23:04:07] I need to finish this puppet stuff before I try dns stuff again, though. heh [23:25:04] hexmode: can you do that? [23:25:12] redirect it [23:29:47] Krinkle: these testswarm instances are killing me [23:29:59] RECOVERY Disk Space is now: OK on cantlogin cantlogin output: DISK OK [23:29:59] Ryan_Lane: plural ? [23:30:03] cantlogin [23:30:25] and testpuppet [23:30:41] I deny everything, haven't been on testswarm-test for a while [23:31:10] testpuppet is *really* slow [23:31:31] like can't run top slow [23:31:31] no idea why [23:33:15] Ryan_Lane: you can pause that VM if you want [23:33:16] I can't even log on it anyway :-) [23:33:32] that doesn't make any sense [23:33:36] is it puppetized? [23:34:32] yeah [23:34:33] can we delete them? [23:34:33] I have set up that VM to actually test puppet [23:34:34] would prefer if you pause it or something [23:34:34] but if it is not possible, yeah delete it [23:34:35] I can pause it [23:34:37] pause would be great [23:34:44] if it does not fix whatever issue you encounter, stop it for good :b [23:35:18] well, pause just means it stays suspended on disk [23:35:45] it won't fix anything [23:54:14] who are you trying to add? [23:54:15] yeah, nvm cloud admin seems irrelevant [23:54:18] sumanah: ^^ [23:54:18] and you're not a sysop. i wonder why i thought you were [23:54:23] I want to make sure I can add them [23:54:26] Ryan_Lane: bsitu. [23:54:40] hm. yeah [23:54:42] worked for me [23:54:57] Ryan_Lane: well, I'm guessing you have A WEE BIT more privs than me [23:55:11] 01/13/2012 - 23:55:11 - Creating a home directory for bsitu at /export/home/bastion/bsitu [23:55:11] I was not in the relevant project so I couldn't add them to the project. [23:55:13] I'm looking at the error logs now [23:55:32] that is, on labsconsole, user:Sumanah is not in the testlabs project, Ryan_Lane. [23:55:47] right [23:55:53] looking at debug logs, I should say :) [23:56:11] 01/13/2012 - 23:56:11 - Updating keys for bsitu [23:56:16] huh. weird [23:56:22] 2012-01-13 23:52:27 labswiki: 2.0a Created a regular filter: (=jeremyb) [23:56:43] i'll say [23:56:46] what TZ? [23:56:57] 2012-01-13 23:51:17 labswiki: 2.0a Created a regular filter: (cn=sumanah) [23:56:57] heh [23:57:25] 2012-01-13 23:51:13 labswiki: 2.0a Couldn't find the user in project: testlabs [23:57:39] * Ryan_Lane goes to check something [23:59:25] o.O [23:59:26] wtf [23:59:34] oh wait [23:59:36] Ryan_Lane: ok, may I hand off the process of giving bsitu a labs account to you? https://wikitech.wikimedia.org/view/Gerrit is the checklist, I assume. [23:59:36] sumanah: sorry [23:59:42] sumanah: bastion is now the correct project [23:59:44] not testlabs [23:59:56] Ah.