[00:23:19] Coren, wasn't S7 supposed to be ready today? :p [00:26:33] petan, ^ [00:39:09] hi [00:39:24] hey :P [00:39:46] tools.wmflabs.org/samoabot/ [00:39:54] how-to? [00:40:25] how far have you got so far? :P if anywhere..? [00:40:58] I have a Labs account, a Tools account and one for my bot [00:41:33] lovely :) all your bots files are on it? :> [00:41:55] not yet :-7 [00:42:43] well :0 [00:42:56] become your tool and copy them all into your tools folder [00:42:57] :) [00:44:58] there is a lovely documentation page at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help :) [00:45:28] just going to finish off the script, and we're ready to go [00:45:33] :> [00:45:44] the main part of the docs that youll find usefull is https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Submitting_Jobs [00:52:08] where to find mwparserfromhell in WMFlabs? [00:54:01] hmmm, /data/project/pywikibot has pywikibot and rewrite [00:54:45] also at /shared/pywikipedia (probably better to use this) [00:55:23] I have to sleep now :< [00:56:32] bye [02:13:13] Coren: do you have time to check an obvious LTA sock on enwiki? [03:03:33] petan: Coren: could one of you install pyexiv2 on the exec nodes, please and thank you? [03:03:49] theo|cloud|away [03:09:56] theo|cloud|away: Probably neither of them is around now. You should probably file a bug in bugzilla, under product "Wikimedia Labs" component "tools". [03:10:24] thanks much...erp, i mean, goodnight [05:22:43] theo|cloud|away ok [05:31:07] done [05:44:47] Warning: There are 2 users waiting for shell, displaying last 2: Blurbz (waiting 1 minutes) Dg2ygx (waiting 1 minutes) [05:54:13] Warning: There are 2 users waiting for shell, displaying last 2: Blurbz (waiting 1 minutes) Dg2ygx (waiting 1 minutes) [05:54:23] * petan is pissed [05:54:24] !ping [05:54:24] pong [05:57:47] Warning: There are 2 users waiting for shell, displaying last 2: Blurbz (waiting 1 minutes) Dg2ygx (waiting 1 minutes) [09:19:52] @notify Coren [09:19:52] This user is now online in #wikimedia-labs. I'll let you know when they show some activity (talk, etc.) [09:31:52] !toolsadmin [09:31:53] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Documentation/Admin [10:03:55] petan: wassup? [10:04:08] Oren_Bochman did I ping you? :P [10:04:14] no [10:04:22] just saying hi [10:04:37] u working in the weekend? [10:15:47] Oren_Bochman_Awa: weekends a time for volunteers ;) [10:16:06] but yes I am in office as well... [10:16:14] * Oren_Bochman_Awa agrees [10:40:16] hmm [10:40:42] petan: I'm currently not going to have internet in my flat in germany! [10:53:26] addshore get some wireless net / mobile [10:53:51] there is surelly very fast mobile internet, at least we have it in Prague [10:54:40] I will have to do some researching in my first week :> [10:54:49] heh [10:54:56] or... [10:55:23] you can get internet from german o2 which I work for... and I could eh, implement some bug that will make you invoices be 0 always :D [10:55:35] HAH! xD [10:55:38] you wouldn't be first :P [10:55:47] interestingly my internet at home is with O2 atm :P [10:55:48] there was a guy that got stuck in system for 2 years or more [10:55:58] was getting no invoices for years [10:56:22] I have only access to german servers :P [10:56:27] :< [11:01:33] addshore: I'm having trouble connecting to the database from my php code [11:01:41] which db? :) [11:01:57] anbd using what code / details? :) [11:02:25] I'm using $connection = mysqli_connect($dbhost,$dbuser,$dpbass,$dbname); [11:03:28] not sure how to connect using the replica.my.cnf [11:06:14] * YuviPanda is playing still because there's a cat around [11:06:23] in php I currently just have a replica.cfg which I load where I have added the details from replica.my.cnf [11:08:20] I get Database connection failed: Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) (2002) [11:08:42] I use enwiki as th edabase name [11:09:12] and the replica.my.cnf details for user and pass and local host as the host [11:09:34] makde that localhost [11:10:31] the host isnt localhost ;p [11:10:43] the host enwiki should work and the database enwiki_p [11:11:03] * Oren_Bochman_Awa trys [11:12:43] I now get "Database connection failed: Unknown MySQL server host 'enwiki' (2) (2005)" [11:13:46] sorry, host is enwiki.labsdb [11:15:33] ok [11:15:38] I found that [11:19:41] now I get permission denied error [11:19:59] Access denied for user 'p50380g50531'@'10.4.0.128' (using password: NO) (1045) [11:20:36] using password: NO [11:20:42] you need to use the pasword :) [11:21:16] just curious you copied this [11:21:17] mysqli_connect($dbhost,$dbuser,$dpbass,$dbname); [11:21:20] should it not be this [11:21:25] mysqli_connect($dbhost,$dbuser,$dbpass,$dbname); [11:21:47] dpbass -> dbpass [11:23:46] I have $dbpass="copied from the replica.my.conf"; [11:24:26] but in mysqli_connect you are trying to pass it $dpbass instead of $dbpass [11:24:42] petan: around? [11:24:45] ok [11:25:30] it does not crass any more! [11:25:39] crash [11:25:57] :D [11:26:05] :-) [11:57:18] YuviPanda yes [12:29:04] petan: can you install redis and mongodb on tools-mc? [12:29:44] hmm [12:29:54] I would prefer first set it up on bots-mc [12:30:03] once it works we can push it to "production" [12:30:12] I don't have any experience with setting it up :/ [12:31:27] andrewbogott_afk ping [12:31:37] paravoid? :o [12:31:43] petan: apt-get should work [12:31:57] YuviPanda mhm... [12:32:03] sure, bots-mc sounds good too [12:32:08] isn't there some setup needed? [12:32:17] we can't use bots-mc until some wmf op wake [12:32:52] petan: not for redis, definitely [12:32:55] i'm unsure about mongo [12:33:01] !log tools petrb: installing redis on tools-mc [12:33:04] Logged the message, Master [12:34:36] YuviPanda that redis server needs to be configured first [12:34:42] there is example config, which is huge [12:34:50] it's going to take some time unless you help me [12:34:53] I just did a brew install redis and that works [12:34:56] I can help you, yes! :) [12:35:02] ok, then ssh to tools-mc [12:35:15] there is /etc/redis/*sample [12:35:17] * YuviPanda sshs [12:35:36] create some proper version of that config and save it to /tmo [12:35:39] /tmo [12:35:41] /tmp [12:35:42] reading [12:35:43] omg [12:35:44] :P [12:35:46] okay [12:35:47] moment [12:43:03] petan: I made it log to /etc/log/redis [12:43:15] petan: where should I put the data? /var/data/redis? [12:43:18] or someplace else? [12:43:20] huh? [12:43:26] you mean the config file? [12:43:28] /tmp [12:43:37] no, I mean where should redis store its persistance files? [12:43:58] ah [12:44:01] hold on [12:44:06] I will create a volume for it [12:45:58] hm... what about /var/redis [12:47:16] I set it to 1gb but it can be extended anytime [12:49:57] YuviPanda how much do you expect redis to use? we have 14gb free on vg00 [12:50:15] I can get some 4gb more... [12:52:40] petan: i don't think it should go on that much [12:52:50] ok [12:52:52] depends on how much people use it [12:53:03] /var/redis created [12:53:03] /var/data/redis is more usable, no? [12:53:09] why? [12:53:25] petan: you aren't supposed to have first level dirs in /var for things [12:53:26] afaik /var/data or /usr/data is a dangerous location :P [12:53:37] even supertux was creating that directory by default :D [12:53:43] it usually conflict a lot [12:53:44] supertux? [12:53:46] yes [12:53:48] the game [12:54:02] oh c'mon :P [12:54:07] ok then /var/lib/redis ? [12:54:08] it's /var/data/redis [12:54:12] eek [12:54:16] this isn't lib no? [12:54:20] mysql is using that too [12:54:21] also where does mysql put its data files? [12:54:24] hmm [12:54:28] by default mysql use /var/lib/mysql [12:54:29] for db [12:54:37] ah [12:54:38] okay then [12:54:45] /var/lib/redis [12:55:10] it already exist [12:56:04] ready [12:56:44] I think it is default location for redis [12:57:43] petan: ah [12:57:44] ok [12:57:57] petan: then the current config file looks good enough to me [12:58:03] where is it [12:58:22] petan: saved to /tmp/redis.conf [12:58:28] does redis use authentication? [12:58:44] petan: it doesn't do per database authentication, no [12:58:47] btw as soon as I enable persistent storage for sharp_memcached it will be redis like :D :D [12:59:07] we can however disable some set of commands easily [12:59:10] in the config file itself [12:59:32] so provided that the config file is not word readable [13:00:03] that config doesn't have /var/lib/redis inside [13:00:15] if that is the default it has ./ [13:00:17] so that should be fine [13:00:20] let me doublec hec [13:00:21] *check [13:00:35] why not just put /var/lib/redis there [13:00:51] that would make sure it doesn't write somewhere else [13:01:04] yeah let me do that [13:01:09] ok [13:01:50] done [13:01:54] ok [13:01:54] try moving + starting it? [13:02:13] ok [13:02:23] petan: shall we disable "CONFIG" [13:02:29] idk [13:02:39] we should [13:02:41] I don't know how redis work [13:02:43] it's like disabling STAT on memcached [13:02:44] okay [13:02:46] aha [13:02:46] let me then [13:02:47] moment [13:02:47] ok [13:08:16] petan: i think that should do it [13:08:29] 9 commands disabled [13:08:40] so if people use a cryptographically secure prefix it should be fine [13:10:35] mhm [13:11:04] petan: remind me - why don't we just kill bots and replace to with a new project called 'tools-infrastructure-dev' or something like that? [13:11:19] (see http://redis.io/commands) [13:11:52] valhallasw because we already have working infrastructure? :P idc... you should maybe ask Coren... [13:12:00] petan: you can copy the infrastructure, right? [13:12:03] that is another option, of course [13:12:11] we can't copy it, but we can recreate it [13:12:53] YuviPanda safe to start? [13:12:57] we can recreate it, we have the technology! :P [13:13:10] petan: yes, after copying over /tmp/redis.conf to appropriate place [13:13:15] I did [13:14:05] good to go i think :) [13:14:36] Reading the configuration file, at line 53 [13:14:37] >>> 'logfile /var/log/redis' [13:14:38] Can't open the log file: Is a directory [13:14:39] :o [13:15:07] hmm [13:15:13] try changing that to just 'redis.log'? [13:15:30] ok [13:15:30] I think it is picking up everything from $CURDIR/filename [13:16:15] Reading the configuration file, at line 251 [13:16:16] >>> 'rename-command DEBUG SEGFAULT ""' [13:16:17] Bad directive or wrong number of arguments [13:16:34] hehe, irc is now setup to push a ping to my phone :> [13:16:47] hmpf, kill the 'SEGFAULT', let's disable all debug commands [13:16:51] (until someone asks) [13:16:53] addshore: irccloud? [13:16:59] YuviPanda how? [13:17:01] comment out? [13:17:06] petan: no, change it to [13:17:09] rename-command DEBUG "" [13:17:33] rename-command DEBUG SEGFAULT "" [13:17:40] to rename-command DEBUG "" [13:17:42] ? [13:18:09] yes [13:18:49] lol [13:18:52] it's starting [13:18:57] like 1 minute :D [13:19:13] what is it written in [13:19:16] it seems so slow [13:19:23] it's C [13:19:26] :o [13:19:26] (IIRC) [13:19:34] what do you mean it is stil starting? [13:19:36] has it started? [13:19:38] well, they say that firefox is written in c++ [13:19:52] in fact it is just a nice interpretor of JS and whole browser is in JS [13:19:54] hmm, i wonder how I can find out [13:20:07] * YuviPanda logs into tools [13:20:09] petrb@tools-mc:~$ sudo /etc/init.d/redis-server start [13:20:10] Starting redis-server: [13:20:35] it still didn't return back to shell [13:20:37] aaah [13:20:38] wait [13:20:43] i'm an idiot [13:20:46] :o [13:21:10] is it defragmenting fs or something :D [13:21:47] petan: search for /deamonize no/ and turn it to /daemonize yes/ [13:21:52] ok [13:21:53] it's running in blocking mode :P [13:22:01] it's on line 17 [13:22:25] done [13:22:38] did it start? [13:29:13] petan: let me check by randomly connecting to it :) [13:29:53] petan: can you also install redis-cli on tools? [13:30:26] petan: works, nevermind :) [13:35:00] petan: thanks [13:36:45] YuviPanda: yup :) [13:36:58] * YuviPanda is reading Twisted documentation [13:37:18] addshore: i'm writing a small server that'll let anyone connect to and receive events from gerrit stream-events [13:37:29] over 0mq, in an easier and more importantly, reliable / persistant way [13:37:37] :D [13:37:58] addshore: this also implements filtering on the server side (possibly), so clients only get the events they're interested in [13:38:12] addshore: if this goes well, we can then do a similar thing for the IRC RecentChanges feeeds [13:38:20] then connecting to them and using them will become super simple [13:38:32] :D [13:42:48] YuviPanda ok [13:42:52] YuviPanda on login? [13:43:06] petan: on login definitely [13:43:19] also on the exec nodes/ [13:43:22] ? [13:44:46] no such package... [13:44:50] where should I get it? [13:47:03] petan: no such package? [13:47:08] no [13:47:17] YuviPanda there is aptitude on -dev [13:47:20] you can look [13:47:39] you mean bots, right? [13:49:20] petan: okay, so you can install redis-server and disable it [13:49:26] or just leave it, it's okay I guess [13:49:38] people can always ssh into tools-mc to checkout redis [13:49:42] :/ [13:49:49] what is that redis-cli for? [13:49:53] what it consist of [13:50:27] petan: it's client [13:50:33] to interactively deal with Redis [13:50:35] as i said [13:50:37] it's okay to ignore it [13:50:43] cli should be command line interface [13:50:50] it is a command line interface client [13:50:51] yes [13:50:53] is it really needed for app's to work with redis? [13:50:55] no [13:50:56] it isn't [13:51:01] it's good for testing [13:51:03] but as i said [13:51:06] it's okay to ignore it [13:51:08] so why should we have it on exec nodes... [13:51:10] ok [13:51:16] what was the second thing you needed [13:51:44] mongodb... [13:52:02] sounds like a database server to me [13:55:04] petan: yeah, let's ignore it for now [13:55:05] and see how our Redis stuff holds up [13:55:13] petan: I think this thing is secure enough [13:55:19] if used with a cryptographically secure prefix [14:16:13] petan: about "[Wikitech-l] [labs] bots project is turning to testing environment of tools project" do I need to erase private data left there now? [14:16:27] nope [14:17:15] but depends on what you consider private data - it is generally easier to get root on bots project [14:17:24] people with root there don't need to sign any nda... [14:17:38] (there is much more of them) [14:18:17] so if by private data you mean data that are supposed to be read only by you and other unknown people should have access to them, you /should/ remove them [14:18:22] but you aren't required to do that [14:18:39] (should not have) [15:23:52] petan: well I mean password files, used for my bot running on bots in the past (now on tools) [15:24:13] yes passwords files aren't very safe on bots project... [15:26:53] sul on tools! :D [15:26:54] http://tools.wmflabs.org/sulinfo/sulinfo.php [15:27:02] o/ [15:33:19] it is very slow [15:35:48] addshore http://www.softpedia.com/get/Internet/Chat/IRC-Clients/Pidgeon-Client.shtml :D :D [15:35:50] wheheee [15:36:16] I just received an e-mail they wrote article about software they found that I made [15:37:50] petan: centralauth db is ready now? [15:38:03] liangent I don't think so :/ [15:38:15] so sulinfo is working via api? [15:38:19] nope [15:38:25] it just doesn't work well [15:38:36] it doesn't need centralauth to look up information about users [15:38:46] it only need it to check if it's attached to central login [15:38:56] try to look up someone [15:39:04] you will see it show all accounts as unattached [15:39:19] but it "works" at least a bit... [15:39:36] it's a small step for one programmer... but for labs :D [15:39:48] petan: :) [15:41:21] can the "List of unattached accounts" gets sorted? [15:41:24] Nice pet an [15:41:35] ;p [15:43:40] :D [15:43:52] liangent: mmmm dunno [16:13:24] addaway, liangent if anyone wanted to participate on that sulinfo project tell me [16:28:00] addaway my bouncer crashed and didn't throw any meaningful error! [16:28:01] I hate this [18:23:07] !log tools petrb: exec-04 96% / usage, creating a new volume [18:23:08] Logged the message, Master [18:23:58] !log tools petrb: temporarily disabling /tmp on exec-04 in order to set up lvm [18:23:59] Logged the message, Master [18:35:59] !log tools root: removed lot of ?audit? logs from exec-04 they were eating too much storage [18:36:01] Logged the message, Master [20:59:07] Ryan_Lane: o/ [20:59:47] Ryan_Lane I wanted something but I always forget... I know I needed extend instance quota on bots project [21:00:02] as well as public IP quota [21:05:00] ok I don't need to extend the quota then :P [21:05:12] @notify Coren [21:05:12] This user is now online in #wikimedia-labs. I'll let you know when they show some activity (talk, etc.) [21:13:10] petan: eh? [21:13:15] you do or dont? :D [21:13:31] I wanted for bots, but if we are going to create new project I don't [21:13:37] ah [21:13:38] right [21:13:48] yeah. I wish it was easier to rename projects [21:13:51] but it's relatively hard [21:13:52] but I need that new project :D I wanted to talk to Coren before though [21:14:28] well, you don't neccesarilly need to rename them in ldap, some kind of alias so that wikitech interface would display different name [21:14:32] should be enough [21:14:44] that's just confusing ;) [21:14:45] like alias (original name) [21:14:56] hmm [21:15:24] we have sul in labs now :D http://tools.wmflabs.org/sulinfo/sulinfo.php but it's slow [21:15:27] much slower than toolserver [21:15:57] I suspect database servers being slow [21:16:03] also s7 doesn't work :/ [21:19:37] Hey folks. I was given an API key to OCLC's worldcat search engine. They want to show us their capabilities, potentially for use in The Wikipedia Library project. I am not sure what the next step is, to access the API through Labs. I have a labs account and links to the API documentation [21:19:37] (http://www.oclc.org/developer/platform/authentication-documentation) (http://www.oclc.org/developer/services/worldcat-search-api) [21:30:12] petan: okay, so now I have a continuous job that is treaming gerrit events into Redis :) [21:30:27] next is to set that up for subscription for everyone else... [21:36:52] PandaWithACat cool [21:39:10] !log tools petrb: db has 5% free RAM eeeek [21:39:12] Logged the message, Master [21:40:33] !log tools petrb: there is no lvm on -db which we need as hell - therefore no swap either nor storage for binary logs :( I got a feeling that mysql will die oom soonish [21:40:35] Logged the message, Master [22:19:48] petan: ping [22:19:58] is it possible to setup a continuos tool that provides 'network services'? [22:20:03] essentially, a 0mq server [22:20:15] so it would need a stable ip [22:21:20] woohoo wsgi coming to labs soon :) [22:21:21] depends [22:21:30]