[00:03:18] Coren, ^ [00:03:59] carl-cbm, https://bugzilla.wikimedia.org/show_bug.cgi?id=47538#c4 [00:05:25] (I have no idea what you're even discussing, but I remember seeing that comment and it looks relevant) [00:23:47] Merlissimo: Are you sure you're not just not /seeing/ the replication because you're in a transaction? [00:24:17] carl-cbm: The exec nodes aren't submit hosts; they couldn't use qstat even if it were installed. [00:24:50] Coren: he's referring to http://ganglia.wmflabs.org/latest/graph_all_periods.php?c=tools&h=tools-login&v=3&m=s5&r=week&z=small&jr=&js=&st=1369957795&z=large [00:25:28] so it's not what *he* sees [00:25:29] Coren: i monitored replag from "outsite". you can also see my runs at 23:20 and 23:55 on http://ganglia.wmflabs.org/latest/graph_all_periods.php?c=tools&h=tools-login&v=1&m=s5&r=week&z=small&jr=&js=&st=1369959833&z=large [00:26:34] at 23:20 run there was a max replag of 219seconds [00:26:55] Hm. It's entirely possible that there is interaction between the way binlogs work and implicit table locks for consistency during a query. Given how quickly the replag goes back down, it's entirely possible the logs get only replayed afterwards. You'd have to ask someone with better understanding of mysql replication (aka Asher) for a definitive answer. [00:27:37] Coren: is there a query-killer running on the labs db? [00:27:49] it just a notice to keep an eye on this possible problem when load rises [00:27:56] * Coren nods. [00:28:03] my test query for importing enwiki which i will need for importing stuff to catgraph just got killed. [00:28:04] I'll certainly raise the issue with Asher. [00:28:12] JohannesK_WMDE: No, there isn't. [00:28:38] JohannesK_WMDE: Do you mean the local DB? [00:29:17] Coren: i mean enwiki.labsdb. [00:29:42] * Merlissimo needs to execute a lot of long running queries [00:30:26] * JohannesK_WMDE too :D [00:30:29] lol [00:34:35] I have two queries running since May 22nd on enwiki_p, and it hasn't caused any rise in replag. Merlissimo, could you post the query? [00:36:28] Hm. According to what doc I can find, this shouldn't be an issue with the current storage engine; there is no table locking by default. [00:38:23] $ echo 'SELECT /* SLOW_OK */ B.page_id, cl_from FROM categorylinks JOIN page AS B ON B.page_title = cl_to AND B.page_namespace = 14 AND B.page_id!=0 AND cl_from!=0;' | mysql --defaults-file=~/replica.my.cnf -h enwiki.labsdb enwiki_p > /tmp/enwiki-dump.dat [00:38:26] [...] [00:38:31] Killed [00:39:41] happened twice. got killed after 2 minutes or so. worked for dewiki though... [00:41:38] JohannesK_WMDE: ... I have no idea why that'd happen; there is no query killer running. [00:41:58] JohannesK_WMDE: But -login hasn't got a lot of memory. [00:42:24] Coren: i'm on -dev [00:42:48] and it should just pipe everything to /tmp. or is that on a ramdisk? [00:42:49] Does mysql format the output, when the query is fed on STDIN? [00:43:09] [876664.107588] Out of memory: Kill process 11649 (mysql) score 717 or sacrifice child [00:43:09] [876664.108112] Killed process 11649 (mysql) total-vm:3795148kB, anon-rss:1901880kB, file-rss:824kB [00:43:09] [876664.148043] mysql: page allocation failure: order:0, mode:0x201da [00:43:16] If it is formatted, it will load the whole result in memory to calculate the column widths. [00:43:21] ah, OOM [00:43:28] * Coren nodsnods. [00:43:41] You'd be better off submitting this to the grid. [00:43:57] i'd be better off not formatting it in any case? [00:44:12] JohannesK_WMDE: That's what I would try. [00:44:13] JohannesK_WMDE: That'd help as well, I'm sure. [00:45:01] if i turn on batch mode, will formatting be disabled? [00:45:53] --batch or --raw will disable formatting [00:46:48] k [00:46:52] Depends on whether you need escaping or not. [00:47:06] scfc_de: http://kpaste.net/31c3c25f0d [00:47:42] Merlissimo: Wow :-). [00:47:55] Holey sheets! [00:48:10] That's not a query, it's the Mother Query! [00:48:24] the query queen [00:48:52] do you want to see my long query codes, too? [00:49:22] JohannesK_WMDE: Just did some testing, and --batch doesn't seem to reduce max RSS when the query is fed on STDIN. [00:49:50] scfc_de: i tried --raw and got oom-killed again [00:51:35] JohannesK_WMDE: Could you try it with a short Perl/PHP/Python script? [00:54:57] JohannesK_WMDE: you should always add -wcAB to mysql arguments if you pipe that in [00:57:00] -w wating if connection limit is reached, -c add commons like your /* sl_ok */ -A speed up database access because of not hashing [01:02:32] and -B for batch. and --raw. still got oom-killed. will try a little python script tomorrow, after some sleep [01:10:49] night all! [01:15:13] Gute Nacht. [01:58:17] Merlissimo: where did that script come from [02:26:54] coren: where is the page listing all tool in tool labs ? [02:27:17] Oren_Bochman: https://tools.wmflabs.org/ [02:27:30] 10x [02:28:18] hmm orwell has a ui I guess it needs some code now :-) [02:28:26] JohannesK_WMDE: Need moar memories! :-) Try the grid, srsly. :-) [02:37:59] Coren: I changed my index.php but the changes are not appearing [02:38:40] Oren_Bochman: Err... there's no caching of any kind. You sure you're hitting the right tool? I see two orwells. :-) [02:38:52] yes [02:38:55] https://tools.wmflabs.org/orwell01/ [02:40:22] and the url is not reponding to ctrl+F5 or ctrl+r [02:43:44] hmm my bad [03:05:56] on typing : become ocounter [03:05:58] I get: sudo: sorry, a password is required to run sudo [03:07:00] what is that about [03:15:16] Oren_Bochman: why are you trying to use sudo? [03:15:34] I'm not I used become [03:15:53] the issue reolved when I logged out and logged into the tool labs [03:16:02] Oren_Bochman: are you part of the ocounter group? [03:16:08] yes [03:16:13] I just created it [03:16:24] but now become works [03:16:45] Oren_Bochman: if you create a project you need to log out/log in for your session to update your permissions [03:16:53] ok [03:17:07] took me a while to figure it out [03:17:17] its just a quirk with Unix file systems [03:18:10] permissions/user groups dont get updated immediately, just during login [03:20:40] thank Betacommand [03:20:52] btw were you in the AMS Hackathon ? [03:21:43] No [03:24:41] still you rock1 [03:45:39] Does anybody have an idea when I can expect Coren to be available again? [03:46:39] Cyberpower678: next week [03:47:00] Wait what? [03:47:05] He was on today? [03:47:11] yeah [03:47:28] So why next week? Is this a joke? [03:48:06] Cyberpower678: no [03:48:14] he is on vacation [03:48:30] Oh. [03:48:46] Could he at least quickly grant me access to archive? [03:49:03] that is Cyberbot, Peachy, and xtools [03:49:10] projects [04:32:47] !toolsdoc [04:32:55] wm-bot: ? [04:32:55] Hi legoktm, there is some error, I am a stupid bot and I am not intelligent enough to hold a conversation with you :-) [04:32:59] !tooldoc [04:32:59] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help [04:33:05] Warning: There is 1 user waiting for access to tools project: Heb (waiting 0 minutes) [04:46:34] Warning: There is 1 user waiting for access to tools project: Heb (waiting 13 minutes) [05:00:03] Warning: There is 1 user waiting for access to tools project: Heb (waiting 27 minutes) [05:13:24] Warning: There is 1 user waiting for access to tools project: Heb (waiting 40 minutes) [05:26:49] Warning: There is 1 user waiting for access to tools project: Heb (waiting 53 minutes) [05:40:19] Warning: There is 1 user waiting for access to tools project: Heb (waiting 67 minutes) [05:51:03] * Hazard-SJ says hello [05:52:12] Coren: Around, by chance? [05:53:18] he's on vacation, according to Betacommand [05:53:48] Warning: There is 1 user waiting for access to tools project: Heb (waiting 80 minutes) [05:54:10] hey [05:54:13] Thanks [05:54:16] in fact he is back [05:54:25] petan: You're actually around :O [05:54:46] !tr Heb [05:54:46] request page: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Access_Request/Heb?action=edit talk page: https://wikitech.wikimedia.org/wiki/User_talk:Heb?action=edit§ion=new&preload=Template:ToolsGranted link: https://wikitech.wikimedia.org/w/index.php?title=Special:NovaProject&action=addmember&projectname=tools [05:54:58] * Hazard-SJ never sees petan around nowadays :/ [05:55:15] I am around like every day :O [05:55:18] check logs [05:55:25] !logs [05:55:25] http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-labs/ [05:55:35] You're idle in here, I know [05:55:46] no I am talking in here [05:55:57] But anytime I pop in you're away ... well, usually :P [05:56:02] according to statistics I am most talkative at all :P [05:56:08] because I live in EU [05:56:25] lol [05:56:30] I just woke [05:56:47] And I should be getting ready for bed :P [05:57:06] But I'm somewhat annoyed :( [05:58:00] petan: Is it possible at all for me to use chown? [05:58:10] where [05:58:26] In local-hazard-bot [05:58:29] yes [05:58:32] there is take [05:58:53] take [05:58:59] :/ [05:59:02] * Hazard-SJ tries [05:59:05] that overtake it [05:59:29] petan: any recursive option? [05:59:46] I think it is recursive by default or not? [05:59:51] maybe try -r if not [06:01:37] It seems to be :) [06:12:37] brb 1h [06:12:45] * Hazard-SJ nods [06:19:28] zz_YuviPanda: ping [06:22:59] * Hazard-SJ gives up on this implementation [06:23:40] * Hazard-SJ says bye [07:00:34] here [08:11:56] Damianz I think you should tarball all the sql backups in your script and zip them [08:12:06] currently it produces huge data [08:12:14] that isn't effective :/ [10:54:57] !log tools petrb: killing process 3060 on -login (mahdiz 3060 1964 88 May30 ? 21:32:51 /bin/nano /tmp/crontab.Ht3bSO/crontab) it takes max cpu and doesn't seem to be attached [10:54:59] Logged the message, Master [12:03:39] Warning: There is 1 user waiting for access to tools project: Jimmyxu (waiting 0 minutes) [12:06:36] !tr Jimmyxu [12:06:36] request page: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Access_Request/Jimmyxu?action=edit talk page: https://wikitech.wikimedia.org/wiki/User_talk:Jimmyxu?action=edit§ion=new&preload=Template:ToolsGranted link: https://wikitech.wikimedia.org/w/index.php?title=Special:NovaProject&action=addmember&projectname=tools [12:07:29] petan: hey is there a page explaining the difference between 'bots' and 'tools'? [12:07:47] jimmyxu: there will be, there is no difference whatsoever [12:07:52] bots is a staging area of tools [12:08:04] that mean it's experimental version of tools project [12:08:09] well, it will be [12:08:15] at least, that is a plan [12:08:20] :P [12:08:28] petan: so technically I should work with my stuffs on bots and then move them to tools? [12:08:49] petan: or should I just stick to tools for now [12:08:54] you can actually work on tools right away. It's a testing environment for changes done on environment [12:09:05] stick to tools [12:09:08] okay [12:11:40] petan: we should really communicate 'use tools' to people who ask about bots vs tools [12:11:56] that is what I do? [12:12:42] right now there are so complicated systems running on "bots" project that can be hardly moved to tools project [12:12:43] petan: the way you described it is somewhat confusing: 'bots is a staging area of tools' etc [12:13:03] well I can't rename the project... [12:13:12] maybe in future once that is implemented to console [12:13:19] sure, but calling it a 'staging area of tools' does not help to reduce the confusion :-) [12:13:24] just stating 'ignore bots, use tools' does [12:13:37] ok but some people are running bots on "bots" [12:13:49] I can't just tell them "ignore bots" [12:13:56] there needs to be explanation why they should do that [12:14:18] except that those people are already in the labs world [12:15:23] basically: if you don't already know what bots and tools are for, you probably don't need to know anything other than 'use tools' [12:15:46] it's like less and more, or pip and easy_install :-) [12:27:15] !log tools petrb: removing project local-jimmyxu - per request on irc [12:27:16] Logged the message, Master [13:02:32] Warning: There is 1 user waiting for shell: Inkbug (waiting 0 minutes) [13:04:05] !log tools petrb: installing bashdb on tools and -dev [13:04:07] Logged the message, Master [13:09:46] The 'Rules' link in tools-login motd seems to be dead... [13:15:13] jimmyxu: the rules don't exist yet :-) [13:15:30] okay then... [13:16:06] Warning: There is 1 user waiting for shell: Inkbug (waiting 13 minutes) [13:29:40] Warning: There is 1 user waiting for shell: Inkbug (waiting 27 minutes) [13:39:07] legoktm: pong? [13:43:05] Warning: There is 1 user waiting for shell: Inkbug (waiting 40 minutes) [13:45:33] !rq Inkbug [13:45:33] https://wikitech.wikimedia.org/wiki/Shell_Request/Inkbug?action=edit https://wikitech.wikimedia.org/wiki/User_talk:Inkbug?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/Inkbug [14:21:28] Hello, Labs! [14:25:28] hi [14:25:34] I have several things to you [14:25:49] hello Coren [14:25:51] Coren: did you read my email I sent you regarding +2 to labs/tools or whatever it's called? [14:26:02] also there is a bunch of stuff that should be merged in puppet [14:26:02] for some reason that sounded like "Hello Newman" in my head [14:26:16] Coren: there are also these 2 exec nodes... [14:26:20] we didn't finish since hackaton [14:26:38] petan: I have. It would make sense. I'll eventually figure out how. [14:26:48] petan: The two new nodes are on my plate for today. [14:26:52] ok [14:27:01] petan: Most of the stuff to merge in puppet I did yesterday. [14:27:06] ok [14:38:03] your branch is behin .... by 60 commits and can be fast... [14:38:08] how do I fix it [14:38:15] !gitexpert [14:38:17] ^^^ [14:38:58] git pull -ff [14:39:31] I tried git pull origin [14:40:06] hmm [14:40:12] git is slow :/ [14:40:20] neither works :( [14:47:15] ok I will do my favorite rm -rf which is the only command that actually works... [14:50:31] I am waiting for a day when git finally will become better than svn [14:50:51] * ConfusedPanda sends petan back to the day git was first released [14:51:00] it still suck [14:51:04] maybe it was better back then? [14:51:07] petan: are you on a branch off your own, or are you working on an existing branch? [14:51:17] I am on production branch [14:51:29] martijnHH btw just for smart guys like you I tarballed the fucked repo [14:51:31] before I deleted it [14:51:51] so if you want to spend your weekend fixing it and then convincing me it's actually easy to fix it, you can! [14:54:40] oh, cool! I want to get some more gitting experience. I'll trade you, I'll spend my weekend trying to do that and tell you what you could have done, you take care off bug 48859 for me ;) [14:55:04] !b 48859 [14:55:04] https://bugzilla.wikimedia.org/48859 [14:56:07] martijnHH I don't understand what you want in that bug [14:56:15] system wide software request? [14:56:32] if you mean to install anything to default labs image that is likely not going to happen [14:56:45] it is supposed to be as small as possible, it contains only software necessary to boot [14:57:47] ugh, I don't know what I'm talking about [14:58:26] I would like it available when I log in to shell. I don't know how to formulate that [14:58:32] to shell where? [14:58:44] labs contains many project with many virtual servers, each for different purpose [14:58:54] tools-login / tools-dev [14:58:59] so tools project! :D [14:59:01] here we go [14:59:03] changing the bug... [14:59:28] the problem to being new to something is that you don't know enough to properly formulate your request [14:59:36] at least, that goes for me [14:59:54] ok this can happen [15:00:05] you want that to live only on -dev or even on -login or somewhere else? [15:01:38] well, it's mainly a build tool, but it would be really nice if I could also use its capabilities in jobs [15:01:54] I'm don't really know what that would entail [15:03:35] !screenfix [15:03:35] script /dev/null [15:07:47] being able to DESCRIBE queries on labsdb would be really useful [15:09:24] martijnHH in that case you want that cluster-wide (on execution nodes as well) [15:09:40] that isn't that simple as long as it's not in a repository :/ [15:10:25] I can install this to -dev right now... [15:13:00] petan: AFAICT, -05 and -06 are now properly on queue. [15:13:07] !log tools petrb: installing /data/project/.system/deb/all/sbt.deb to -dev in order to test it [15:13:09] Logged the message, Master [15:13:16] JohannesK_WMDE: Need moar memories! :-) Try the grid, srsly. :-) [15:13:22] (in re yesterday) [15:14:06] JohannesK_WMDE: Known limit is mysql >=5.0; can't explain without having SELECT right on the underlying table, which we can't do. I'd be glad to run an explain for you though. [15:14:24] !log tools petrb: installing default-jre in order to satisfy its dependencies [15:14:25] Logged the message, Master [15:14:39] Coren: the import query works now that i used the normal _mysql bindings directly, without the added python bloat. but yes, i will need a vm with MOAR RAM for everything. but that was clear from the start [15:15:28] JohannesK_WMDE: Even if I give you a new instance, it'll still be gridengined for reliability. Only I'll make a special queue for you. :-) [15:15:50] JohannesK_WMDE: So that they can live together, and have lots of little baby processes. :-) [15:16:02] yay! :p [15:16:04] martijnHH it's on -dev [15:16:15] man sbt [15:16:17] meh [15:16:19] nvm [15:16:58] what does that actually mean Coren? "will still be gridengined" -- does that mean i will have gridengine available? yes, i would want that :) [15:17:01] Coren howcome we have no JRE anywhere on tools :P [15:17:45] Coren, hi. [15:17:59] petan: Watcha mean? java version "1.7.0_21" [15:17:59] OpenJDK Runtime Environment (IcedTea 2.3.9) (7u21-2.3.9-0ubuntu0.12.04.1) [15:18:12] mhm... there was none on -dev [15:18:13] Cyberpower678: Hoy hoy. [15:18:29] Do you think you could do something for me? :) [15:18:38] petan: Hm. Might be an oversight. [15:18:59] Coren: what was the reason we couldn't use these 2 exec nodes? [15:19:47] petan: I'm not entirely certain. I think I was just too distracted to turn them on right while at the Hackaton. I just added them normally to the queues and fired 'em up and it worked. :-) [15:19:59] Coren, Do you think you could do something for me? :) [15:20:01] ok :) [15:20:30] Cyberpower678: That what they pay me the big^H^H^Hmediu^Wreasonably-sized bucks for. [15:20:44] Cyberpower678 I think he could but you first need to tell him what... [15:21:05] Coren, you said you could let me access archive in the replication. [15:21:16] Do you think you could set that up? [15:21:33] Cyberpower678: That's not going to be instantaneous; it'll need some back-and-forth with legal. ~2 weeks, I'd expect. [15:21:58] Well so much for migrating X!'s edit counter and adminstats for now. [15:22:50] Cyberpower678: I need to set up a researcher-equivalent right with the same approval structure for Tool Labs first. [15:23:21] researcher has access to archive. [15:23:34] I know. That's why I need to set that up. :-) [15:24:25] Cool. [15:24:36] Coren, also give CU and Oversight. :D [15:24:47] * anomie updates http://tools.wmflabs.org/anomiebot/packages.html now that Coren and petan have the new exec nodes up [15:25:03] Cyberpower678: That's never going to happen. :-) [15:25:18] Then I'll just hack Wikipedia. [15:25:59] anomie can you tell me how it works? [15:26:05] anomie I would like to see the script which does it :> [15:26:10] seems handy [15:26:20] Coren, how do you connect to the DBs again? [15:26:49] Cyberpower678: Easy-sauce version: sql [15:26:58] Like: sql enwiki [15:27:03] petan: Sure. Code is in /data/project/anomiebot/pkgs. It just submits a job on each node that does dpkg --get-selections, then filters the results after they all complete. [15:27:13] Coren: what did you mean by "will still be gridengined" [15:27:29] JohannesK_WMDE: That the instance will be a grid engine exec host. [15:27:35] it seems that lot of stuff has changed in default images [15:27:43] there is no iptables for example [15:27:54] Coren, where's the replag table? [15:28:10] I am wondering howcome I installed ksh on exec-06 and not on 05 o.O anyway I think that ksh and csh should be default everywhere... [15:28:25] it's just a common shell [15:28:31] Cool. I can see the cu_log table. :p [15:29:30] and what does that imply exactly (or: why do you want that), Coren? [15:30:01] JohannesK_WMDE that the user processes (like bots) will run on it [15:30:40] Coren, you still there? [15:31:07] doesn't that kind of run counter to the possibility of permanently allocating CPU cores, Coren/petan? [15:31:12] JohannesK_WMDE: So that it remains managed through the same scheduler, and job status can be seen the same way as everything else. [15:31:36] JohannesK_WMDE: I can dedicate the host to a specific queue(s), for that tool. [15:31:41] Cyberpower678: Yes? [15:31:55] Coren, where's the replag table? [15:32:03] ok... [15:32:17] Cyberpower678: There isn't one yet. Nor is there a metainformation table. Also "yet". [15:32:32] Coren, ETA? [15:32:58] Cyberpower678: Metainformation soonish, next week I guess. replag is low priority; maybe the week after that. [15:33:33] So the edit counter will still need to wait a few weeks before it gets deployed on labs then. [15:34:27] o_O Why in blazes would an edit counter need a replag table? [15:34:28] !g 66264 [15:34:28] https://gerrit.wikimedia.org/r/#q,66264,n,z [15:34:43] Coren I think this needs a merge :o [15:34:57] Coren, the replag is informative when people look at the edit count of a user. [15:35:15] It's helped a lot for me too. [15:35:20] I was surprised that mysql-client wasn't there as all exec nodes have it [15:35:27] except for latest 2 [15:36:12] Cyberpower678: It's completely pointless in this setup. Why would you even /want/ to know that the editcounter is 5 or 20 seconds behind? Nobody needs that level of detail, it's certainly not worth stalling a tool to wait for it. [15:36:19] YuviPanda: i figured it out :D [15:36:34] Coren, not that specific. [15:36:51] When replag is high, that's when the message gets dispalyed. [15:37:00] oh, what was the deal, legoktm [15:37:06] Coren: It probably was more important on the toolserver where replag tended to get hours behind. [15:37:10] Like on toolserver when it was replaged for 48 hours that one time. [15:37:29] Coren: why do we have apache servers running on exec nodes? [15:37:56] petan: There shouldn't be. Some oddity, I expect, with image experimentation for a while. [15:38:06] I have no idea how well labs handles replag but I think it would still be nice to have the replag notice there when there ever is a time replication stops all of a sudden. [15:38:12] hmm, they are even in latest version so it's likely somewhere in puppet [15:38:15] Cyberpower678: More useful would be the timestamp of the last edit. [15:41:07] petan: -2'ed. Asking for mysql-client when we use mariadb-common == trouble. :-) [15:41:37] ok I was thinking it's aliased [15:46:23] !log tools petrb: replacing mysql with maria on exec nodes [15:46:24] Logged the message, Master [15:47:05] !log tools petrb: replacing mysql with maria on exec nodes [15:47:06] Logged the message, Master [15:47:10] damn [15:47:12] :D [15:47:21] bash history being counterproductive [15:48:48] at some point that sql tool should be probably everywhere as well [15:48:58] what if someone wanted to make a shell script which use it and submit to grid? :P [15:50:02] YuviPanda: actually i do need help with something. i was using the sh module, and when i did sh.svn.up() it failed, then I did svn = sh.Command('/usr/bin/svn') that failed too [15:50:13] it said it didnt exist or something when running w/ jsub [15:50:22] legoktm: PATH issue perhaps? [15:50:32] legoktm: does svn exist in the exec nodes? [15:50:50] i ssh'd to tools-exec-1 and svn worked there at /usr/bin/svn [15:51:14] ohhhh [15:51:16] not all of them have it [15:51:21] local-pywikibot@tools-exec-02:~$ svn [15:51:21] The program 'svn' is currently not installed. To run 'svn' please ask your administrator to install the package 'subversion' [15:51:26] * legoktm pokes petan ^ [15:51:34] can we get svn installed on all exec nodes? [15:51:52] legoktm yes, but what for? :P [15:51:54] just being curious [15:52:01] legoktm: hmm [15:52:02] legoktm: :D [15:52:59] petan: i wrote a script that automatically runs unittests after new pywikibot commits [15:53:03] ok [15:53:11] so it has to run svn up to update, then svn info to get the revision # [15:55:33] !log tools petrb: installed subversion to exec nodes 4 legoktm [15:55:35] Logged the message, Master [15:55:44] :D [15:55:46] ty [15:55:57] legoktm: :D [15:56:00] should work now [15:56:07] legoktm: also I like sh.svn('up') better, for some reason [15:56:52] petan: i also got it to send an message to #pywikipediabot via wm-bot :D [15:57:04] aha great :P [15:59:59] * legoktm grumbles about not enough memory [16:00:10] legoktm: did it get killed? [16:01:02] actually everything worked except sending the message via IRC> [16:01:22] sh.nc('bots-labs', '64834', _bg=True, _in="#pywikipediabot "+msg) [16:01:58] can somebody please help me to get a tool running on tool labs? [16:02:19] Pyfisch: That's what we're here for. What problem are you running into? [16:02:44] Coren: getting an internal server error. [16:02:58] Pyfisch: can you run it fine via shell? [16:03:08] if its python, try enabling cgitb [16:03:20] yes it is python [16:03:40] brb [16:03:58] i have enabled cgitb but I only get internal server error [16:04:14] Pyfisch: What is the name of your tool? [16:04:39] it is "wdrdr" [16:04:49] the file is in cgi-bin [16:05:16] and it is (not) accessiable over http://tools.wmflabs.org/wdrdr/cgi-bin/index.cgi?item=q1 [16:05:36] Pyfisch: The script needs to be owned by your tool, not you. :-) [16:05:55]