[00:00:04] It fact, it might be a good idea to make bots into that staging area once tools- is ready. [00:00:19] petan: I'll install mono for you so you can play with it tomorrow. :-) [00:00:30] that might be possible needs more talk [00:00:36] ok [00:07:25] petan: on bsql01 the SELECT user FROM mysql.user; list is rather interesting.. [10:33:04] @notify addaway [10:33:05] This user is now online in #huggle so I will let you know when they show some activity (talk etc) [10:33:15] @notify ceradon [10:33:15] I will notify you, when I see ceradon around here [10:36:01] @seenrx mad [10:36:01] petan: Last time I saw [1]madhatter they were leaving the channel #wikimedia-tech at 10/12/2012 8:34:34 PM (154.14:01:27.1801400 ago) (multiple results were found: Ravingmad_, komaduan, madcaplaughs, ahmad, madcaplaughs_ and 7 more results) [10:36:10] @seenrx madman] [10:36:10] petan: I have never seen madman] [10:36:13] @seenrx madman [10:36:13] petan: I have never seen madman [10:36:25] @seenrx MadMan [10:36:26] petan: I have never seen MadMan [10:36:38] wtf is his nick [13:26:54] addaway [13:27:01] which problem [13:27:03] on mysql? [13:27:12] ? [13:27:19] you sent me something with user [13:27:23] mysql.users [13:27:24] was sorted, someone was using the wrong password ;p [13:27:24] table [13:27:27] oh! [13:27:29] ah [13:27:41] interesting how there is a 'test' user and mutiple users with the same name..? [13:27:47] who? [13:27:57] only root has multiple that because they have different hosts [13:56:01] !log bots petrb: temporarily changed some parameters so that load get distributed better [13:56:06] Logged the message, Master [14:00:11] [bz] (NEW - created by: Chris McMahon, priority: High - normal) [Bug 46166] Add automated browser tests to beta labs (tracking) - https://bugzilla.wikimedia.org/show_bug.cgi?id=46166 [15:07:25] petan: Could you add me to the bots-project? [17:13:42] @notify Jan_Luca [17:13:42] I will notify you, when I see Jan_Luca around here [17:28:53] Change on 12mediawiki a page Wikimedia Labs/Tool Labs/Notepad was modified, changed by MPelletier (WMF) link https://www.mediawiki.org/w/index.php?diff=660504 edit summary: [+27] Eeew! Mono! [17:33:58] petan: IIRC Jan Luca is "jan" on Labs. [17:34:11] ok [17:34:13] ty [17:34:33] @labs-user Jan [17:34:33] Jan is member of 5 projects: Bastion, Centralauth, Gerrit, Tools, Wikiversity-sandbox, [17:48:46] @labs-user Krenair [17:48:46] Krenair is member of 2 projects: Bastion, Openstack, [17:50:12] Trailing commas, [17:55:03] scfc_de fix it then [17:55:27] :P [18:03:05] petan. where's the source? [18:03:14] Type @commands for list of commands. This bot is running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.10.6.8 source code licensed under GPL and located at https://github.com/benapetr/wikimedia-bot [18:05:48] !logs [18:05:48] logs http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-labs [18:05:50] hey Jan_Luca [18:05:59] petan: Hi [18:06:46] !log bots petrb: +Jan to bots [18:06:47] Logged the message, Master [18:07:09] petan: Thank you! [18:07:13] yw [18:14:27] petan: Does bots use SGE or the the successor Open Grid Scheduler? [18:14:43] ogs [18:16:57] petan: C#? No, I don't think so :-). [18:17:07] Didn't you use the standard gridengine-* packages? [18:17:26] yes [18:17:30] is that sge? [18:18:41] petan: OGS/GE is based on Sun Grid Engine [18:18:43] I think, Coren had something written about that on his TODO page. Moment. [18:21:59] petan: I get this error: ERROR 1370 (42000): execute command denied to user ''@'%' for routine 'system.create_db' [18:22:30] Jan_Luca ok sec [18:22:48] Jan_Luca can you check if there is .my.cnf in /mnt/secure/jan [18:22:59] petan: (I'll never get my head around the information split between meta, mw.org and wikitech - anyway ...) https://www.mediawiki.org/wiki/Wikimedia_Labs/Tool_Labs/TODO: "Scheduling and job management: Open Grid Scheduler[1] is a current candidate [...] Ubuntu currently packages the immediate predecessor, which is user-side compatible. [...]" [18:23:26] aha [18:23:32] but we are using wikimedia repo [18:23:35] who knows what is there [18:23:35] petan: There is no dir /mnt/secure/jan [18:23:42] Jan_Luca ok let me fix it [18:24:00] I don't :-). I think the differences are probably small, anyway. Gotta go. [18:24:03] Jan_Luca that script which create access to db is being ran every hour [18:24:22] petan: OK [18:24:35] will run it now k [18:24:46] petan: Thank you [18:30:09] Jan_Luca try now [18:30:29] petan: It works. Thank you very much! [18:30:47] yw ;) [18:34:28] petan: How I can create a new dir in /data/project? [18:34:44] Jan_Luca that needs to get fixed too, sec [18:35:17] try now [18:35:51] mhm :/ [18:35:55] petan: I get Permission denied [18:35:57] doesn't seem to work [18:36:08] ok, I will create it for now but this need fix [18:36:12] what folder you need? [18:36:17] wmlinky [18:37:14] or try to create it inside [18:37:16] userdata [18:37:24] /data/project/userdata [18:37:56] that works [18:38:01] cool [18:38:55] !log bots petrb: /data/project/ is a mess we need to clean it up [18:38:59] Logged the message, Master [18:42:58] !log bots petrb: fixed 1 security bug and 2 other bugs in scripts/mysql_backup.sh [18:43:01] Logged the message, Master [18:43:21] !log bots petrb: btw when you create these things log it please and document them [18:43:23] Logged the message, Master [21:02:56] Change on 12mediawiki a page Wikimedia Labs/Tool Labs/TODO was modified, changed by 69.70.195.70 link https://www.mediawiki.org/w/index.php?diff=660554 edit summary: [+191] Updates [21:03:34] petan: BTW, mono should be usable now on tools- [21:03:47] petan: The cluster nodes only have runtime, though. [21:04:26] ? [21:04:33] is it mono-complete? [21:04:47] petan: mono-runtime on the nodes, mono-complete on -login [21:04:58] I don't need mono on login at all [21:05:02] I only need it on nodes [21:05:23] petan: You don't but other maintainers may well need to have the dev tools if they are going to maintain .net apps. :-) [21:08:35] Change on 12mediawiki a page Wikimedia Labs/Tool Labs/TODO was modified, changed by 69.70.195.70 link https://www.mediawiki.org/w/index.php?diff=660555 edit summary: [+109] Also mail [21:25:27] * Damianz wises people would login... ips are so anti-colab [21:36:54] * Damianz pats Ryan_Lane [21:42:01] * Ryan_Lane yawns [21:42:07] I started writing that email yesterday [21:56:06] lol [21:56:44] It's interesting because it breaks the puppet mentality but is a huge blocker to actually using 'clustered computing'... which is part of the reason bots is all single hosts of stuff currently [21:58:16] what's a blocker? [21:59:28] [bz] (NEW - created by: Antoine "hashar" Musso, priority: Normal - normal) [Bug 36422] easily reload all apaches - https://bugzilla.wikimedia.org/show_bug.cgi?id=36422 [21:59:56] not having project wide 'system' users (ok, so you can get around it now by just making a full user for it... but that sucks and is more risky sec wise). [22:00:05] yeah [22:00:29] I think this is a good solution to the problem [22:03:48] * Damianz wonders if hasher wants to make that happen [22:15:07] Damianz: wants to make what happen? salt? [22:15:11] I just responded to that [22:15:19] I think what I'm proposing is doable [22:15:25] mhm [22:15:36] without upstream features [22:15:42] runners make everything easier :) [22:15:45] I'll read your response as I know what I'd like to do [22:16:15] hm. I'd still need to make the peer_run configuration, though [22:16:17] and that kind of sucks [22:16:23] since you need to restart the master for that [22:17:07] fuck [22:17:10] salt docs are ugly [22:17:20] I guess I could just allow all instances to call the runner [22:17:33] and then apply authn in the runner [22:17:44] err [22:17:45] authz [22:18:56] Ahhh I never knew up/down was a runner [22:19:01] yep [22:22:58] btw when addaway is back tell him that multiple root's on mysql is ok :P [22:23:12] it's default in ubuntu... [22:23:57] * Damianz goes back to hugging zuul into shape [22:24:13] lol [22:24:23] * petan goes to sleep now [22:24:43] Ryan_Lane I am wondering if we really need a dedicated server for mysql [22:24:51] maybe we would be fine with dedicated storage [22:24:57] so that IO wouldn't be so bad [22:25:11] eh? [22:25:16] because performance of boxes on labs is quite good [22:25:18] what do you mean? [22:25:21] only problem is with storage [22:25:26] that's why all these things are slow [22:25:29] IO is bad [22:25:39] CPU and RAM seems fine to me [22:25:42] IO through virtualization is bad [22:25:48] and IO on a shared host is bad [22:25:55] well, but you can mount external storage directly to virtual host [22:26:04] which would make it fast [22:26:09] not really [22:26:18] there's a lot of IO overhead with virtualization [22:26:21] I have a vps and it has incredibly fast storage [22:26:33] I don't know how they do that but it works [22:26:40] the storage we have is 8 fast SAS disks [22:26:41] per host [22:26:53] maybe the way how it's connected? [22:27:08] doubtful, it's using virtio [22:27:13] is it directly mounted to instance or it's through some other host [22:27:18] that said, it's qcow2 [22:27:20] not raw disks [22:27:25] ah [22:27:27] that slows things down quite a bit [22:27:39] it's directly mounted to the instance [22:27:58] we overprovision our storage quite a lot [22:28:01] maybe raw disk directly mounted to instance would be just as fast as whole dedicated box [22:28:05] we can't do raw disk [22:28:19] well, I meant only for sql servers, not for all instances [22:28:37] meh. why bother? we can just have dedicated boxes [22:28:48] it sounds cheaper to me :P [22:28:55] storage only vs whole server [22:29:04] it's not much cheaper [22:29:38] anyway, once people *really* start hitting them with queries, I think you'd change your mind ;) [22:29:42] [bz] (REOPENED - created by: MZMcBride, priority: Normal - normal) [Bug 36885] wmflabs.org does not resolve - https://bugzilla.wikimedia.org/show_bug.cgi?id=36885 [22:29:55] Just buy a blade chassis, stick like 256gb of ram in there, some ssds and then I'll use your mysql server :D [22:29:56] I think addshore is hitting them pretty much now [22:30:13] petan: not like the toolserver folks hit them [22:30:28] right, but they are using it in a different way [22:30:38] they are querying production clone [22:30:42] not own databases [22:31:31] I don't think in bots / tools project there will be many bots that would be more expensive than addshore's [22:31:55] basically he's eating over 90% of all resources now [22:33:00] * Ryan_Lane shrugs [22:33:22] dedicated boxes will perform better and won't be affected by other instances [22:44:52] Ryan_Lane: can we discuss , pls ? [23:53:54] does anyone know why https doesn't work on http://bots.wmflabs.org/ ? [23:54:13] that is, https://bots.wmflabs.org/ [23:59:04] they don't have a server running on 443? [23:59:50] still, I don't think they would have access to the certificate...