[01:46:47] Coren, you there? [01:53:34] Theo|away, you about? [01:53:42] Hey [01:53:47] Sort of :p [01:54:09] That tool for the redlink will be stationed in the cyberbot project. [01:54:32] Okay [01:54:39] Can you grant me access when you get a chance? [01:54:42] No hurry though [01:55:29] I'm still trying to figure out my concerns of then sealing the config files so you can't access them. [01:55:57] Why bother? [01:56:35] If you get granted access to the cyberbot project, you'll have to my account everywhere, as well as cyberbots'. [01:56:50] You use the same password on your bots and your main account? [01:56:52] And given what happened to Riley... [01:56:54] That doesn't sound secure [01:57:25] They're different. But I have my main account config file in there too, for test wiki. [01:57:31] which I am a sysop on. [01:57:44] Oh, okay [01:58:04] I strongly advise against storing your main account password in a text file, tbh [01:58:39] Cyberpower678: Can't you just create another account for testing? [01:59:36] scfc_de, I've got a butt load of accounts now, when I tested Peachy and forgot to stop the script. :p [02:00:12] But I test with my main account on test.wiki [02:00:19] since it has sysop. [02:00:53] Just get a test sysop account [02:01:18] This all sounds like a disaster waiting to happen [02:01:31] It's worked for me so far. [02:01:45] The config folder is locked to outsiders. [02:01:58] Nobody can access them. [02:02:17] Except for wmf roots [02:02:21] Just saying... [02:02:28] Well yea... [02:02:37] I just trust Coren though. [02:02:39] You shouldn't consider contents stored on a remote server that you have no control over secure [02:02:49] And everyone else should you make a mistake. [02:04:36] Theo|away, I'm going to setup a repo. The interface to manage the redlink script will have a deploy option for you and me specifically. [02:04:52] Sure, whatever floats your boat [02:05:37] So you will still have access to the codes and that way, we can keep them updated in one place without conflict. [02:05:39] :) [02:13:31] Gotta love a crappy network. [02:13:37] Stupid LTE. [02:51:25] My Poor CPU [02:55:06] Betacommand, what happened? [02:55:34] Cyberpower678: 100% usage for the last 18 hours [02:55:43] And? [02:55:57] Cyberpower678: thats a lot of work [02:56:07] 9_9 [02:56:24] I've done that to my computer a lot. [02:56:53] Cyberpower678: continuous max? [02:56:54] If you got a good fan to vent the heat, it should be fine/ [02:57:25] Betacommand, I batch processed 5 seasons of HD videos in a row. [02:57:49] Cyberpower678: Im doing just about the same :P [02:58:08] my media collection is massive [02:58:18] And I collect Pokemon and so when I was collected them all, it was a total of 12 seasons I had my computer work on. [02:58:36] 3TB baby. [02:58:46] Cyberpower678: lol [02:58:50] I collect only the best quality material. [02:58:59] Im in the 12TB range [02:59:05] RESPECT. [02:59:21] I don't have that much time unfortunately. :p [03:00:02] Ive got somewhere around 5 years of continuous media streaming [03:00:16] X!'s edit counter on toolserver Executed in 23.53 second(s). [03:00:54] On labs, it Executed in 1.39 second(s). [03:01:03] Cyberpower678: no deleted info [03:01:11] yet. [03:01:20] SSDs, and a lot less server stress [03:01:27] But that's still a huge difference. [03:01:48] * Cyberpower678 is in love with SSDs <3 [03:02:17] He has 768GB of that kind of storage on his laptop. [03:02:20] Cyberpower678: Im in the process of migrating three of my home servers into one, with an SSD [03:02:46] Windows 7 boots in 1.3 seconds. [03:02:53] Download server, media server, and web server [03:03:16] I've got a really tiny, basic server. [05:51:15] Warning: There is 1 user waiting for shell: Euvl (waiting 0 minutes) [06:04:40] Warning: There is 1 user waiting for shell: Euvl (waiting 13 minutes) [06:18:09] Warning: There is 1 user waiting for shell: Euvl (waiting 26 minutes) [06:25:05] !rq Euvl [06:25:05] https://wikitech.wikimedia.org/wiki/Shell_Request/Euvl?action=edit https://wikitech.wikimedia.org/wiki/User_talk:Euvl?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/Euvl [06:49:05] scfc_de: ping [06:50:59] legoktm: Pong. [06:51:28] so uh, i just saw [06:51:29] You have new mail in /var/mail/legoktm [06:51:39] On Tools? [06:51:43] yeah [06:51:46] and theyre all cron messages from local-dbreps [06:51:49] and then [06:51:53] N 6 Mail Delivery Syst Thu May 23 00:30 50/2178 Mail delivery failed: returning me [06:51:59] a bunch of those [06:52:05] Yep. Everything's normal :-). [06:52:18] lol, ok [06:52:21] Susan hasn't created a Tools account, and jsub is very noisy at the moment. [06:52:27] :P [06:52:34] ... account *yet* ... [06:52:50] if i create a .forward file in my user home, will it work? [06:52:55] Yes. [06:54:16] legoktm: BTW, on Toolserver popular_items.py fails with "IndexError: tuple index out of range". [06:54:43] ugh :/ [06:54:47] ill look into that tomorrow [06:54:54] legoktm: legoktm: BTW, on Toolserver popular_items.py fails with "IndexError: [06:54:55] tuple index out of range". [06:54] [06:54:55] ERC> legoktm: Ooops. I meant: Traceback (most recent call last): File "/home/project/d/b/r/dbreps/bin/popular_items.py", line 91, in main() File "/home/project/d/b/r/dbreps/bin/popular_items.py", line 83, in main report = mk_report(db) File "/home/project/d/b/r/dbreps/bin/popular_items.py", line 64, in mk_report text+= table_row.format(qid, count) [06:56:07] thanks [06:56:22] do you know of a quick way i can clear all my mail on labs? i doubt any of it is relevant [06:57:13] "mail", "d *", "q" [06:57:59] yay thanks :D [07:14:35] Coren is probably sleeping? (3:13L?) [07:15:14] Yesterday I could connect to p50380g50450__checkwiki_p as my Tools user, now (after the restart?) I cannot. [07:15:19] hi [07:16:33] petan: https://bugzilla.wikimedia.org/show_bug.cgi?id=49069 [07:17:12] Do you know anything about that? [07:17:21] scfc_de: it's fixed resolved [07:17:34] I don't know anything about it :o [07:18:10] k [07:33:09] Coren ping [07:33:18] Coren https://bugzilla.wikimedia.org/show_bug.cgi?id=49058 [07:33:26] hello [07:33:30] hi [07:37:09] scfc_de that bug is where? [07:37:13] replicas? [07:37:15] or db [07:37:27] I can't see your u* account on -db [07:37:33] there are no such users [07:37:45] and every db has %_p to read [07:37:48] on everyone [07:50:39] petan: why's there no `user_daily_contribs` table in the replica [07:50:51] I have no idea [07:51:03] I have no access to replicas whatsoever neither I set it up [07:51:10] you need to ask binasher or Coren [07:51:14] k [08:05:36] petan: I should keep notes :-). I was very certain that in the past I connected to tools-db as u2267 and created a database u2267__something, but I can't find any evidence for that :-(. [08:11:56] that is weird... there is no such user [08:21:41] Warning: There is 1 user waiting for shell: Gitsan (waiting 0 minutes) [08:23:23] !rq Gitsan [08:23:23] https://wikitech.wikimedia.org/wiki/Shell_Request/Gitsan?action=edit https://wikitech.wikimedia.org/wiki/User_talk:Gitsan?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/Gitsan [08:24:55] petan: Probably a memory error in my brain :-). [08:32:08] @search status [08:32:08] Results (Found 3): monitor, osm-bug, stats, [08:32:12] !stats [08:32:12] http://tools.wmflabs.org/?status [08:33:05] jimmyxu: see http://www.gossamer-threads.com/lists/wiki/wikitech/347292 that extension might have been disabled [08:35:24] drdee: the replica on TS still has fresh data. However I'm not sure if this is in the progress of being phased out [08:35:41] ok [08:35:46] can;t help you any further [08:37:32] !log tools petrb: installing python-memcache on exec nodes [08:37:34] Logged the message, Master [09:50:06] petan: tools-exec-cg ? :D [09:50:12] no idea [09:50:18] no entry in SAL, Coren is quiet [09:50:28] probably some secret thing [09:50:34] mhhhm, it looks like its a grid node :P [09:50:35] but it seems quite unconfigured to me [09:50:47] because it's missing most of packages we have on other nodes :/ [09:51:27] what about tools-mc ? :P [09:51:31] oh memcache [09:52:50] lol this is best picture of me I found so far https://commons.wikimedia.org/wiki/File:Erik_Hackathon.jpeg [09:53:16] hah [09:53:16] i hve seen a better one but I dont know where :P [11:25:17] Coren, metawiki_p is also broken. [11:56:06] Warning: There is 1 user waiting for shell: Aadithvmenon (waiting 0 minutes) [12:09:31] Warning: There is 1 user waiting for shell: Aadithvmenon (waiting 13 minutes) [12:23:01] Warning: There is 1 user waiting for shell: Aadithvmenon (waiting 26 minutes) [12:34:13] What is lab's IP? [12:36:29] Warning: There is 1 user waiting for shell: Aadithvmenon (waiting 40 minutes) [12:50:02] Warning: There is 1 user waiting for shell: Aadithvmenon (waiting 53 minutes) [12:52:53] Cyberpower678 hold on [12:53:08] ? [12:53:38] !ip is 208.80.153.128/25 (labs) [12:53:38] Key was added [12:53:41] this one [12:58:38] !rq Aadithvmenon [12:58:39] https://wikitech.wikimedia.org/wiki/Shell_Request/Aadithvmenon?action=edit https://wikitech.wikimedia.org/wiki/User_talk:Aadithvmenon?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/Aadithvmenon [13:13:03] Good $moment, labs. [13:13:19] Coren what is that -cg thing [13:20:48] Coren, metawiki_p is broken too. [13:20:58] Define "broken"? [13:21:05] Missing tables. [13:21:57] Huh. I wouldn't have expected that one to be atypical (the only DBs that had been problematic to date are the ones that had unique extensions or schema changes) [13:22:05] * Coren will check. [13:22:58] Coren, Table 'metawiki_p.revision_userindex' doesn't exist [13:26:19] How odd! metawiki has missing stuff in some of its tables! [13:26:52] ... that can't be right. [13:28:26] !log tools petrb: installing csh on cluster [13:28:28] Logged the message, Master [13:29:53] Aha! [13:30:12] metawiki is on s7; that's an old (broken) version that dates a few years before it was moved. [13:30:13] Coren what is that -cg thing [13:30:15] :P [13:30:18] catgraph [13:30:22] hmm [13:30:30] I can't submit jobs there [13:30:38] hence the package check can't finish... [13:30:40] That's normal, it has a dedicated queue. [13:30:49] I know, but I can't submit even to that one [13:30:51] :P [13:30:54] I mean using local-admin user [13:31:10] there is a check that verify if exec nodes are consistent [13:31:25] but if this one is somehow special maybe it doesn't need to be [13:31:35] That queue is limited to the catgraph tool. :-) It doesn't need to be consistent. [13:31:38] ok [13:31:39] Incidentally... [13:31:56] Coren, what's so wrong with S7? [13:32:23] Some of the stuff you've made consistent were inconsistencies caused by the original image; fixing them is only useful if you puppetize the change. :-) [13:32:31] Cyberpower678: It's not available yet is what. :-) [13:32:43] Huh? [13:35:15] Coren, what's up with Wikidata? [13:35:16] Caution: Replication lag is high, changes newer than 1 weeks, 3 days, 20 hours, 43 minutes, 25 seconds may not be shown. [13:36:21] Cyberpower678: I wouldn't know -- someone more in tune with prod might be more useful there. [13:37:00] ok. [13:37:09] Cyberpower678: There; you shouldn't get broken tables error with metawiki_p anymore. Because the wrong metawiki database has been removed from the list of aliases. :-) [13:38:07] Guessing from the missing columns, that "metawiki" database was a remnant dating sometime before 1.18 [13:38:33] revision didn't exist before 1.18? [13:39:44] No, there were missing columns in some tables that were added in 1.18, so that old database dated from at least before that. [13:48:14] Coren, so now you deleted the entire thing? [13:48:39] My error is getting flooded now. [13:48:42] :/ [13:49:03] It was never supposed to be there in the first place; my view creation system got confused by the old database that should have been dropped years ago. [13:49:52] not yet available: arwiki cawiki eswiki fawiki frwiktionary hewiki huwiki kowiki metawiki rowiki ukwiki viwiki centralauth [13:50:28] metawiki and centralauth. O_o [13:50:45] The most important of 'em all. :p [13:50:52] centralauth is the reason for the holdup. The security review should be done this week. [13:51:16] security review? [13:51:22] ... but I admit this is the very first time I hear *anyone* call meta "important", let alone "the most important". :-) [13:51:53] Coren, it's where all of the central maintenance happens. [13:52:01] :o [13:52:10] I was always thinking it's a place where nothing ever happens [13:52:11] Cyberpower678: I am aware. I wasn't saying it was unimportant, just that it gets no respect. :-) [13:52:36] I was mistaking meta with sandbox all these years... :/ [13:54:19] petan, say what now? [13:54:26] Seriously? [13:54:31] XXXXXXDDDDDDD [14:11:46] Warning: There is 1 user waiting for shell: Gryllida (waiting 0 minutes) [14:17:41] hi [14:18:33] hi [14:19:26] https://wikitech.wikimedia.org/wiki/Category:Shell_Access_Requests is huge, how do I provide reasons for shell account request on my page in that category? [14:19:37] (and how long does it usually take to get processed?) [14:19:47] it takes like 10 minutes or less [14:19:51] @requests [14:19:51] Warning: There is 1 user waiting for shell: Gryllida (waiting 8 minutes) [14:19:54] is that you? [14:20:01] !rq Gryllida [14:20:01] https://wikitech.wikimedia.org/wiki/Shell_Request/Gryllida?action=edit https://wikitech.wikimedia.org/wiki/User_talk:Gryllida?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/Gryllida [14:20:32] yes [14:20:37] done [14:20:38] :) [14:21:03] still made it within 10 minutes :P [14:21:05] according to bot [14:21:24] ok, thanks; do you work for wikimedia or do volunteers process those requests? [14:21:33] volunteer [14:21:53] we have a lot of monkeys for that [14:22:36] eg. these requests are processed by both employees and some volunteers [14:22:38] do I need to do any paperwork before logging in and running whatever wikimedia-related scripts I'd like? [14:22:59] not that I know of, but shell access doesn't give you much [14:23:07] you need to be member of some project first [14:23:26] what does project membership give? [14:23:41] shell itself just allow you to login to bastion hosts, which are just a login servers to other instances, which are parts of projects [14:23:42] feel free to stab me with a well sized manual if there is one :) [14:23:56] !documentation [14:23:56] https://labsconsole.wikimedia.org/wiki/Help:Contents [14:24:00] yes here is one [14:24:17] thanks for the keywords and the link, I'll take a look [14:24:23] I don't know what you want to do on labs, but I have a feeling you will want to become member of tools project [14:24:27] !toolsdocs [14:24:28] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help [14:24:42] which is a biggest project where most of wikimedia related tools and bots live [14:24:53] Can I become a member of the tools project too? [14:24:58] other projects are quite specific, mostly for stuff that doesn't fit in that [14:25:00] bawolff: sure [14:25:07] bawolff what is your labs username? [14:25:15] Brian_Wolff [14:25:33] * bawolff just wants something that has php, apache and is web accessible... [14:25:42] bawolff are you sure? :P [14:26:07] @labs-user Brian Wolff [14:26:07] Brian Wolff is member of 1 projects: Bastion, [14:26:07] petan: My username on wikitech is Brian_Wolff, I think my shell name is "bawolff" [14:26:11] indeed [14:26:13] you are correct [14:26:22] * bawolff didn't create my account [14:26:45] Successfully added Brian Wolff to tools. o/ [14:27:03] :) [14:28:02] lol, I like the ascii logo in the motd [14:28:07] hehe [14:30:35] petan, toolserver has a powered by toolserver button. [14:30:43] Does labs have one of those? [14:31:00] I logged in and see a "*** System restart required ***" among other things, is that normal? [14:31:41] and yes please, add me to tools project too [14:32:02] gry: yes that is normal... puppet install new kernels no matter if it's needed or not, but fortunatelly never reboot server [14:32:25] !tr Gryllida [14:32:25] request page: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Access_Request/Gryllida?action=edit talk page: https://wikitech.wikimedia.org/wiki/User_talk:Gryllida?action=edit§ion=new&preload=Template:ToolsGranted link: https://wikitech.wikimedia.org/w/index.php?title=Special:NovaProject&action=addmember&projectname=tools [14:32:40] done [14:32:52] I sent you another message which contains even more links :P [14:33:36] petan, toolserver has a powered by toolserver button. [14:33:36] Does labs have one of those? [14:33:42] Cyberpower678 idk :o [14:33:46] where should it be? [14:33:54] we are powered by unicorns [14:34:19] last time I added unicorn to main page I got slapped for breaking some copyrights or whatever [14:37:21] petan: people didn't want "religious" things used [14:37:30] aha [14:37:34] I had no idea it's religious :D [14:37:43] it was just a cute unicorn [14:38:27] it's similar to flying spaghetti monster, apparently [14:38:41] petan: for example, I want to run an irc bot for mediawiki that was not at toolserver (it was self-hosted before). where do I run it and would I need to request a project, an instance or something else for it in addition to what I already have? [14:38:48] petan, You're saying you don't have a powered by labs button or a powered by labs powered by unicorns button? [14:39:19] gry: you would run it on tools project application grid, you don't need to request anything [14:39:36] Cyberpower678 not that I know of [14:40:08] petan: ok. what is an 'application grid'? [14:40:10] Ryan_Lane but unicorn is still a logo of labs or not? [14:40:11] petan, then I'll make one. [14:40:16] !toolsdocs [14:40:16] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help [14:40:20] petan: that specific one? no [14:40:28] but a unicorn? yes :) [14:40:45] gry: there is a good explanation, but if you want I can explain it myself which would be far worse than what is there [14:40:47] petan: http://en.wikipedia.org/wiki/Invisible_Pink_Unicorn [14:40:54] ok [14:41:05] I think we made it green :D [14:41:12] heh. same same [14:41:17] then it's invisible and green [14:41:37] ok then have someone create unique huge cute unicorn :D [14:41:43] petan: thanks, the link has a 'grid engine' section that'd keep me busy comprehending for a longer while [14:42:05] that is it [14:45:50] Ryan_Lane: puppet keep putting that evil popularity contest back, is it possible to get rid of it? :P [14:46:07] popularity contest? [14:46:11] the ubuntu package? [14:46:12] that is some package [14:46:13] yes [14:46:19] I can't imagine that puppet is doing that [14:46:23] I can [14:46:29] it probably install ubuntu-standard [14:46:35] which is virtual package depending on it [14:46:54] everytime I remove it, puppet come and try to put it back [14:47:05] bad thing :( [14:47:20] i don't care but this package produces lot of errors in mail [14:47:28] that is annoying [14:49:29] !log tools petrb: installing sbt in order to fix b48859 [14:49:31] Logged the message, Master [14:50:10] wheeeeeeee Coren [14:50:23] we are apparently missing java runtime... everywhere! o.O [14:50:57] petan: No we're not. [14:51:11] marc@tools-exec-01:~$ java -version [14:51:11] java version "1.7.0_21" [14:51:17] it say: depends on java2-runtime; however: [14:51:18] Package java2-runtime is not installed. [14:51:54] ok let me try to force install it... [14:52:00] maybe it will work then [14:52:17] No, no, no. [14:52:24] What are you trying to install? [14:52:33] https://bugzilla.wikimedia.org/show_bug.cgi?id=48859 [14:52:35] sbt [14:52:58] it is already installed on -dev [14:53:00] it seems to work [14:53:40] * Coren grumbles. [14:54:20] I don't particularily want runtime to be installed instead of runtime-headless; the former creates bazillion dependencies. [14:55:17] I /definitely/ want only headless on exec nodes. [14:55:32] Coren I don't think that maintainer of package did pick correct dependencies, so maybe this will work with headless as well [14:55:53] Then you really should rebuild the .deb with -headless as a dependency before you install this. [14:55:57] they were likely lazy so they just used whole runtime as dependency instead of picking the required sw only [14:56:02] Perhaps even point out the bug upstream. [14:56:12] maybe I should first check if it really works with headless first, and then rebuild it :P [14:56:20] Yes, checking would be good. :-) [15:09:43] Coren: https://bugzilla.wikimedia.org/show_bug.cgi?id=49104 pliz [15:09:45] :D [15:10:00] I don't think it's actually hard to implement but I don't know where the sources of interface are [15:30:44] what's tools-dev? [15:30:46] seen in motd [15:33:24] liangent: it's like tools-login [15:33:36] except you use it for stuff that may be very resource intensive on that instance [15:33:48] so, if you need to compile something giant, use tools-dev [15:34:31] Ryan_Lane: i really need to login to the catgraph vm without jumping through hoops :) [15:34:51] what do you mean jumping through hoops? [15:34:58] using ssh. [15:35:01] you can login directly to it via proxycommand [15:35:27] apparently it is possible - Kai_WMDE just told me he can log in to this normally: https://wikitech.wikimedia.org/wiki/Nova_Resource:Mediawiki-custom-de [15:35:29] that isn't documented for tools project because it's not generally necessary to use [15:35:59] proxycommand. hmmm. coren told me i have to 'become catgraph' first, then schedule an interactive job on the node [15:36:02] is your catgraph stuff in tools? [15:36:05] ah [15:36:08] wait. [15:36:14] you need your tool to ssh into something? [15:36:15] that's a bit complicated. i would like to just ssh there directly [15:36:34] no, the tool doesn't use ssh. but i need to. [15:36:59] well, the tool runs on a grid and is scheduled [15:37:06] JohannesK_WMDE: As I've told you previously, I'd be happy to support your workflow if I understood its use case. :-) [15:37:07] it's not simple to just ssh into it [15:37:15] JohannesK_WMDE: if I understand correctly, you want to connect, and be logged in as catgraph user, instead of having to become catgraph afterwards? [15:37:32] valhallasw: no, why [15:38:23] Coren: i want to login there, start top, vmstat, or whatever else to minitor things, for example. and i run graphserv in a screen, which i want to resume from time to time to interact with it. [15:39:09] that project "mediawiki-custom-de" is parallel to tools labs, by the way [15:39:18] not residing in tools labs [15:39:19] Ah, this is where we part ways. You should not run tools in a screen; that provides no redundancy, no monitoring and no scheduling. :-) [15:39:20] and this should be possible without 1) sshing to -login, 2) 'become'ing catgraph, 3) scheduling an interactive job. i would like to just ssh there. [15:39:50] JohannesK_WMDE: Perhaps your use case model is more apropriate to a project of your own rather than Tool Labs, then. [15:40:09] Providing other projects with access the the replicas is in the short term plans. [15:40:11] In theory, you should also be able to schedule an interactive session using SGE [15:40:38] valhallasw: Yes, but interactive sessions require you to be... interactive. :-) And are - more importantly - not restartable. [15:40:55] ooohhh, you guys just want access to the replicas from that project? [15:41:10] Coren: it should be possible to start a screen from a non-interactive terminal, too, I'd think [15:41:11] well, yeah, i would of course start it from /etc/init.d if i had root. but at this point, i think a screen is sufficient - for one thing, i need to interact with it, on the other hand, i work on the code often and have to recompile & restart a lot, so i don't need to make some init.d script (yet) i think. [15:41:49] JohannesK_WMDE: You're missing my point. You wouldn't start it from init.d /either/. On tool labs, you hand your job to the scheduler and let /it/ start and keep it up. [15:41:56] for example, i recently added the find-cycles command. and that needed a recompile+restart [15:42:29] honestly it may be saner for catgraph to not live in tools [15:42:38] but rather be a resource that tools can use [15:42:39] Ryan_Lane: it might [15:42:42]