[01:09:14] $ become static [01:09:14] sudo: sorry, a password is required to run sudo [01:09:16] ? [01:26:36] Krinkle: log out and in again? [01:26:58] no difference [03:00:30] (03PS1) 10Legoktm: Add .gitreview [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/147014 [03:00:46] (03CR) 10Legoktm: [C: 032 V: 032] Add .gitreview [labs/tools/extdist] - 10https://gerrit.wikimedia.org/r/147014 (owner: 10Legoktm) [10:09:52] Krinkle: "getent group tools.static" => "tools.static:*:51367:ireas" => No krinkle there :-). [10:10:09] Yeah, I found that out later. [10:10:13] But the error is rather confusing [10:10:37] I know plain sudo has the same weird error, but 'become' may want to abstract that [10:17:52] That's a good idea (though probably not many users will hit that path). [11:00:38] 3Wikimedia Labs / 3tools: become should complain more clearly if the user is not (in this session) a member of the tool - 10https://bugzilla.wikimedia.org/68156 (10Tim Landscheidt) 3NEW p:3Unprio s:3enhanc a:3Tim Landscheidt At the moment, if you try to become a tool that you are not a member of, the... [11:33:57] Sandaru here? [11:39:50] 3Wikimedia Labs / 3tools: become should complain more clearly if the user is not (in this session) a member of the tool - 10https://bugzilla.wikimedia.org/68156#c1 (10Tim Landscheidt) For testing, I just added myself to the coursestats tool /after/ logging into tools-login: | scfc@tools-login:~$ getent grou... [11:56:49] 3Wikimedia Labs / 3deployment-prep (beta): HTML Tags are stripped from extension output - 10https://bugzilla.wikimedia.org/66516 (10physikerwelt) p:5Unprio>3Normal [12:37:58] (03PS1) 10Tim Landscheidt: become: Make more user-friendly [labs/toollabs] - 10https://gerrit.wikimedia.org/r/147096 (https://bugzilla.wikimedia.org/68156) [12:38:22] (03PS2) 10Tim Landscheidt: become: Make more user-friendly [labs/toollabs] - 10https://gerrit.wikimedia.org/r/147096 (https://bugzilla.wikimedia.org/68156) [12:42:20] (03CR) 10coren: [C: 04-1] "Does not handle root; see comment." (031 comment) [labs/toollabs] - 10https://gerrit.wikimedia.org/r/147096 (https://bugzilla.wikimedia.org/68156) (owner: 10Tim Landscheidt) [12:49:38] (03PS3) 10Tim Landscheidt: become: Make more user-friendly [labs/toollabs] - 10https://gerrit.wikimedia.org/r/147096 (https://bugzilla.wikimedia.org/68156) [12:49:48] (03CR) 10Tim Landscheidt: become: Make more user-friendly (031 comment) [labs/toollabs] - 10https://gerrit.wikimedia.org/r/147096 (https://bugzilla.wikimedia.org/68156) (owner: 10Tim Landscheidt) [12:51:09] 3Wikimedia Labs / 3tools: Performance problem on database server s5 using commonswiki - 10https://bugzilla.wikimedia.org/67602#c5 (10Marc A. Pelletier) No further efforts trying to fix federation will be spent as we are upgrading to MariaDB 10 which will provide commons locally and do away with federation en... [13:04:01] (03CR) 10coren: "Just a bit of nitpicking left (use $EUID versus $(id -u))" (031 comment) [labs/toollabs] - 10https://gerrit.wikimedia.org/r/147096 (https://bugzilla.wikimedia.org/68156) (owner: 10Tim Landscheidt) [13:32:17] (03CR) 10Incola: [C: 032 V: 031] add .gitreview [labs/tools/maintgraph] - 10https://gerrit.wikimedia.org/r/146939 (owner: 10Ricordisamoa) [13:33:05] (03CR) 10Incola: [C: 032 V: 031] load the latest minified version of d3 from toollabs:static [labs/tools/maintgraph] - 10https://gerrit.wikimedia.org/r/146941 (owner: 10Ricordisamoa) [14:03:01] Betacommand: available ? [14:05:58] matanya: yeah whats up? [14:08:01] hi Betacommand. i'm tying to do some sql queries to labs db [14:08:19] Ok [14:08:27] how can i gather the information and see how many people in the last 10 days registered and did more than 10 edits to he.wiki [14:08:49] and compare that to May metrics [14:10:01] matanya: do you may stats? [14:11:14] what do you mean Betacommand [14:11:46] matanya: do you already have the stats for may, or do you need to calculate those too? [14:11:54] need them too [14:12:10] matanya: thats going to be harder [14:12:19] why ? [14:13:13] matanya: using non-current values for edit count makes the query harder [14:13:20] 3Wikimedia Labs / 3(other): "Deprecated" error on devwiki/pagemigrationwiki (http://pagemigration.wmflabs.org/wiki/Special:RecentChanges) - 10https://bugzilla.wikimedia.org/66040#c4 (10Southparkfan) 5NEW>3RESO/DUP This bug seems to be a duplicate of bug 65776, closing this one. *** This bug has been mar... [14:13:56] right now I can just query user registration time and user editcount [14:14:04] well, May stats are partially available. the last 10 days are more interesting for me [14:17:44] matanya: select * from user where user_registration > '20140707000000' and user_editcount > 10; [14:18:15] so 31 is the answer [14:18:21] Not exactly 10 days but close enough [14:20:52] Betacommand: Unless I'm mistaken, matanya needs (users who registered in may) and have >=10 edits, not (users who have registered and have >=10 edits) in may [14:21:33] you are right Coren [14:21:43] Coren: that would make it easier then [14:21:57] the idea is to know the impact of the war in israel, and see if the rumors that people forced to stay home make more people sign up to wikipedia and edit [14:22:36] matanya: you can use user_registration > '20140707000000' to set bounds on the data returned [14:23:13] it is seems the answer is no. if for the 10 days of the war 31 people joined and edited more than 10 times, and entire May is 155, so the proportion shows the answer is no. [14:23:55] matanya: that just means that those user have make 10 edits in the last 3 months [14:24:18] Betacommand: that = ? [14:24:21] Betacommand: Yeah, "people who sticked around after registering" [14:24:32] vs 10 edits in 10 days [14:24:57] oh, i see [14:25:00] indeed, so you'd have to join users who registered in may with counts of revisions they made in may [14:25:00] with the numbers from may, it might mean only 2 edits a week [14:25:46] matanya: Besides, the premice sounds fishy; I could see why being stuck at home might increase users, but I should think that being stuck at home *because war* is likely to have people not be willing to concentrate on writing an encyclopedia. [14:25:49] yuvipanda: which is why I said its more complex [14:25:54] right [14:26:45] so if i separate those? last 10 days registered and edited more than 10 times vs registered in May and edited more than 10 time in May [15:04:29] andrewbogott: it is up, but problems with audio / video [15:04:57] 'problems with audio / video' describes every teleconference system ever :( [15:06:10] no, i mean, java issues. not something really broken in the software [15:06:31] you can start a meeting, and ask people to join, and then java starts crying [15:07:01] andrewbogott: not sure exactly when yet, but is was 2am when i was playing with it, so gave up for the night [15:08:31] Sounds like a fair bit of progress for one day of work [15:11:27] use chrome to access gitsi.wmflabs.org [15:11:39] * jitsi.wmflabs.org [15:13:10] andrewbogott: if you don't mind spending like 15 minutes with me to go further, i'd appreciated that [15:13:38] Sure, give me a minute... [15:14:28] (I'm trying to figure out how to have ensure a service is running and also remove the init.d script. Works the first time, subsequent runs the service complains (reasonably) about not being able to find the init script) [15:15:59] andrewbogott: don't use init as provider? e.g. use upstart ? [15:16:33] The idea is to prevent a user from ever starting the service. So I want no service script at all. [15:33:41] IIRC I think matanya's right in that you need to explicitly use "upstart" as the provider. But it's been a long since I looked into that deeper. [16:09:55] jsub -once -continuous -N job.name -mem 6800m /data/project/projname/blah 2> /dev/null [16:09:55] how ti surpress teh creation of out files when using jsub? [16:13:27] -j y -o /dev/null? [16:18:00] gifti: hi, thanks. [16:18:47] ... and, if you do that, don't complain if you don't get error messages why your job couldn't get started, failed, etc. [16:19:34] hm, and with -stderr? [16:20:44] and possibly -quiet [16:22:19] do you mean like: -j y -o /dev/null -stderr /phat/ ? [16:24:05] no, read the documentation [16:27:48] gifti: the doc is crap.... [16:28:01] i always read teh doc bevor asking :P [17:09:49] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c2 (10Andrew Bogott) This patch sets /etc/mailname to contain $::fqdn. Is there ever situation where we'll need more flexibility than that? https://gerrit.wikimedia.org/r/#/c/147150/ [17:10:20] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962 (10Andrew Bogott) 5NEW>3PATC [17:15:20] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c3 (10Tim Landscheidt) I think Tony with his VERP GSoC project might be affected by this. Tony, in your Labs instance, would it do you any harm if /etc/mailname was fixed to something.eqiad... [18:32:20] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c5 (10Andrew Bogott) I could potentially change the patch so that it only modifies the file if it contains exactly 'labs-vmbuilder-precise.eqiad.wmflabs'. I'd prefer the generic solution, t... [18:35:35] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c6 (10Tony Thomas) (In reply to Tim Landscheidt from comment #3) > I think Tony with his VERP GSoC project might be affected by this. Tony, in > your Labs instance, would it do you any harm... [19:01:54] !log deployment-prep possibly breaking labs by cherry picking an apparmor patch that affects mysql https://gerrit.wikimedia.org/r/#/c/147027/ [19:01:56] Logged the message, Master [19:33:52] where there any changes to the database servers on tool labs in the last weeks? the performance decreased significant [19:36:37] and there is some real strange performance behaviour. Example: "SELECT page_id, timestamp FROM dewiki_templatedata_pages WHERE timestamp>=20140716234009 LIMIT 0,10;" -> this one really lasts veeery long (don't know how long, but seems to be hours. If I replace the precise columns ("page_id", "timestamp") with "*" it is very fast (0.1 s). The expected behaviour would be, that the query is faster with given columns... [19:49:01] scfc_de: Could you pls restart enwp10 webservice. It died and Theopolisme's restarter seems to be non-functional [19:50:25] hedonil: Read labs-l; Theopolisme should use bigbrother. [19:50:36] Oh wait. He is. [19:50:49] 2014-07-17 19:21:47 warn: Too many attempts to restart job 'lighttpd-enwp10'; throttling [19:50:52] * hedonil chuckles. "big brother"... [19:51:03] That thing crashes every other hour. [19:52:04] I started it by hand, but he's really gotta take a look at what's wrong with it [19:52:39] Coren: yeah, it ran OOM at 8GB and (mod_cgi.c.590) cgi died, pid: 29105 [19:52:51] but I need it currently to test some things [19:53:11] Coren: yep, back again, thx [19:53:22] hedonil: You should probably ask Theopolisme to make you a maintainer so you can start it by hand. [19:53:32] But, incidentally, OOM at 8G? What does that thing /do/? [19:53:41] hehe idk [19:54:15] memory leak I suppose, will watch it some minutes [19:56:42] Before the night from July 14th to July 15th a script of mine mostly ran ~5 min; since then a mysql query runs for hours and I didn't change anything... so are there any database problems known for tools-db? [20:03:25] appper: No, but you may have simply been bitten by the fact that this is a shared resource and someone else held locks for a while. [20:39:46] !log logstash Updating self-hosted puppet checkout after not paying attention to this project for way too long. [20:39:49] Logged the message, Master [20:41:19] 3Wikimedia Labs: centralauth_p is missing tables - 10https://bugzilla.wikimedia.org/66533#c3 (10Andrew Bogott) Was this resolved? [20:43:34] 3Wikimedia Labs: Service diamond creates 500+ MByte /var/log/diamond/diamond.log - 10https://bugzilla.wikimedia.org/66458#c1 (10Andrew Bogott) On a randomly chosen instance, I see: -rw-r--r-- 1 diamond nogroup 9810916 Jul 2 23:59 archive.log.2014-07-02 -rw-r--r-- 1 diamond nogroup 9810168 Jul 3 23:59 ar... [20:44:19] 3Wikimedia Labs: centralauth_p is missing tables - 10https://bugzilla.wikimedia.org/66533#c4 (10Marc A. Pelletier) It's moot, given the imminent transition to MariaDB 10. Keeping the bug open for tracking, however, until that's done. [20:51:50] 3Wikimedia Labs: Service diamond creates 500+ MByte /var/log/diamond/diamond.log - 10https://bugzilla.wikimedia.org/66458#c2 (10Tim Landscheidt) 5NEW>3RESO/FIX Yes, Chase and Yuvi fiddled with diamond quite a bit with the conclusion (IIRC) that it doesn't log at all on Labs. [20:56:05] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c7 (10Tim Landscheidt) (In reply to Andrew Bogott from comment #5) > I could potentially change the patch so that it only modifies the file if it > contains exactly 'labs-vmbuilder-precise.e... [21:23:05] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c8 (10Andrew Bogott) (In reply to Tim Landscheidt from comment #7) > (In reply to Andrew Bogott from comment #5) > > I could potentially change the patch so that it only modifies the file if... [21:26:35] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c9 (10Tim Landscheidt) (In reply to Andrew Bogott from comment #8) > [...] > I could make a patch in the firstrun script that sets this for new > instances. Of course that would leave all t... [21:34:06] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c10 (10Andrew Bogott) (In reply to Tim Landscheidt from comment #9) > (In reply to Andrew Bogott from comment #8) > > [...] > > I could make a patch in the firstrun script that sets this for... [21:39:16] andrewbogott: My patch may have broken wikitech notifications. I don't have any new notifications since 07-11 and I deleted 2 instances and built a third today. :( [21:39:36] bd808: It isn't deployed yet; scheduled for Monday morning [21:39:49] so if something has changed… it's not that [21:41:08] andrewbogott: Well that's good I guess. [21:41:33] And I just got a notification for the instance create so maybe I'm just paranoid [21:42:50] 3Wikimedia Labs: Service diamond creates 500+ MByte /var/log/diamond/diamond.log - 10https://bugzilla.wikimedia.org/66458#c3 (10Yuvi Panda) It does log, but only logs errors. We killed the archive handler that logged all the *metrics* being sent, which was causing the huge log files. [21:51:13] bd808: I received some notifications in the past few days, so it seems okay^Wunchanged. [21:52:10] Hello [21:52:26] scfc_de: :) I didn't get delete notifications that I thought I would, but yeah I was just in too much of a hurry for the create notification. [21:52:51] I am getting ssh: connect to host tools-dev.wmflabs.org port 22: Connection timed out error while scp some files. [21:53:16] scfc_de: If my patch works on Monday you may be pleasantly surprised [21:53:16] I could do it successfully earlier today but suddenly it's not happening [21:54:08] Can I please get some help from any admin since I have some project deadlines... [21:54:49] bd808: I have followed your patches with excitement :-). [21:55:16] tantan: Works for me. Can you ssh to tools-login.wmflabs.org? [21:56:47] scfc_de yes [21:57:34] scfc_de I have ran a few queiries in the db and got some files. I am trying to copy them to my local machine. [21:58:33] scfc_de I can login to tools_login and tools_dev both [21:59:04] 3Wikimedia Labs: /etc/mailname is set to "labs-vmbuilder-precise.eqiad.wmflabs" - 10https://bugzilla.wikimedia.org/64962#c11 (10Tim Landscheidt) Okay, then I think Puppet is the right solution, and I assume merging the firstrun patch as well doesn't hurt? [22:03:35] tantan: Just scp'ed tools-dev.wmflabs.org:/public/dumps/public/dewiktionary/20140625/dewiktionary-20140625-pages-meta-current.xml.bz2 (~ 90 MBytes), and it worked fine. As a work-around, you could always copy the files you want to download to the public_html directory of one of your tools and use standard web downloaders that resume aborted downloads. [23:08:03] Coren: you about? [23:18:06] !log deployment-prep Puppet broken for deployment-bastion by labs specific logic in misc::deployment::vars. [23:18:08] Logged the message, Master [23:21:04] Betacommand: Sorta. What be up? [23:21:33] Coren: is there a way to block a UA at the proxy level so my access logs dont show it? [23:21:49] @scfc_de my files are in the directory which contains the cnf file [23:22:13] Betacommand: You mean block a connection from that UA or remove the User-Agent header itself? [23:22:51] Coren: Ive already blocked the UA, but Im still getting my access log flooded with those requests [23:22:51] @all, I am trying to scp a file to my local machine but getting this error. Can anyone please help. ssh: connect to host tools-dev.wmflabs.org port 22: Connection timed out [23:23:00] I have a project deadline. [23:23:37] Its the dam tweetmebot spider [23:23:59] Betacommand: I take it it disobeys robots.txt? [23:24:24] Coren: I dont have access to labs robots.txt [23:25:16] tantan: It works for me; perhaps the problem is on your (network) side? [23:25:40] tantan: You're not trying to use that name to scp /within/ Labs are you? [23:26:08] Betacommand: Lemme check if it even read robots.txt. [23:28:02] Betacommand: Can you give me a the UA please? [23:28:57] Coren: one moment [23:32:29] Coren: Mozilla/5.0 (compatible; TweetmemeBot/3.0; +http://tweetmeme.com/) [23:41:10] Betacommand: What I find interesting is that, technically, Tweetmeme is defunct. [23:41:14] Reedy: was it you who re-started Jenkins the other day? I think it's in trouble again. [23:41:20] Coren: I know [23:41:42] chrismcmahon: I did a couple of times, yeah [23:41:44] Reedy: I can shut Jenkins down but I don't think I can restart it. :-/ [23:42:19] chrismcmahon: sudo -u jenkins /etc/init.d/jenkins start [23:42:20] IIRC [23:42:35] is the sudo needed.. [23:42:37] * Reedy looks [23:43:03] sudo /etc/init.d/jenkins start [23:43:27] Reedy: this is gallium, right? [23:43:30] * chrismcmahon tries it out [23:43:34] yeah [23:43:38] @Coren, I dont understand why I am getting this error since it worked for me this morning. [23:44:17] tantan: What do you mean by "my files are in the directory which contains the cnf file"? [23:45:34] @scfc_de, I am sorry I am very new to wikilabs. :( [23:46:12] From my sayantan-13@tools-login:~$ prompt when I am doing "ls" I can see the files [23:46:44] tantan: You cannot scp to that host name from tools-login; but you also wouldn't *need* to do so; your home is shared between both hosts already. [23:47:04] (In fact, you could destroy the files by attempting to copy them over themselves via NFS) [23:47:06] @scfc_de it's when I am trying to do the following scp -v sayantan-13@tools-dev.wmflabs.org:may31-jun14-revisions.txt may31-jun14-revisions.txt I am getting the error [23:47:13] no dice, can't seem to get to gallium [23:47:39] @Coren, in that case how I scp the files to my local directory? [23:47:46] tantan: Also, I can guarantee that couldn't possibly have worked before. Perhaps you copied files from your /local/ computer? [23:48:10] tantan: I don't understand your question. Anything that is in your home is already accessible from either hosts; there is nothing to copy. [23:48:44] Coren: AFAIUI, he wants to download them to his own machine. [23:49:05] tantan: Can the files be made publicly available, or do they contain private data? [23:49:38] @scfc_de, they can be made publicly available. [23:49:43] tantan: Which computer are you attempting to write that scp command on? [23:50:26] @Coren, on my own machine. @scfc_de is right. I am trying to get those files to my local machine. [23:51:18] @scfc_de the only thing is i need to do this with other files too so I need to know how I can do it on my own. [23:52:44] @scfc_de, @Coren, I was able to something similar this morning: scp sayantan-13@tools-dev.wmflabs.org:may31-june28-users-revisions.txt may31-june28-users-revisions.txt If you are having access problems, please see: https://wikitech.wikimedia.org/wiki/Access#Accessing_public_and_private_instances may31-june28-users-revisions.txt 100% 78MB 2. [23:56:03] tantan: I can see no reason on this side why it wouldn't work; and I can copy files back and forth with no issues. [23:56:17] tantan: After logging out and in again, you should be able to copy files on Tools (with "cp" on the Tools command line) from your account to /data/project/tantan-www/public_html and they should appear . [23:56:33] Coren: Me, too. No clue why it fails for him. [23:59:13] Betacommand: datasift has been blocked at the network layer. [23:59:31] Coren: datasift?