[00:06:57] New patchset: Platonides; "Make dir a normal fd." [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/70110 [07:16:05] zz_YuviPanda . [07:54:34] zz_YuviPanda you have admin on toolsbeta, getting yourself root is a first task you need to pass to prove you are ready for it :P [07:57:23] !log tools petrb: 50527 4186 22830 1 Jun23 pts/41 00:08:54 python fill2.py eats 48% of ram on -login [07:57:26] Logged the message, Master [07:58:29] henna ping [07:58:38] henna: can you consider moving this job to grid? :P [07:58:46] or -dev at least... [08:08:04] grrr Cannot login in AWB o-O [08:49:44] a930913: still want help with flask on cgi? [08:53:14] a930913: you can see https://github.com/addshore/dumpscan/blob/master/dumpscan.py [08:55:52] petan: the second sentence confused me. [08:56:01] petan: admin != root? [08:56:08] YuviPanda no [08:56:12] ok [08:56:14] so I have admin [08:56:15] but not root [08:56:17] YuviPanda admin == being able to grant root [08:56:23] .. [08:56:23] ahh [08:56:47] so I don't have root but I can grant myself root? :) [08:56:51] yes [08:57:08] and if you can't do that simple task you probably don't deserve it :D [08:58:48] :P [08:58:49] right [08:59:02] petan: is toolsbeta puppetized? [08:59:09] like, are the various components built from puppet? [08:59:10] or hand? [08:59:37] toolsbeta is using same stuff as tools [08:59:42] same puppet classes [08:59:53] check the puppet config of instances [08:59:59] as admin you apparently can change instances too [09:00:03] ah, sweet [09:25:26] YuviPanda: Ta. I got it working, though gave it up after I realised it wasn't a daemon. [09:25:35] ? [09:25:38] petan: Is py3 -dev installed? [09:25:47] ? [09:25:57] what is that? [09:26:03] is that a library or some program? [09:26:16] why there is space between py3 and -dev [09:26:33] YuviPanda: Flask cgi is loaded per webpage hit. [09:26:44] that's a CGI issue, yeah :) [09:26:54] everything is :) [09:26:54] petan: Because I'm tired :p [09:26:58] should be better with wsgi, maybe even fastcgi [09:27:05] a930913 I have no idea what you want [09:27:15] YuviPanda: I've settled for running a daemon that is queried by the cgi. [09:27:21] are you worried about 'performance implications' of loading flask everytime? [09:27:26] hmm, how are you running the daemon? [09:27:27] petan: python-dev for python 3. [09:27:30] continuous task? [09:27:39] YuviPanda: Atm, in a screen on login. [09:27:53] Coren doesn't like that, IIRC :) [09:28:06] YuviPanda: Hey, I'm developing it. [09:28:10] hehe [09:28:14] :) [09:28:21] When it works, I'll migrate it to the grid :) [09:28:28] I'm going to have to do some celery stuff to have a core framework [09:28:39] of how to do this kinda 'hand off stuff to grid and then wait' tools [09:30:08] petan: I tried pip installing for python 3, and it complained of no Python.h [09:30:13] ok sec [09:30:19] you want that on -dev I guess [09:31:16] petan: i think he's looking for the package 'python3-dev' [09:31:25] yes I knwo [09:31:26] which gives header files for python3 [09:31:27] ok [09:32:23] installed [09:32:32] petan: I've only used -login. I never really learnt what was out there, nor how to connect. [09:32:52] -dev is a place to build stuff at [09:33:08] -login is only supposed to be used for ssh / management of grid jobs etc [09:39:58] petan: Ah, kk. [09:40:19] Well, now to work out what's causing the next error D: [09:44:16] a930913: summarized: don't use -login, use -dev, unless -dev is overloaded and you quickly need to start a job :-p [09:44:57] valhallasw: Might want to make that clearer on the wiki, if it isn't already. :) [10:08:11] !log wikidata-dev wikidata-dev-9 installed "WikibaseDataModel" in "/srv/devrepo/w/extensions" [10:08:13] Logged the message, Master [10:11:14] !log wikidata-dev wikidata-dev-9 installed "WikibaseDataModel" in "/srv/devclient/w/extensions" [10:11:15] Logged the message, Master [10:11:52] !log wikidata-dev wikidata-dev-9 updated crontab to pull "WikibaseDataModel" [10:11:53] Logged the message, Master [10:19:33] petan: ping [10:19:37] how do I give myself root? [10:19:44] sudo nor su work, so can't really visudo [10:20:35] YuviPanda: probably by adapting sudoers in the puppet config? [10:20:41] not sure, though [10:20:56] or something something ldap [10:21:07] well, that would require someone to merge that, and petan said I just need to be member of admin group [10:21:07] hmm [10:21:36] maybe there is a root service group or something? [10:22:46] hmm [10:22:49] * YuviPanda looks at wikitech [10:25:06] hmm I can edit config in wikitech [11:07:01] !sudo [11:07:01] You have sudo in any project that you are a member of, excluding global projects (like bastion). Your sudo password is your labsconsole wiki password. [11:07:06] mhm [11:07:19] YuviPanda: did you try sudo groups? :P [11:07:21] in wikitech [11:07:30] grr, didn't click that [11:07:43] YuviPanda I recommend you to read labsconsole help [11:08:43] !log Granted yuvipanda root on toolsbeta [11:08:43] Granted is not a valid project. [11:09:08] !log toolsbeta Granted yuvipanda root on toolsbeta [11:09:10] Logged the message, Master [11:09:15] mmm :) [11:13:52] petan: that !sudo message makes no sense- I doubt I have sudo at tools :P [11:14:18] valhallasw: heh, that makes sortof sense for labs in general, tho [11:14:21] not for tools-labs [11:14:28] valhallasw it's from times when labs started [11:14:34] old several years I guess [11:14:43] yeah [11:14:51] so please update it to make sense ;-) [11:15:03] valhallasw: [[WP:BB]] [11:15:06] something like 'where you are in the root sudo group' [11:15:07] "If you are wondering if you have sudo, you don't" [11:15:14] valhallasw broken? fix it! [11:15:25] petan: except I don't know how it works [11:15:34] valhallasw it's broken because you didn't fix it :P [11:15:41] read help, educate yourself :P [11:15:45] This bot is running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.10.8.15 source code licensed under GPL and located at https://github.com/benapetr/wikimedia-bot [11:16:06] petan: 'it' = sudo on labs, not wm-bot [11:16:07] geez. [11:17:03] it's pretty enough to link it to https://wikitech.wikimedia.org/wiki/Help:Sudo_Policies [11:17:15] !sudo is https://wikitech.wikimedia.org/wiki/Help:Sudo_Policies [11:17:15] This key already exist - remove it, if you want to change it [11:17:23] !remove sudo [11:17:34] ugh. Whatever. [11:18:12] !sudo del [11:18:12] Successfully removed sudo [11:18:14] !sudo is https://wikitech.wikimedia.org/wiki/Help:Sudo_Policies [11:18:15] Key was added [11:18:29] happy now? You could have, of course, done that in about 1/10th of the time it took me. [11:20:18] * YuviPanda gives valhallasw ice cream [11:20:28] Thanks. [11:32:41] valhallasw yes I could, but in that case you would never learn it :P [11:32:56] that's how wikipedia works - it's broken? fix it [12:13:35] addrest, ping [12:43:19] http://deployment.wikimedia.beta.wmflabs.org/w/index.php?title=Special:AbuseLog&wpSearchFilter=8 [12:43:32] i hav locked som spambots global som day ago. [12:43:45] But the spambots continuing to spamming o_O Bug? [12:46:00] *ping* Hashar [12:53:36] Steinsplitter: pong [12:55:53] see above, hashar. seems to be a bug [13:00:58] Steinsplitter: sounds to me they are still trying [13:01:04] despite receiving a warning and a disallow [13:01:25] jepp. But the accounts ar globaly locked? [13:01:25] the page http://simple.wikipedia.beta.wmflabs.org/wiki/User:Christinab2l has no content [13:01:54] jep. The filter :). But locked accounts cannot perform any actions.? [13:01:57] I am not sure what the 'Disallow' action is [13:02:03] maybe it does not block the user [13:02:10] also global blocking might not be working [13:02:48] Disallow = do not save the page [13:03:23] what is strange is that http://simple.wikipedia.beta.wmflabs.org/wiki/User:LolaHarrington [13:03:32] does not seem to be an account on the simple wiki [13:03:41] jepp. [13:03:51] hashar: cu is disalowed on beta? [13:03:58] cu ? [13:04:07] CheckUser [13:04:13] ah yeah [13:04:14] it is [13:04:17] as well as oversight [13:04:29] that scared some people apparently [13:04:50] :----( [13:05:08] at least the filter block the edits \O/ [13:05:21] jepp [13:07:01] hmpf... seems to be a bug in the global filter (Lola Harington). i ask later hoo_man [13:07:16] Does a puppetmaster need a special security group? [13:26:05] Steinsplitter: LolaHarrington is a global user on labswiki apparently [13:26:34] ah [13:41:35] scfc_de try to use take on /tmp/take-test [13:41:38] cd /tmp/take-test [13:41:40] take * [13:41:48] on -login [13:42:01] !log toolsbeta Created toolsbeta-puppetmaster. [13:42:03] Logged the message, Master [13:42:45] "05-role-role--labs--tools--webproxy: You need to share a group with the file" -- and? [13:42:56] (That's not "file not found".) [13:43:10] !log toolsbeta Made scfc root. [13:43:19] Logged the message, Master [13:44:04] scfc_de hold on [13:44:24] try now [13:45:16] Okay, I see that error. [13:51:19] "take 05-role-role--labs--tools--webproxy 40-infrastructure-banner 40-infrastructure-banner-beta 99-footer addtool continuous" works, "take 05-role-role--labs--tools--webproxy 40-infrastructure-banner 40-infrastructure-banner-beta 99-footer addtool bin continuous" doesn't. [13:51:57] "take bin" works, "take bin continuous" doesn't. [13:54:02] So we fchdir() to the subdirectory in line #130, but don't change back. [14:26:43] scfc_de: still reading through puppet docs, but I guess now that we have a puppetmaster locally, we can put our scripts in there and have the other instances update from that? [14:29:22] YuviPanda: I have read the Puppet docs, but have to read up on puppetmaster::self :-) (and especially how to set clients to another puppetmaster in the instance config). I'm waiting for Ryan_Lane or andrewbogott to come online. [14:29:33] ah, wonderful [14:32:19] @notify Coren [14:32:19] This user is now online in #wikimedia-labs. I'll let you know when they show some activity (talk, etc.) [14:39:33] New patchset: coren; "take.cc: Fix argument evaluation (directories)" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/70170 [14:47:37] !log toolsbeta petrb: rebooting exec-01 to fix the grid weird info [14:47:39] Logged the message, Master [14:47:55] !petan [14:47:55] Petr Bena - http://enwp.org/User:Petrb [14:48:02] hey [14:48:11] !Coren [14:48:12] Coren is dead. petan killed him. He now roams about as a zombie. [14:48:19] :D [14:48:45] petan, why? ^ [14:49:28] I'm not really here today (statutory holiday) but I thought I might point out the mysterious appearance of s7. [14:49:51] Coren|DayOff, really? :):):):) [14:50:22] Coren|DayOff you promised to review new take :o [14:50:25] Not centralauth itself though, that one is still missing its views (the automated system couldn't work for it). The wikis on it are though. [14:50:35] it's just like 200~ lines of code :3 [14:51:35] Coren|DayOff, what about legal? [14:52:47] petan: well, it is his day off :) [14:53:17] YuviPanda :( [14:56:37] petan, you are hereby promoted to Acting Coren. [14:56:46] Coren|DayOff urgent PM [14:57:00] petan, now get to work. :D [14:58:17] Cyberpower678: :DS [14:58:22] without S, actually [14:59:17] valhallasw, I just gave my Peachy framework the ability to update itself. [15:02:32] hi, my bot instance (bots-salebot) is not responding... can an admin restart it? [15:02:40] (do I have the rights to do that myself?) [15:03:51] gribeco: Are you talking of the Bots project? [15:04:13] scfc_de: the instance is part of the bots project, yes [15:07:12] gribeco: I think only admins can do that. [15:07:49] gribeco, you could try Coren but... [15:07:51] !Coren [15:07:51] Coren is dead. petan killed him. He now roams about as a zombie. [15:09:26] petan: could you reboot bots-salebot and/or Coren? [15:10:33] petan can't reboot Coren, I think [15:10:48] here [15:10:53] gribeco yes [15:11:03] petan: cool, thanks [15:11:13] gribeco you really want to move that bot there to tools project... [15:11:28] ok [15:11:51] * Cyberpower678 reboots petan [15:15:11] !log bots rebooting salebot instance [15:15:13] Logged the message, Master [15:17:16] New patchset: Tim Landscheidt; "take.cc: Fix argument evaluation (directories)" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/70170 [15:20:00] petan: looks a lot better now; thanks! [15:21:17] petan: Are all tools supposed to have new replica.my.cnf? Or just users? [15:26:02] oh, sorry, my mistake [15:26:06] all users [15:26:08] tools are fine [15:26:11] I will update my email [15:27:17] anomie ^ [15:27:53] hashar: do you have any thoughts on why this might be happening in beta? https://bugzilla.wikimedia.org/show_bug.cgi?id=49911 [15:29:31] chrismcmahon: maybe search is dead :) [15:30:10] hashar: it's the expand/collapse on the left side missing that worries me more [15:31:19] that sounds like two different bugs to me [15:31:36] the expand collapse might be a regression in the Vector skin / extension [15:31:42] or in the mediawiki configuration [15:31:49] the search is probably because search is dead [15:37:53] !log deployment-prep upgrading -searchidx01 and refreshing puppet manifests [15:37:55] Logged the message, Master [15:38:18] !git operations/puppet [15:38:18] For more information about git on labs see https://labsconsole.wikimedia.org/wiki/Help:Git [15:38:50] petan, S7 seems to be frozen. When I did sql metawiki, everything stopped responding. [15:39:06] And the edit counter still isn't working with it either. [15:39:43] petan, [15:39:44] ERROR 2003 (HY000): Can't connect to MySQL server on 'metawiki.labsdb' (110) [15:39:44] Make sure to ask for a db in format of _p [15:40:10] labsdb.metawiki_p ? [15:40:20] hashar: does manybubbles know that search is dead? [15:40:37] hashar, hih? [15:40:44] huh? [15:40:45] long live search? [15:40:46] Cyberpower678: maybe metawikiwiki? [15:40:48] sumanah: no and he doesn't have to care about it :-) [15:40:58] sumanah: that is the old lucene install being broken in beta :-D [15:40:58] oh ok [15:41:02] nevermind [15:41:12] sumanah: thanks =) [15:41:13] I do care - it makes me want to replace it even more! [15:41:21] Cyberpower678 I would try just meta, or find out what the name is [15:41:26] you will see it in /etc/hosts [15:41:41] these special wiki's like meta were aliased in sql so that it's easier to get to them [15:41:51] local-xtools@tools-login:~$ sql metawiki [15:41:51] ERROR 2003 (HY000): Can't connect to MySQL server on 'metawiki.labsdb' (110) [15:41:51] Make sure to ask for a db in format of _p [15:41:51] local-xtools@tools-login:~$ ^C [15:41:51] local-xtools@tools-login:~$ sql metawikiwiki [15:41:54] This is unknown db to me, if you don't like that, blame petan on freenode [15:41:54] manybubbles: yeah it is a bit unstable in beta, the java daemon dies from time to time and I don't think we have any monitoring :-] Will fix it up manually [15:41:56] Make sure to ask for a db in format of _p [15:42:00] local-xtools@tools-login:~$ sql meta [15:42:02] This is unknown db to me, if you don't like that, blame petan on freenode [15:42:02] Cyberpower678 I said "meta" [15:42:04] Make sure to ask for a db in format of _p [15:42:06] ok [15:42:09] let me find out [15:42:22] From what I'm seeing, it's metawiki [15:42:40] !log deployment-prep restarted lucene-search-2 on searchidx01 [15:42:42] Logged the message, Master [15:42:45] Cyberpower678 sql metawiki works to me [15:43:02] which is s7 [15:43:05] that doesn't work :P [15:43:07] java.io.IOException: Server returned HTTP response code: 403 for URL: http://de.wikivoyage.beta.wmflabs.org/w/index.php?title=Special:OAIRepository&verb=ListRecords&metadataPrefix=mediawiki&from=2001-01-01 [15:43:08] oh yeah [15:43:32] petan, Like I said. S7 is frozen. [15:44:09] not anything I can fix :/ [15:44:13] Coren|DayOff, S7 mysteriously appeared broken. [15:44:21] :p [15:44:50] chrismcmahon: search is fixed up :-) [15:45:13] ok thanks hashar. those issues showed up at exactly the same time last week [15:45:26] !log tools petrb: changed colors of root prompt productions vs testing [15:45:28] Logged the message, Master [15:46:59] chrismcmahon: for the sidebar "triangles", I do got them using Safari 6.0.5 and http://en.wikipedia.beta.wmflabs.org/wiki/Main_Page [15:47:13] or not hmm [15:47:56] chrismcmahon: yeah I got some javascript errors in my console :( [15:48:02] Exception thrown by skins.vector.js: 'undefined' is not a function (near '...;}}).collapsibleTabs();});(function($){v...') [15:48:21] hashar: nice. I don't know that I saw that reliably last week. [15:48:29] looking again [15:48:41] though using ?debug it doesn't show [15:48:42] http://en.wikipedia.beta.wmflabs.org/wiki/Main_Page?debug [15:48:52] so that must be some page which is staleld [15:49:00] I guess the bits cache is not properly purged [15:50:37] petan: does wm-bot also allow logging for service groups? [15:50:47] huh [15:50:49] logging of what [15:51:11] oh wait, that's labs-morebots that does that, now wm-bot [15:51:43] and logging as in !log [15:51:54] logging as !log is morebots [15:52:03] wm-bot is just relaying !log messages from terminal [15:52:13] so that I don't need to switch windows [15:52:19] I am very lazy if you didn't notice yet [15:52:24] I am famous for that [15:53:16] hashar: bits cache? [15:53:53] chrismcmahon: sorry was merely speaking to myself. The bits cache hold the javascript and css files, that is not related to the issue [16:02:35] chrismcmahon: ahhh [16:02:39] Vector has not been updated :-] [16:14:19] hashar: well, I'm glad it's not a production bug, but it is breaking a lot of my builds :) [16:20:40] I am not able to get onto parsoid.wmflabs.org -- get kicked out right away [16:28:35] paravoid, can you help me? [16:39:14] New review: Platonides; "I considered whether take folder in the non-recursive case should change the folder owner or not (or..." [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/70059 [16:41:01] scfc_de: do poke me when you start doing stuff with the puppetmaster :) [16:43:14] New review: Platonides; "(1 comment)" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/70058 [16:44:05] hello ops folks, once again .. I am not able to get onto parsoid.wmflabs.org -- get kicked out right away because of this: "Unable to create and initialize directory '/home/ssastry'" .. Roan and Gabriel cannot get on either because of similar issues. [16:45:23] * subbu goes back to writing tests [16:47:41] chrismcmahon: will attempt to have a look at it tongiht [16:47:51] chrismcmahon: seems to be some cache related issue. I might just clear the cache [16:54:27] YuviPanda: Will do. [16:54:31] ty [17:04:06] New patchset: Platonides; "take.cc: Fix argument evaluation (directories)" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/70170 [17:04:59] New review: Platonides; "PS3: Fixed space indentation, and skipped fchdir() when unneeded." [labs/toollabs] (master) C: 1; - https://gerrit.wikimedia.org/r/70170 [17:07:58] Can someone tell me the external hostname of the replication databases? [17:09:36] Cyberpower678: You can't access them from outside. [17:09:48] scfc_de, grrrr. [17:11:02] Cyberpower678: Why do you want to do that anyway? You have been shown multiple times how to use ssh tunnels. [17:11:24] scfc_de, I can't seem to get them to work. [17:14:25] hi, it seems that gluster seems to have issues again- we still get login failures on parsoid-spof [17:16:39] after a reboot of the VM services that depend on gluster did not come back up again either [17:19:06] andrewbogott, Ryan_Lane: ^ [17:19:14] gwicke: try now [17:19:22] I fixed it for you about 10 mins ago [17:19:53] thanks, trying.. [17:20:14] fixed, thanks! [17:20:17] yw [17:20:27] we'll be switching to NFS for homedirs this week [17:20:31] and project storage later on [17:22:33] back to good old NFS ;) [17:22:56] well, till we find a DFS that does multi-tenancy reliably :) [17:23:30] Ryan_Lane: I am trying to puppetize Redis for tools labs, do take a look at https://gerrit.wikimedia.org/r/#/c/70212/ and its dependency when you can :) [17:23:39] Ryan_Lane: Re puppetmaster::self, can (and how do) I direct other instances to pull their puppet config from a project puppetmaster, i. e. "toolsbeta-puppetmaster". I see a variable $::puppetmaster, but I don't see how I could set it in the instance configuration (Special:NovaInstance...). [17:24:11] YuviPanda: well, it has a −1 from jenkins ;) [17:24:21] hmm, what did I do now [17:25:10] YuviPanda: Missing "{"? [17:25:22] No. [17:25:41] no [17:25:58] scfc_de: I haven't set this up, I don't know [17:26:02] Ryan_Lane: https://gerrit.wikimedia.org/r/#/c/70064/3 is a dependent commit, and doesn't have a -1 :) [17:26:05] i'm looking at the -1 now tho [17:26:14] it may be that we need to add a global variable for this [17:26:33] andrewbogott: Did you write puppetmaster::self? [17:26:35] I'll ask ottomata when he gets on [17:27:30] Ryan_Lane: ottomata designed puppetmaster::self? [17:27:50] he changed it to allow instances to point to a central instance [17:28:08] Ryan_Lane: -1 gone :) https://gerrit.wikimedia.org/r/#/c/70212/ [17:28:22] scfc_de: I... had $s there. [17:28:27] Ryan_Lane: Okay. [17:32:57] /data/project is mounted by automount? [17:33:07] yes [17:35:03] Ah, I missed setting a role for that instance. That makes sense. [17:36:29] YuviPanda: I added a comment [17:37:27] is Labs already considered stable enough for productiony stuff? last year, we used a dedicated server for WLM, can it be different this year? [17:37:50] MaxSem: labs is not meant for production at all [17:37:59] Ryan_Lane: hmm, I was a little confused about .each vs .each_pair, so opted for safer alternative. [17:38:03] * YuviPanda 's ruby is nonexistant [17:38:04] Ryan_Lane: fixing [17:38:27] MaxSem: if this is a production need, then it should go in production [17:38:36] Ryan_Lane, yet is hosts Tool labs:) [17:38:38] tool labs is not production ether [17:38:41] *either [17:39:06] it has slightly better support than TS, but not the same as production [17:39:40] yep, so WLM was just a way to run TS stuff on a more stable server [17:40:50] if I recall correctly, you guys were doing some weird crap with moving data from TS to a production machine [17:41:01] pretty sure we weren't fans of that [17:41:08] we neither;) [17:41:23] labs isn't the answer for that [17:41:25] we aren't doing any of that, IIRC. [17:41:30] this year, I mean [17:41:35] ah. good [17:41:41] they're moving all their stuff to WikiData [17:41:44] so, what's the need for labs hosting, then? [17:42:06] tfinc confirmed that we aren't doing it about 5 seconds ago :P [17:42:06] so [17:42:12]