[00:48:19] andrewbogott: hey [00:48:21] andrewbogott: basically SpecialPage::getTitle is deprecated in core now, and extensions should use SpecialPage::getPageTitle instead. If you don't change it, nothing will break right now, it'll probably be a few versions/years before the function is removed [00:48:38] Ah, yes, but... [00:48:42] Well, this is moot now :) [00:49:25] getPageTitle is only supported in the very latest version of MW. So your patch required an upgrade. [00:49:45] I'm not sure what the best option was there, other than maybe mentioning that in ALL CAPS in the comment. [00:49:54] Did Ryan merge it? Or someone else? (did I?) [00:51:37] ah [00:51:39] sorry :/ [00:52:28] I just ran a script to submit commits for every extension [00:52:29] https://gerrit.wikimedia.org/r/#/c/103962/ [00:52:34] siebrand merged it [00:53:11] Hm, that means that every extension requires the tip of MW? [00:55:25] no [00:55:27] some do [00:55:35] just depends who maintains it [00:55:59] Well, I mean -- if you submitted similar patches to many extensions, then... [00:56:03] all of those do, now :) [00:56:06] petan, ping [01:00:51] not all of them were merged [01:04:06] (03PS1) 10Tim Landscheidt: Fix anchor links from status to list page [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106173 [01:04:55] (03CR) 10Tim Landscheidt: "Already hotfixed." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106173 (owner: 10Tim Landscheidt) [02:53:54] Coren: If you happen to still be around this late, I have two questions for you. (1) Is there a reason task@tools-exec-09 is "temporarily not available" but continuous@tools-exec-09 is available? (2) Is job 1895011 actually doing something with all that CPU it's burning, or is it stuck in a loop? [12:17:23] !ping [12:17:47] !ping is pong [12:24:20] !ping [12:24:21] !pong [12:24:28] !ping [12:24:28] !pong [13:44:05] !log tools Cleared error states of continuous@tools-exec-05, task@tools-exec-05, task@tools-exec-09 [13:44:06] Logged the message, Master [13:54:46] may be going to Bangladesh as well for a few days [14:58:03] Can someone remind where the documentation to wm-bot is? [15:00:06] @doc [15:00:12] @documentation [15:00:15] !doc [15:00:15] What's up doc? [15:00:19] !docs [15:00:20] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help [15:00:27] !botdoc [15:00:36] hedonil, ping [15:00:50] Cyberpower678: pong! [15:01:02] hedonil, where can I find the bot documentation [15:01:33] Cyberpower678: i think I have no clue, sry [15:01:50] Coren, petan anomie ^ [15:02:08] @help [15:02:08] I am running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.20.2.0 my source code is licensed under GPL and located at https://github.com/benapetr/wikimedia-bot I will be very happy if you fix my bugs or implement new features [15:02:10] Cyberpower678 ^ [15:02:21] He speaks. :p [15:02:38] petan, can I redirect the sulinfo tool? [15:02:43] what [15:03:06] redirect what where and why [15:03:09] You started an sulinfo project on tools. [15:03:13] yes I did [15:03:30] With a forked copy of quentinv57's sulinfo script. [15:03:44] well, I copied it before it gets deleted from toolserver [15:03:48] it was rather a move than a fork :P [15:04:01] Well, now it's hosted on quentinv57-tools [15:04:15] With me as a co-maintainer. [15:04:16] why don't you just update it on sulinfo project? [15:04:26] Because I have no access. [15:04:27] it's not that hard to execute cp [15:04:31] well you don't [15:04:34] because you never asked for it [15:05:12] petan, can I have access? [15:07:32] yes, you have it, but I think it's better to create a service group for every service, rather than createing a service group "bob's tools" and keep all services under it [15:07:53] I'm only respecting quentin's decision. [15:08:10] I think his decision is wrong [15:08:36] I prefer keeping them in one place. [15:08:41] that's wrong thing [15:08:47] it will make it extra hard to maintain [15:08:51] How? [15:09:03] the point of tools labs was to have tools isolated from each other [15:09:09] this way tool A can damage tool B [15:09:09] I think it's difficult knowing where your tools are if they are all seperated. [15:09:15] because they are both hosted under same user [15:09:43] well, they are all in /data/project ;) [15:09:54] /data/project/$tool [15:09:57] I haven't had any issues with that yet. [15:10:03] you will have [15:10:24] for example, when your tool A start crashing webserver, all your tools will be down [15:10:41] So? [15:10:51] so the whole system will be less reliable [15:11:06] Fix the cause of the crash and it won't happen again. Doesn't add any difficulty. [15:11:07] services should be isolated from each other so they don't have access to resources they don't need [15:11:15] it adds outages lol [15:11:33] service A shouldn't be able to crash service B [15:11:39] this way it is [15:11:41] Or better yet, make it so it doesn't crash in the first place. [15:11:43] it's a bad security design [15:11:50] it's also harder to balance the load [15:12:10] because all services will be considered as 1 service by grid [15:12:17] xtools seems to be doing just fine. [15:12:18] they all will be hosted by 1 webserver only [15:12:24] the webserver load will never be balanced [15:12:31] Also, some maintainers want to join to work on one tool, but not all of Quentin's. [15:12:41] +1 [15:13:10] instead of granting access to tool A you need to grant access to all of the tools, which imposes security threat [15:13:29] giftpflanze: I need to specify a tool which will have access to your node. Which do you want? [15:13:31] When it comes to that, I will seperate them, but for now, they're staying together. [15:13:58] Cyberpower678: And petan is correct, you /should/ separate the tools. There is no resource cost in splitting them, and it makes all of them easier to find and manage. [15:14:14] Coren: i have a hard time just adding somebody to an LDAP group, and i have done it multiple times before.. it's odd [15:14:17] *easier? [15:14:24] yes [15:14:33] there is even a section for every tool on tools mainpage [15:14:42] mutante: Odd. Details? [15:14:42] with own description, name and list of maintainers [15:14:58] it's a service group not services group [15:15:23] Coren: root@formey:~# ldaplist -l passwd adrian [15:15:23] I would have trouble heving to keep in mind all of those service groups. [15:15:28] BTW, Magnus Manske did a fantastic job separating his tools. [15:15:51] Coren: the user exists per ldaplist, but when i want to add him to group "wmde" it says he does NOT exist [15:16:02] adrian doesn't exist, and won't be added to the group. [15:16:41] well, this is all about a design, keeping the together is not a best design, despite it may work. And keeping the wrong desing on beginning will only make it harder to redesign them later [15:16:59] Coren, petan: It's not that you don't have me convinced, but for xtools for example, needs some major code changes first. The tools are interdpendant of each other in some cases. [15:17:30] xtools has a wierd setup. [15:17:33] Cyberpower678: One does not preclude the other. Betacommand has a number of tools that share code. [15:18:16] Coren, I'll give it some serious thought. [15:18:48] mutante: Trying to see what's up. [15:18:56] Coren: thank you [15:18:57] Coren: what about me? [15:19:05] Betacommand, nothing. [15:19:24] Coren: root@formey:~# modify-ldap-group --addmembers adrian wmde [15:19:27] Betacommand: You were just cited as a positive example. I know, right? :-) [15:19:51] Coren: I dont split my tools [15:19:57] Coren: it's ticket 6604 , not asking for changes without one :) [15:20:18] * petan takes away badge "exemplary user" from Betacommand [15:20:32] mutante: I'm not trying to do the group thing, but to figure out why you can't. ;-P [15:20:43] kk [15:22:02] mutante: And I failed at that; while trying to see what happens it just worked without issue. [15:22:06] wmde:*:1006:jkroll,addshore,adrian,jeroendedauw,hsnater,tgritschacher,aude,lydia [15:22:18] seems it worked [15:22:20] Coren: ehhh .. :) [15:22:33] thanks, but a it weird [15:22:33] strange [15:22:56] mutante: I'm annoyed because I have no idea why it might not have worked for you. :-( [15:23:14] and i used it before and the command was from bash history [15:23:22] oh well, let's see if it ever comes back [15:23:31] otherwise we ask coren :P [15:24:00] aude: resolving ticket then, i'll paste the group members [15:24:06] no daniel k :/ [15:24:20] he never asked or replied about being in the ldap group [15:24:31] anyone interested in helping me fork and support http://tools.wmflabs.org/betacommand-dev/cgi-bin/replag into its own tool? [15:25:38] Betacommand: Should be fairly straightforward, but wouldn't it be better if it actually queried true slave replication status instead? [15:26:06] ^ That's what I'd really like to see in a replag tool [15:26:37] Coren: not at times, remember when a lock caused a mass backup and the slave status was wrong? [15:27:15] Betacommand: Hm. Well then it's not really a question of replication. But I see your point. [15:27:49] Coren: it caused what was in effect replag [15:28:16] * Coren would have thought Seconds_Behind_Master was a more useful metric for the general case though. [15:28:50] Isn't this in Ganglia already? (BTW, Coren, you closed that bug report prematurely, as currently the report script runs unpuppetized under my personal account.) [15:30:04] scfc_de: Ah, that should be fixed. :-) [15:31:32] Coren: I wanted to puppetize this as a property of the DB server, not the client, as the information is probably useful for all MySQL slaves, not just Labs ones, but never got around to work it out with Asher. [15:38:04] @seen petan [15:38:04] Cyberpower678: I have never seen petan [15:38:19] "petan " isn't here :P [15:38:43] petan, I can't get @seen to work in a different channel. Do I need to do something? [15:38:49] yes [15:38:52] @seen-on [15:38:52] Seen was already enabled [15:55:08] petan, why are there so many wm-bots [15:55:20] where [15:55:48] 3WM-BOT [15:55:53] #wm-bot [15:56:10] to bypass the freenode limits [15:56:21] wm-bot is now in 160 channels [15:56:26] Oh. [15:56:36] Is it possible they can talk to each other? [15:56:38] there is a limit for number of channels per user as well as messages per second [15:57:00] so if there was only 1 irc session for bot it would be very slow to respond to people [15:57:09] !ping [15:57:09] !pong [15:57:17] this command would probably take several seconds :P [15:57:51] they don't talk to each other because they are all 1 process, which opens multiple connection to freenode. You see 5 bots, but in fact it's just 1 bot [15:58:03] look [15:59:11] @instance #wikimedia-labs wm-bot2 [15:59:11] This channel I never heard of :'( [15:59:19] @instance wm-bot2 #wikimedia-labs [15:59:20] Changed default instance of #wikimedia-labs to wm-bot2 [15:59:23] you see [15:59:35] @instance wm-bot #wikimedia-labs [15:59:35] Changed default instance of #wikimedia-labs to wm-bot [15:59:55] Are the tools from https://toolserver.org/~vvv available on Tools Labs? Does anybody know? [16:00:52] petan, it has bug though [16:00:54]

es.wikiquote.org : 88.20.234.108 edited Juan - P�gina redirigida a [[Juan el Evangelista]] - http://es.wikiquote.org/w/index.php?diff=271179&oldid=271177 [16:00:54]

es.wikisource.org : C�rdenas edited Plantilla:La mejor espigadera - http://es.wikisource.org/w/index.php?diff=622551&oldid=622539 [16:00:54] @seen wmtbot-8 [16:00:54]

huh: I have never seen wmtbot-8 [16:00:59] ??? [16:01:34] Cyberpower678: I don't see what is a bug in there [16:01:46] but if you want to report a bug use bugzilla :P [16:02:19] petan, @seen wmtbot-8, as it's in the channel returns that it hasn't seen this user. [16:03:15] that user said "@seen wmtbot-8 " [16:03:21] there must be no extra space on end of line :P [16:03:35] if you don't like it, fix it [16:03:45] or create a ticket [16:03:59] my memory suck so hard I probably forget this whole conversation in next 2 minutes :D [16:04:18] :D [16:05:44] petan, perhaps a trim function may be in order? [16:06:14] yes that would fix it [16:10:48] petan, does the bot log privately? [16:16:01] !searchlogs [16:16:09] !logsearch [16:16:09] http://bots.wmflabs.org/~wm-bot/searchlog [16:25:34] Cyberpower678: no it doesn't [16:26:06] petan, searchlog is broken btw. [16:26:17] btw Cyberpower678 there is php interface that gets log from sql [16:26:24] http://tools.wmflabs.org/wm-bot/logs/ [16:26:36] it's open source, you can implement some better search in that :P [17:21:54] (03PS1) 10Tim Landscheidt: Fix build and run-time dependencies [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106281 [17:24:25] Coren: the answer to your question is: giftbot [17:24:27] (03Abandoned) 10Tim Landscheidt: WIP: Fix build and run-time dependencies [labs/toollabs] - 10https://gerrit.wikimedia.org/r/104951 (owner: 10Tim Landscheidt) [17:25:42] Coren: will all jobs be submitted on that node or only when i specify it? [17:30:16] giftpflanze: /Normally/ when I dedicate a node to a tool, I request that only that node then be used for all jobs. [17:30:34] giftpflanze: (But you'll need to specify the queue explicitly either way) [17:30:57] so i can still send the normal jobs to the normal queues? [17:31:48] giftpflanze: Yeah; but if your tool is doing two very different thing it'd be better if you split it so that only the one that needs the special queue has it and vice versa. [17:32:02] (03PS1) 10Tim Landscheidt: Work around pbuilder not properly setting $USER [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106283 [17:42:57] (03CR) 10Tim Landscheidt: [C: 04-1] "take got included earlier, so I need to add shlibs again." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106281 (owner: 10Tim Landscheidt) [17:44:13] (03PS2) 10Tim Landscheidt: Fix build and run-time dependencies [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106281 [17:46:55] (03PS2) 10Tim Landscheidt: Work around pbuilder not properly setting $USER [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106283 [18:15:27] (03CR) 10Tim Landscheidt: "Ready to merge." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106281 (owner: 10Tim Landscheidt) [18:15:36] (03CR) 10Tim Landscheidt: "Ready to merge." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106283 (owner: 10Tim Landscheidt) [18:33:40] * Coren mutters unkind things about how long puppet can take in its first runs. [18:42:19] I think Puppet installs one package at a time, so this might explain this to some degree. [19:44:02] Greetings. I just requested shell access at wikitech.... wmflabs.org [20:25:26] hcohl_: Done. In which projects will you work on Extension:Math? [20:31:12] scfc_de: Hi there. I am developing a mediawiki archive of mathematical formulae and I am using many features of the math extension. I am working with Moritz Schubotz who is developing the extension and helping me debug certain aspects of caching and other important details. [20:36:22] hcohl_: That sounds interesting. Is there an existing project on Labs that this is related to, do you need a new project, do you want to develop this in the Tools project? [20:40:57] scfc_de: No, there is an existing project. [20:41:25] It's called "math" [20:44:24] hi, I tried to login to bastion.mmflabs.org, but it says permission denied [20:44:37] (bastion.wmflabs.org) [20:46:22] hcohl_: Re the Math project, you need to contact one of the members listed on its wiki page then to add you to their project. [20:46:38] Are you using Linux or Windows to access bastion.*w*mflabs.org? [20:46:47] linux [20:46:57] he says he already added me to the project [20:49:03] Is your Labs Unix username the same as your username on your own system? Otherwise, you need to "ssh $LABS_USERNAME@bastion.wmflabs.org". [20:56:07] yes I did ssh hcohl@bastion.wmflabs.org [20:58:28] And you uploaded the public key corresponding to the private key you're using to wikitech/Preferences? [20:59:02] I tried to [20:59:41] ... but? Is it listed on the preferences page? [21:00:42] I don't see anything on the preferences page [21:02:19] hcohl_: On https://wikitech.wikimedia.org/wiki/Special:Preferences#mw-prefsection-openstack, how many keys are listed for you? [21:03:52] it says there's a key [21:04:33] one public SSH key [21:05:52] And is this the same as ~/.ssh/id_rsa.pub on your system? [21:06:10] i literally copy and pasted this text file to the open window [21:06:50] it just let me in! [21:07:07] thanks! [21:07:59] !! [21:07:59] petan needs a new hobby :P [21:09:19] hcohl_: NP. /public/keys/hcohl was (first) created 21:05Z, so I don't know why previous changes on your preferences page didn't come through. [21:10:50] oh well, it all works now [21:11:08] thanks so much [21:39:31] !log wikimania-support Updated scholarship-alpha to d45c259 [21:39:33] Logged the message, Master [21:40:21] hi, I have problems to connect to my old instances on wikitech... i.e the server says --------------------------- Error --------------------------- Disconnected: No supported authentication methods available (server sent: publickey) --------------------------- Authentication log (see session log for details): Using username "physikerwelt". Authentication failed. --------------------------- OK Reconnect (3 s) Help - [21:41:52] furthermore the server version seems to be deprecated even though it's ubuntu-12.04-precise (deprecated 12-16-2013) and I should add that it has the puppet master self role [21:54:21] according to https://wikitech.wikimedia.org/wiki/Puppet there is no puppet server for labs anymore? [22:06:33] physikerwelt: Ryan updated the base images recently, so IIRC "deprecated" just means that the instance was created with the old image. Have you tried rebooting the instance to see if the auth prob may just be a hanging system? [22:54:58] physikerwelt: I'm quite distracted just now, but I can try to help you debug. What's the project and instance name? [22:55:29] latexml-test [22:55:33] on math [22:56:10] there is also a problem with the newly created instance drmf... I can logon there but puppet does not get a valid certificate [22:58:15] physikerwelt: how long ago did you create it? [22:58:18] drmf I mean [22:58:29] I hour or so [22:58:34] maybe two [22:58:51] ok, which would you like me to debug first? [22:59:04] there is noting on so I could drop and recreate without any trouble [22:59:25] I think drmf is more relevant since I did not do anything special were [22:59:26] there [23:01:25] The initial Puppet run can take quite some time. [23:01:51] I tried sudo puppetd -tv [23:02:57] this leads to err: Could not request certificate: The certificate retrieved from the master does not match the agent's private key. Certificate fingerprint: 18:E5:81:B2:80:DA:3C:A8:9B:42:2C:3C:36:D8:1B:89 To fix this, remove the certificate from both the master and the agent and then start a puppet run, which will automatically regenerate a certficate. [23:04:30] andrewbogott: thanks for fixing latexml-test now I can log into that machine again [23:04:40] !log wikimania-support Updated scholarship-alpha to be1d5a7 [23:04:41] Logged the message, Master [23:04:43] hm… I didn't do anything :) [23:05:05] ok maybe it's just because of the reboot [23:05:31] ok, then, on to the next one... [23:06:08] An instance I recently created took (at least) two subsequent Puppet runs to clear that certificate error. Never debugged what went wrong there, though. [23:06:40] physikerwelt: can you log in to drmf? [23:06:46] yes [23:06:49] Oh, I guess you said you could. Is that the same project? [23:07:04] yes [23:07:40] In the error message it's said that I have to run puppet cert clean i-00000906.pmtpa.wmflabs [23:07:43] on the master [23:08:27] but I don't have acess to the master (which is ok). [23:09:19] OK, I'm going to fix this by hand… please let me know if this happens again on a future instance. [23:09:37] ok thank you [23:09:56] I can create a further instance right now for testing that [23:10:48] sure, would be good to see. [23:13:28] ok same problem for the new instance test-puppet-cert [23:13:57] hm, weird. [23:14:03] OK, I'll see if I can reproduce [23:14:22] still doing a puppet run on drmf, we'll see if it survives. [23:18:40] physikerwelt: Are you selecting role::mediawiki-install::labs on the instance creation page? [23:18:57] yes [23:19:21] ok -- I suspect that that's the issue… that interface is a bit dumb since it encourages you to configure instances before their initial puppet run. [23:19:30] But I'll do some tests to verify that that's the issue. [23:19:41] Meanwhile… drmf seems happy, you can go ahead and delete test-puppet-cert. [23:20:42] ok the other idea I had in mind might be an error somwhere in my custom puppet group math [23:21:12] Generally you want to get a clean puppet run before you do any configuration work -- that way it can set up certs and accounts and such with out the custom complications. [23:23:07] ok I'll take care of the warning the next time (in the past that was not an issue) and with vagrant I also enabled roles prior to the first run... so sorry for that fault [23:24:12] I'm wonderding that math-test2 is the only instance with puppet status ok (according to the instance list) [23:24:15] Hm, actually I just made a new instance without any boxes checked and it has the same failure :( [23:24:36] Anything running self-hosted puppet will probably show up as 'unknown' [23:25:50] well, wait, actually maybe my new instance is fine -- too early to tell [23:25:53] Ok. I know about the self hosted instances... but I was expecting that drmf would get an ok [23:26:46] Is drmf still not reading 'ok'? [23:26:54] I'd think it would by now. [23:27:18] yes it is ok but in the instance list the status is unknown [23:27:36] I mean that list https://wikitech.wikimedia.org/wiki/Special:NovaInstance [23:32:03] ... and the role role::mediawiki-install::labs did not have any effect on the server [23:35:14] ok, let me catch up... [23:35:30] firstly, my test instance (with no boxes checked) came up ok. so that seems to answer that. [23:36:03] As for drmf not showing up 'ok', I'm not sure what that's about but I predict it'll stabilize in an hour or so. [23:36:13] Now, the question about role::mediawiki-install::labs... [23:36:15] lemme log in and see [23:36:58] can you force a manual puppet run on the test instance without error [23:37:08] yeah, just did [23:38:02] But, you're right, no apache. Hm... [23:38:34] for me the output of puppetd -tv looks really strange [23:39:09] I can not the the instance nor the cluster name (math) there [23:40:27] ok, I think that... [23:40:52] for math-test2 the correct information is displayed there... for drmf another project is printed out [23:40:59] your best bet is probably to burn down this instance, start a new one and make sure you get a clean puppet run before adding packages. Probably that will be more useful than trying to debug [23:41:19] another project is printed out… where? [23:42:50] physikerwelt: is the 'fresh start' approach OK with you? It sounds like you know how to do that... [23:42:59] when I ran puppetd-tv on drmf [23:43:02] sure [23:43:44] thanks. I'm willing to dig in but I think that'll get you what you need the fastest. [23:45:04] ok I think the wiered output was caused by your manual fix... so it doesn't matter [23:46:22] yeah, possible... [23:46:39] Will you be up for a while? I'm about to vanish but should be back later on. [23:46:39] how log should I wait before I start to configure something? [23:46:54] I'd give it 10 minutes, then log in and make sure you can do a clean puppet run before configuring. [23:47:20] ok thank you that's all I need to know [23:48:02] If this still doesn't work, then I'm interested! ping me and I'll get the message when I get back. [23:48:06] thank you very much for your help... I'll come back to you tomorrow if there are still problems (I don't expect further problems) [23:48:30] sure, no problem. [23:48:39] I'll have weird hours tomorrow too, but hopefully will be able to check in. [23:48:48] (traveling, visiting family, etc this week)