[23:59:59] awjr: I don't need to do any of this today, so feel free to go ahead and do what you need to [00:00:31] maybe you could try pushing it and see if you have the same problem [00:00:47] AaronSchulz do you know anything about message cache on the cluster? [00:01:18] bits and pieces [00:02:32] ok well i was asked to make sure a non-en i18n message update was pushed out - i made sure the i18n changes were in master, got everything into the deployment branch and scap'd a while back. Special:Version shows the extension at the right sha1, but the message is not updated [00:02:51] i thought scap refreshed the message cache, so i figured that was all i needed to do [00:02:57] but that does not seem to be the case. [00:03:22] is this for an extension? [00:03:32] AaronSchulz yeah [00:03:44] this is the message: http://en.wikipedia.org/wiki/MediaWiki:Zero-rated-mobile-access-banner-text-grameenphone-bangladesh/bn [00:04:28] this is what it looks like currently in the i18n file: [00:04:28] 'zero-rated-mobile-access-banner-text-grameenphone-bangladesh' => '$1-অপেরা মিনি থেকে বাংলা উইকিপিডিয়া ব্রাউজিং ফ্রি', [00:05:08] why does that MediaWiki: page exist? [00:05:40] i believe all messages can be viewed that way [00:06:12] MediaWiki: [00:06:39] nevermind, I thought it had a history [00:08:29] kaldari i wonder if you couldn't commit that config change because you're not in the deployment gorup [00:10:31] AaronSchulz i wonder if just running scap again would resolve the issue [00:10:43] I should be in the deployment group [00:10:50] that's the 'mortals' list right? [00:11:43] dunno about mortals but i meant https://gerrit.wikimedia.org/r/#/admin/groups/21,members [00:12:27] bah, I don't have review rights on the configs :( [00:13:00] awjr: every editor I have shows that differently due to bidi ;) [00:13:27] bidi? [00:13:42] like the cigarettes? [00:14:19] oh bidi = bi directional [00:14:26] fml [00:14:52] rtl [00:14:59] awjr: I would touch the i18n file and rescap [00:15:08] not to be confused with "Ride the Lightning" by Metalica [00:15:49] AaronSchulz what's weird is i see other messages that were introduced in the latest commit [00:16:50] but i shall try the scap [00:18:14] kaldari: it's supposed to 102 tomorrow and 106 here on friday [00:18:20] i think you'd like it [00:18:56] nice [00:20:23] looks like no one's around to add me to the group :( [00:20:30] it would be cool if there were a way to update the localisation caches without scap'ing [00:20:32] :( [00:20:40] awjr: I see you're in the group though [00:21:05] iyam [00:21:23] any chance you can review https://gerrit.wikimedia.org/r/#/c/9490/ [00:21:34] yep [00:22:02] gah chrome is spazzing [00:23:00] kaldari: merged [00:35:21] AaronSchulz are you able to see cronjobs configured on fenari by chance? [00:36:07] I think anyone can see them, though a lot of them are probably on hume [00:36:16] oh [00:37:26] how would i view them on fenari? [00:37:35] the usual ways i know require r00t [00:38:18] I can see the l10update one in /etc/cron.d [00:39:16] oh hey sweet [00:39:33] although it is commented out [00:46:59] awjr: it was moved to puppet [00:47:10] see misc-servers.pp [00:48:12] AaronSchulz: bingo [00:48:14] thanks! [02:24:41] petan: ping [07:34:17] bahh [10:34:28] MaxSem: I have it [10:39:11] The power? [12:29:29] apergos: Around? [12:29:50] yes [12:29:53] multichill: [13:35:48] RoanKattouw: Someone installed Vector extension. Collasible tabs is broken, except on SpecialPreferences and SpecialLogin [13:36:01] Site/user script issue, right ? [13:38:08] Hah [13:38:17] Yeah almost certainly [13:38:42] Well I mean a site/user script is causing it, it may or may not be at /fault/ [13:38:53] it might just be exposing a bug in CT [13:40:17] CT has bugs? [13:41:51] lol [13:42:22] You could file a bug and have the reporter, assignee, and subject be the same ;-X [14:56:23] I've got to hide from Danny_B|backup now... [14:56:58] do you guys have a new license for PhpStorm? [14:57:10] the current license for the mediawiki project is going to expire in 18 days... [14:58:41] I've not applied for it yet, will get on that later [15:01:36] living w/o PhpStorm would be a disaster [15:03:04] yes... so... i'm just making sure it's on someone's todo list [15:03:14] * Daniel_WMDE looks at robla [15:04:31] some volunteer got that [15:04:39] or was it Reedy? [15:04:51] Meeee [15:05:05] I'll apply now [15:09:00] TimStarling: Is the reason you just mentioned regarding logging classes also the cause of the issue where stewards using their rights on local wikis to do global actions to not be able to go to the "global-foo" logs on Meta-Wiki? [15:09:07] (and instead end up in the local logs) [15:10:52] that's part of the reason I guess [15:11:40] basically most MediaWiki code uses wfGetDB(DB_MASTER) to get a connection for writing [15:11:58] LogPage is one example of that [15:12:12] okay, that makes sense [15:12:50] Could be interesting to try and experiment overwriting that from a subclass for the global* special pages that CentralAuth implements (or maybe even in core using a wiki-id) [15:13:02] LocalFile is one of the exceptions, it has a getMasterDB() function [15:13:16] so in theory you can do remote write operations with it [15:13:22] except that it calls LogPage [15:14:09] Daniel_WMDE, JeroenDeDauw: the invalidation stuff is basically just another job. the jobs just need to not be popped until all clients have picked up the job. if the JobQueue class was changed so that we can write into different queues, then we could switch out the database writes with queue writes, and your code would stay the same [15:14:12] Hm.. what's it used for (LocalFIle::getMasterDb) [15:14:20] Or just future proof [15:14:27] Ryan_Lane: it already has different queues [15:14:36] it does? [15:14:40] that's how we deliver emails in a timely way without having them wait for refreshLinks [15:15:37] It has exactly two queues, yes [15:15:44] This is implemented in jobs-loop.sh [15:26:52] hm. it isn't abstracted in a way that's usable for this, as far as I can tell [18:03:41] hackaton starts tommorow at 17:00? [18:34:57] question about MW session handling - when browsing as anon, on pages rendered by MobileFrontend, locally I get a consistent session id on every page view. in production, on the cluster, every mobile page view seems to give me a new session. why might that happen? [20:37:49] kaldari: Hi [20:38:05] on minute, finishing up a deployment... [20:38:14] ok [20:48:57] Reedy: around? [20:49:08] mmm [21:37:57] drecodeam: Howdy [21:38:04] deployment is finished [21:38:18] kaldari: in Berlin, all set for the hackathon tomorrow [21:38:25] excellent [21:38:37] BTW, I created a new remote branch for UploadWizard... [21:38:47] ya needed to talk to you about that onl [21:38:49] *only [21:39:39] the new branch in gerrit is called 'flickr' [21:39:52] checking it out [21:39:52] under the UploadWizard project [21:40:34] if you want, you can do your development under that branch and then we can create another one the geo features later [21:40:47] for the geo features... [21:40:56] https://gerrit.wikimedia.org/r/#/admin/projects/mediawiki/extensions/UploadWizard,branches [21:41:49] kaldari do I have access to it ? [21:41:59] good question [21:43:45] I think you should, but you may not be able to merge from it [21:43:47] Any devs/ops about who are involved in the IPv6 testing? There's some concerns being raised here: https://meta.wikimedia.org/wiki/User_talk:Eloquence#IPv6_rumors [21:44:01] I saw those [21:44:13] I'll poke Erik about that in the morning [21:45:21] RoanKattouw: back from zyankali? [21:45:46] I didn't go to Zyankali [21:46:09] error: insufficient permission for adding an object to repository database .git/objects [21:46:09] fatal: failed to write object [21:46:09] fatal: unpack-objects failed [21:46:11] drecodeam: You can try pull and submitting to it and let me know if you have any trouble. You'll probably want to edit your .gitreview file to change it to the default for the time being. [21:46:42] kaldari: you'll need to set/change your umask [21:46:55] drwxr-xr-x 2 kaldari wikidev 16 2012-05-31 20:22 fa [21:47:38] drecodeam: tomorrow is my 20% day, so I'll have more time to test it out then [21:47:54] kaldari: Yeah set your umask to 0002 [21:47:57] RoanKattouw: ^ fancy fixing? [21:48:00] We need to have a better way of handling umask stuff [21:48:03] probably easiest for you todo it as root [21:48:04] ^^ [21:48:07] Reedy: Path? [21:48:08] will do [21:48:08] indeed [21:48:21] /home/wikipedia/common/php-1.20wmf4/.git/objects [21:48:42] chmod g+w -R /home/wikipedia/common/php-1.20wmf4/.git/objects [21:48:43] kaldari: Alright. [21:48:44] I guess.. [21:49:10] drecodeam: any idea what your schedule will be like tomorrow? [21:49:10] done [21:49:54] thanks [21:55:24] kaldari: I am going on a sightseeing till 5pm and then I will be at the hackathon till 11pm according to the germany time [21:58:06] matanya: hello to far east [21:58:21] hi Danny_B|backup [22:02:57] hey hackers, what time is it in Berlin right now? [22:03:04] midnight! [22:03:16] OK, so that will work out well then [22:03:20] RoanKattouw: fancy fixing wmf-config too? I can't write to initialise/common settings... [22:03:24] +9 [22:05:06] Fixed [22:11:44] umask added to profile. Sorry if I locked everyone out of the files :P [22:12:02] hello RobH [22:12:36] heyas, just putting my laptop online, now gonna go pass out, jetlaggggg [22:12:39] ;] [22:12:44] * RobH is in berlin [22:12:56] RobH: you're not the only berliner ;-) [22:13:07] ya here? =] [22:13:16] not in mercure though [22:13:20] the main hotel filled up, had to overflow the ops folks to another hotel =[ [22:13:23] in novotel [23:38:23] Kaldari: If LastModified is already merged into trunk... we don't need a separate deployment window do we? [23:38:40] well, sort of... [23:39:05] we merged everything into truck and pushed to the wmf4 deployment branch... [23:39:56] this means the code is live on test.wiki and will get pushed out to en.wiki on Monday (without any effort from me). However, after it's pushed out, it still needs to be turned on in the configs, but that's all I'll need to do on Monday. [23:40:12] ok [23:40:25] we would have pushed it to en.wiki today, but scap was broken [23:40:40] which is one of the deployment tools [23:41:56] plus it let's us do more testing on test.wiki first