[00:56:16] Hola kiero hacer algo para superarme [06:19:34] hmm I wonder if I can run screenconnect with linuxulator [06:19:46] and get rid of the debian jail [06:37:17] Remilia: sorry, what's the context? [06:39:49] legoktm[m]: I have this weird software called ConnectWise which uses its own fork of mono and only has it built for Linux [06:40:12] right now I run it in a bhyve VM [06:40:37] so I have been wondering if I could make a jail with linuxulator [06:49:25] Is that related to MediaWiki? [08:53:26] legoktm[m]: oh no, sorry, wrong channel :(( [08:53:43] legoktm[m]: my irc client switched the channel tabs around :( [08:54:30] thank you for pointing it out [09:05:25] Remilia: no worries, it happens :) was mostly curious/confused [09:08:21] legoktm[m]: at least it is not as bad as sometimes happened with irssi where I had muscle memory for alt-# and then during reconnects channels would swap around haha [09:08:45] :p [16:13:48] hi all. trying to move an existing mediawiki installation from postgres/apache to postgres/nginx, starting out with copying the web tree and connecting to the same db. Connections from the commandline works, but in the webui, I only get '[825c88976ec6a0fcb9a345dd] 2021-07-10 16:03:41: Fatal exception of type "LogicException"' and nothing in the logs. I've tried to turn on php logging without any more [16:13:55] help. Any ideas? [16:57:19] RoyK: in general https://www.mediawiki.org/wiki/Manual:How_to_debug may be helpful, add "$wgShowExceptionDetails = true;" to your LocalSettings.php to show the full error message [18:52:11] Last time I was here, I mentioned how the OATHAuth extension was causing users who had never set up a 2FA code to suddenly receive the prompts. I deleted the data from the table, re-enabled the extension, and asked people to "have at it." I received another complaint of an erroneous 2FA prompt yesterday, and when I checked the database, ID 0 was back in that table. I'm starting to think this may be related to CleanupUsersWithNoId not completing [18:52:12] successfully. We had some issues related to that when upgrading from 1.29.1 to 1.35.2. Is it possible that there are accounts which somehow get user ID 0 written? [18:52:54] Or maybe I should try playing with that maintenance script again? [19:30:57] RoyK: if you use pretty URLs, try disabling them temporarily, as nginx does not support .htaccess if I recall right [19:32:54] Remilia: I found an even easier solution - ditching nginx for now [19:33:18] apache isn't really the problem, modphp is [19:34:04] RoyK: php-fpm is the answer [19:34:33] I run 5 MW instances with Apache + php-fpm [19:34:40] (and PostgreSQL) [19:35:49] Remilia: I know - that's what I'm using [19:36:06] postgres was in there already, but with modphp - that shit hadn't been changed for ages [19:36:26] I generally only use mysql at gunpoint [19:36:56] MW had a lot of pgsql issues a relatively short while ago [19:37:46] none that I saw - I don't administer very large wikis [19:38:02] but I do prefer postgresql for most db stuff [19:38:04] RoyK: most issues were fixed by 1.34 [19:40:58] RoyK: T230618 as an example haha [19:40:59] T230618: ActiveUsersPager's new query logic breaks PostgreSQL compatibility - https://phabricator.wikimedia.org/T230618 [19:50:52] * RoyK mumbles something about QA [20:20:37] heh [20:32:47] QA? sounds like an exotic dessert ;-) [20:35:10] but in all fairness the last time I tested PGSQL it was fairly stable [23:21:22] I've been a MediaWiki user for over a decade. Just discovered the Lockdown extension and it's been so useful! If there's a dev over here, just wanted to say thanks! [23:44:17] KuatreroDelSur: that would be duesen_