[07:34:24] <BurningPrincess1>	 Getting these errors when I try open my wiki: https://pastebin.com/FcJV0req and this is my localsettings php: https://pastebin.com/FFaTb1CF and I using pgp 7.4 and Ver 15.1 Distrib 10.1.47-MariaDB, for debian-linux-gnu (x86_64) using readline 5.2
[07:34:25] <BurningPrincess1>	  
[07:40:38] <Majavah>	 BurningPrincess1: which port is your mysql server running on?
[07:41:11] <BurningPrincess1>	 port 3306
[07:41:51] <Majavah>	 then $wgDBserver = "localhost:8889"; is your problem
[07:42:16] <Majavah>	 you've set mediawiki to try to use 8889 for mysql
[07:42:56] <BurningPrincess1>	 Ah, turns out I had to restart mysql, but thank you anyway
[14:25:04] <RhinosF1>	 tgr|away: what about https://www.mediawiki.org/wiki/Manual:$wgHTTPTimeout
[14:25:12] <RhinosF1>	 That's 30 total
[14:25:31] <tgr_>	 what about it?
[14:25:46] <RhinosF1>	 tgr_: would that also impact?
[14:26:52] <tgr_>	 if Commons is down, you can't possibly connect to it, right?
[14:27:07] <tgr_>	 I guess it could depend on the type of the outage
[14:28:05] <tgr_>	 in any case, that task would need more details on what happened on your side
[14:31:23] <RhinosF1>	 tgr_: my understanding is yesterday and definately the DDoS were connect timeout
[14:31:42] <RhinosF1>	 Like ERR_CONNECTION_TIMED_OUT
[14:32:05] <tgr_>	 that should happen after $wgHTTPConnectionTimeout seconds
[14:32:11] <tgr_>	 which is 5 by default
[14:32:42] <RhinosF1>	 Right
[14:34:27] <RhinosF1>	 Hang on
[14:34:35] <RhinosF1>	 Let me think, I'm tired and headachey
[14:35:19] <RhinosF1>	 You're connecting to the backend (commons api) fine during the DDoS and I think yesterday. It's just high latency.
[14:36:26] <tgr_>	 Ideally your logs should tell you what happened. PHP timeout due to slow queries? connection timeout? response timeout?
[14:37:59] <RhinosF1>	 It'll be response
[14:38:14] <RhinosF1>	 I can log into graylog tonight and try and check
[14:47:21] <RhinosF1>	 tgr_: curl error 28
[14:47:36] <RhinosF1>	 But varnish might be a factor too as that's a 30s timeout from mw* to cp*
[14:47:58] <RhinosF1>	 So yeah I think varnish is to blame because people randomly copy defaults
[15:28:11] <tgr_>	 if that would be the case you wouldn't get a connection timeout since varnish would open the connection, I imagine
[15:29:05] <tgr_>	 in any case 28 is connection timeout and that would happen after five seconds
[15:29:28] <tgr_>	 so either the page is opening lots of connections, or it cannot handle connection errors
[15:31:32] <RhinosF1>	 tgr_: it's healthchecks that don't wait longer than 30 seconds at fault
[15:31:40] <RhinosF1>	 Essentially, I need to lower the timeouts