Conversation
Notices
-
Embed this notice
we're back again for the 90th time today, I upgrade psql and did some other stuff maybe I can relax maybe I can stop looking at terminals
-
Embed this notice
@grumbulon maybe maybe
if you have any trouble you can always hmu I had to figure this shit out well enough to do manual db maintenance because I didn't have enough disk to run the automatic shizz
-
Embed this notice
@s8n I will if it keeps crashing I think I've got it but if it keeps fucking crashing I will def do that
-
Embed this notice
@pernia mine is actually cyan bigot
-
Embed this notice
@grumbulon look at the blackmirror (dysopia refernce [terminals are black])
-
Embed this notice
@grumbulon my shit was getting slammed by so many requests for threads I had to double my cpu power (core count was:2 now:4) to keep up. I was keeping it under control by reducing the pool sizes in the pleroma admin console. During busy hours plr was giving a lot of 500 errors because postgres just couldn't keep up with db reads
-
Embed this notice
@pernia @grumbulon cyan is light blue you dumb nigger
-
Embed this notice
@grumbulon what the hell is cyan what are you gay? Thats just pink but gayer ur a FAGGOT
-
Embed this notice
@pernia @cesare you are a child
-
Embed this notice
@cesare @grumbulon these are the only colors that arent made up:
blue green red yellow brown black white
everything else is gay bullshit
-
Embed this notice
@xue @pernia @cesare > tom scott
not a great source
-
Embed this notice
@grumbulon @pernia @cesare maybe, but he is right
tom scott told me that
-
Embed this notice
@s8n are there any negative effects of lowering the queue? I've dropped it pretty far down and my resource usage is quite low now but I don't actually know what these things do
-
Embed this notice
@grumbulon if the pools are too small then your instance will run out of oban runners and start failing to serve things like notifications or threads depending on which are all used-up. I was able to see things start to choke and die when I lowered them too much. I lowered them all by like 50% and then I lowered some of them more until things started eating shit and raised them back up until it kind of stopped. Eventually I gave up and threw more cpu at the problem and raised them all again until blink stopped complaining
-
Embed this notice
@s8n yeah I set them absurdly low just to see if it will prevent a crash, I plan on adjusting them slightly up from 5 tomorrow assuming it doesn't crash overnight
-
Embed this notice
@s8n yeah I've been tailing and grepping syslog for sev error for the last few hours, once it crashes again (assuming it will) I'll verify if it's the same as before I adjusted the queues or not
-
Embed this notice
@grumbulon you can set the logger to syslogd or whatever and set the log level to error if you want to catch the exception that's making it eat shit, but whoever wrote the error messages is a retard and you will have to sift through piles of retarded shit