Bug Bad Gateway and Site wont load correctly

[Nurti]

The site constantly gives me the error Bad Gateway, especially when I try to open a few new tabs from it (like one or two).
The new tabs get the error. Then I reload the maintab I have open for the site and I get the same error; I check site with isitdownrightnow.com and sometimes it says its down (in the last few days like five times), more often its online.
When I start a new thread and the site crashes twice when I click on save its getting annoying.

Then I reload it for like four times afterwards and it works again with luck or it looks like this and I have to reload again:
(sometimes a few pictures wont load either and I gotta reload again then they might show up or not)

Browsers tested: Mozilla Firefox and Microsoft Edge

Also when I go to another tab and then come back to the main tab I have open for this site
or start a new thread, click on preview or post thread I get this error popup very often:

Next up when I try to edit my thread the pictures which are in it wont load so its hard to remove them?, only when I go to preview they will load and there comes the above error popup again like ten times and also it seems BBcode doesnt always work: sometimes when I enter for example [i m g ] https://imagehere.png [/i m g ] (without the spaces) it only shows this in bold letters when I saved: "imagehere.png", then I edit it and put the same code inside again and now it works? Also something which is kinda irritating for me, why is there a "Delete Account" next to the "Logout", its wild.

 

[Checkmate]

There is a problem with NFS on kernel 4.x that I haven't yet to figure it out. NFS client will hang, and speed become absymal (4.0MB/s) between client and server, preventing the client to read the php files from the server, and cause a bad gateway timeout.

May have to move away from the centos forks (almalinux) altogether and jump ship to debian. As centos is all about stability, and I'm not getting that in almalinux 8.

So what happened is that NFS client sometimes hang, making nginx and php-fpm couldn't read the php files on the storage server anymore:
nginx reports error 11: resources unavailable
The mounting still works, but even a simple list (ls) command becomes slow.
The mounting is very slow when trying to navigate, read a file, write a file, even if it's a 0 bytes txt file
Restarting the mount, the nfs client, the nginx, or php-fpm will not solve the problem. The only solution is to restart the entire server (the client, not the nfs server).
This problem never existed in CentOS 7 which use kernel 3.x (EOL anyway, and won't work well with newer CPU). I will need more time to figure this one out.

 

[Yosef30]

Have same problems. But it's nice it's known about and hopefully can see a fix down the line. Never had issues with site before the move so it's making me miss it a bit.

 

[Checkmate]

Update:
This has been fixed. It is stable for >2 weeks.

We kindly request your understanding as we continue to work on optimizing the backend. As such, you may see this problem again in the future, occasionally, but it is unrelated to this NFS bug as written above.

We will regularly post optimization updates in the Updates sticky thread for transparency.