PHP-FPM Random 502 bad gateway error with no load at all on the server

Please fill in any relevant information that applies to you:

• CentOS Version: CentOS 7 64bit ?
• Centmin Mod Version Installed: 123.09beta01
• Nginx Version Installed: i.e. 1.11.10
• PHP Version Installed: i.e. 5.6.30 and 7.0.15
• MariaDB MySQL Version Installed: 10.1.21
• When was last time updated Centmin Mod code base ? : today

Hi

I am getting a lot of 502 errors

Server load is very low and i can't see anything to seem related and can cause that

---

I think there is a bug somewhere .....

Reading on the net i got some info and i don't know if something is related to Centminmod like:

The packages setup must be install on some order otherwise can cause that 502

They solve it by moving some code:

Some others recommend:

fastcgi_buffers 16 512k;
fastcgi_buffer_size 512k;

Other users recommend to use sockets for connection....

Another thing is to increase timeout (which one?)

Some nice info:

https://www.datadoghq.com/blog/nginx-502-bad-gateway-errors-php-fpm/
Nginx + PHP-FPM = 502 Bad Gateway

Sockets are faster than TCP (true?)

bug in pthreads ?


From my logs i can see this error:

Code:

tail -20 /home/nginx/domains/domain.com/log/error.log

Code:

upstream sent unexpected FastCGI record: 3 while reading response header from upstream

Any ideas?

Thank you

 

I am using the latest version 7.0.17 and 5.6.30 and i have on both servers this issue

I know that there was a related fix for phpfpm but i am using the latest versions......

Maybe related to aio threads that i use?

I really don't know why i am getting this the last 1-2 months

 

Don't know why but there was no erorrs at phpfpm log files and that makes me think that it may be related to Nginx...

And there the only thing that i have enable as extra is the aio threads....

I am thinking also that the issue may be related to:

Code:

try_files        $uri $uri/ /index.php;

as the 502 errors most of the times are related to the IPB admin panel that uses links like:

Code:

/forum/?page=1

and may need to use another option like:

Code:

try_files $uri $uri/ /index.php?args;

or

Code:

try_files $uri $uri/ /index.php?q=$request_uri;

or even:

Code:

try_files $uri $uri/ /index.php?q=$uri&$args;

What other ipb 4 users use? @Revenge , @Jimmy ?

Thanks

 

Thanks but i didn't find anything that helps

Maybe this one isn't patch yet for php 5.6.30 ?

PHP :: Bug #69860 :: php-fpm process accounting is broken with keepalive

but on the other hand i have that issue on latest php 7.0.17

 

Phpfpm and Nginx debug mode ON

I got some info in any case that anyone has an idea :

Code:

2017/03/20 20:09:06 [error] 8745#8745: *9489 upstream sent unexpected FastCGI record: 3 while reading response header from upstream, client: 131.161.11.204, server: domain.com, request: "GET /profile/125721-zach-binx/?tab=gallery HTTP/1.1", upstream: "fastcgi://127.0.0.1:9002", host: "www.domain.com"

2017/03/20 20:09:06 [warn] 8745#8745: *9489 upstream server temporarily disabled while reading response header from upstream, client: 131.161.11.204, server: domain.com, request: "GET /profile/155721-zach-binx/?tab=gallery HTTP/1.1", upstream: "fastcgi://127.0.0.1:9002", host: "www.domain.com"

The upstream server temporarily disabled while reading response header

 

I did but nothing.....

I will plan to downgrade phpfpm and nginx to older versions (2 months ago) that all was ok and see....

I am sure that it is specific to the server software as i can see the same errors on IPB, Wordpress, Opencart....

 

All problems solved by disabling multiple pools

I got tricked the first time when i was disabled them as i got some 502 but now i check and all are related to one simple think so maybe something block it as i thing Centminmod uses a kind of protection for that file as it was used for Ddos attacks:

Code:

"POST /xmlrpc.php?for=jetpack&token=ajPjE%40i03yH%21FYYKMoUa%23%2364jVM8Q%24d7%3A1%3A0&timestamp=1490292995&nonce=I3ymrBuB2N&body-hash=K1jPFZ%2BIik361eqqJintTEayeMg%3D&signature=x6W0BUMgDvB4px8WGKWxQJv5mtw%3D HTTP/1.1" 502 166 "https://www.domain.com/xmlrpc.php?for=jetpack&token=ajPjE%40i03yH%21FYYKMoUa%23%2364jVM8Q%24d7%3A1%3A0&timestamp=1490292995&nonce=I3ymrBuB2N&body-hash=K1jPFZ%2BIik361eqqJintTEayeMg%3D&signature=x6W0BUMgDvB4px8WGKWxQJv5mtw%3D" "Jetpack by WordPress.com" "192.0.100.89" rt=0.100 ua="127.0.0.1:9000" us="502" ut="0.100" ul="0" cs=-

No other 502 errors than that !

Do i have to edit something related to it?

Thank you

 

Don't remember if i was use menu 22 for this Wordpress site to be honest so is it normal to get that 502 error and should i ignore it?

 

I do not use anymore multiple pools and i got this one.