plesk 502 bad gateway

Vald

Hi everyone
I have plesk installed (nginx+php-fpm) on my cloud vps and I am running into a problem while testing the site load with blitz.io. After 80 hits the site returns 502 bad gateway, so I guess somewhere is set up to serve up this number but i can't find where to increase it.

I tried few options including adding this in the nginx config

proxy_buffer_size 128k;

proxy_buffers 4 256k;

proxy_busy_buffers_size 256k;

nothing works so far, maybe its the apache side but I don't know if this can happen from serving request to nginx.

Does anybody have a suggestion?

Bayu

Seems not nginx, but you can try to increase php-fpm pool setting

Vald

ok so for every site there is a config which says do not edit here, where would be the place to change these settings and which options should I increase?

; ATTENTION!

;

; DO NOT MODIFY THIS FILE BECAUSE IT WAS GENERATED AUTOMATICALLY,

; SO ALL YOUR CHANGES WILL BE LOST THE NEXT TIME THE FILE IS GENERATED.



; If you need to customize this file, use either custom PHP settings tab in
; Panel or override settings in /var/www/vhosts/system/mysite.com/conf/php.ini.

; To override pool configuration options, specify them in [php-fpm-pool-settings]

; section of /var/www/vhosts/system/mysite.com/conf/php.ini file.

[mysite.com]

; Don't override following options, they are relied upon by Plesk internally

prefix = /var/www/vhosts/system/$pool

user = mysite

group = mysite

listen = php-fpm.sock

listen.owner = myroot

listen.group = mygroup

listen.mode = 0660

; Following options can be overridden

chdir = /

; By default use ondemand spawning (this requires php-fpm >= 5.3.9)

pm = ondemand

pm.max_children = 5

pm.process_idle_timeout = 10s

; Following pm.* options are used only when 'pm = dynamic'

pm.start_servers = 1

pm.min_spare_servers = 1

pm.max_spare_servers = 1

; php.ini custom configuration directives

php_value[open_basedir] = "/var/www/vhosts/mysite.com/:/tmp/"

xDutchy

This would probably be a good place..

vladimir said:

; If you need to customize this file, use either custom PHP settings tab in
; Panel or override settings in /var/www/vhosts/system/mysite.com/conf/php.ini.
; To override pool configuration options, specify them in [php-fpm-pool-settings]
; section of /var/www/vhosts/system/mysite.com/conf/php.ini file.

Also; you might not want to run php-fpm as root

Vald

@xDutchy silly me for not reading the rest of it.

@xDutchy said:

>

Also; you might not want to run php-fpm as root

this was set by plesk not by me

xDutchy

vladimir said: this was set by plesk not by me

That's quite... curious?

I'd love to find a vurnerable WordPress install on such a Plesk server, having fun all day by uploading php scripts and running shell executes

Vald

whats more funny

specify them in [php-fpm-pool-settings] ; section of /var/www/vhosts/system/mysite.com/conf/php.ini file.

this file does not exist in the directory

Vald

this does not make me happy at all, going downhill after 82 concurrent users

forthcloud

@vladimir Did you check top and free -m when you were performing the test?

Vald

@forthcloud yes and there is only 3mb increase of memory, usually idles at 160mb of 1GB, i have 3 websites and only the one under load returns 502, the other 2 work fine.

forthcloud

@vladimir said:

Can you please check your Apache Logs and PHP-FPM logs?

The following commands should also help:

echo 'NGINX_ULIMIT="-n 4096" >> /etc/sysconfig/nginx
grep 16384 /etc/nginx/nginx.conf worker_rlimit_nofile 16384;
/etc/init.d/nginx restart
/usr/local/psa/admin/sbin/httpdmng --reconfigure-all

jar

plesk

Found the problem

forthcloud

@vladimir said:
this file does not exist in the directory

Replace the "mysite.com" with your own domain.

/var/www/vhosts/system/mysite.com/conf/php.ini

Vald

@forthcloud That was fine, i just replaced in the post so my domain stays private

however my nginx config file was lacking worker_rlimit_nofile option

I added it now and Im going to test again.

thanks

Vald

still same

nginx 502 Bad Gateway

and same message
The first timeout happened at 5 seconds into the test when the number of concurrent users was at 82.

I feel I am hitting in the wrong place all the time

forthcloud

@vladimir said:

What type of content are you hosting?

Vald

@forthcloud php sites

I know this is comming from plesk since I have worked on many enterprise solutions and I know what I am doing, the problems lies somewhere in plesk but where exactly... the configs are mixed up and scattered all over so Im having a hard time finding the right place.

one thing I know for sure now is that plesk isolates configs and processes for each site since only one site goes down and the others work fine.

forthcloud

@vladimir said:

Have you checked /var/log/sw-cp-server/error_log? If you can submit a ticket, we will log in to your server and fix it for you.

Vald

@forthcloud that log is empty, and although I am very grateful to you i ordered unmanaged service, i don't expect you to fix these issues for me, on the other hand I also want to learn how to fix problems my self :P

btw the nginx error log is pretty heavy - 30mb

2014/06/08 18:38:37 [error] 30933#0: *55067 connect() to unix:/var/www/vhosts/system/mysite.com/php-fpm.sock failed (11: Resource temporarily unavailable) while connecting to upstream, client: 54.***.**.9, server: mysite.com, request: "GET / HTTP/1.1", upstream: "fastcgi://unix:/var/www/vhosts/system/mysite.com/php-fpm.sock:", host: "mysite.com"

the problem lays somewhere here

forthcloud

@vladimir Try to edit /etc/nginx/nginx.conf file and add following value, then restart.

send_timeout 120;
proxy_read_timeout 120;
proxy_connect_timeout 120;

Vald

@forthcloud still same result

I also turned off nginx to serve php scripts by apache and the result is 82 users = 500 HTTP Internal Server Error, Instead of Bad Gateway

memory used 360mb

UPDATE: this guy has same problem http://forum.parallels.com/pda/index.php/t-289980.html

forthcloud

@vladimir said:

http://stackoverflow.com/questions/10003978/php-fpm-and-nginx-502-bad-gateway
http://stackoverflow.com/questions/8772015/502-gateway-errors-under-high-load-nginx-php-fpm

Vald

--UPDATE--

problem solved a while ago, just increase php-fpm max child processes.