Author Topic: how to switch from mpm_event_module to mpm_prefork  (Read 204 times)

0 Members and 1 Guest are viewing this topic.

Offline
*
how to switch from mpm_event_module to mpm_prefork
« on: July 18, 2017, 04:36:26 PM »
Hi,
Ive got to solve this..

Installed 3 servers with  CWP version: 0.9.8.259 on Centos 6.9 2  unsing Apache 2.4.26 each server has 2 cores & 8 gb RAM (all the same)

Their main task: just webserver 1 site, mysql is installed in another server and Im using CloudFlares CDN and a LoadBalancer.. all works fine.. but im having this issue...

Initially Ive left the three servers with default configs, so each server only has apache and mpm_envent_module with default config..

Suddenly apache stop to work, the cpu usage drops near to 0% (when running normaly they are between 10-15% cpu use).. this happened on differnt times on the three servers..

so I stop apache service..(takes a minute or two to stop) and when try to start it just stands there... so I have no other choice to reboot server..and then return to normal..  all this from from ssh.. not from cwp.
 
Checking apache error logs. Ive spoted this "[mpm_event:error][pidxxxx]Script timeout before returning headers: index.php
[mpm_event:error]AH00484 server Reached MaxRequestsWorkers setting, Consider raising the MaxRequestsWorkes setting]

so, Ive uncoment '#' at the Include/conf/extra/httpd-mpm.conf at httpd.conf and then edited the httpd-mpm.conf.. there Ive set this values.. ThreadsPerChild 50 and MaxRequestsWorkers 800 (default was 400)

restarted apache.. hoping this might solve the issue.. but after a few hours got the same behaviour..

so, in only one of the servers Ive installed Ngnix and apache... hoping this solves it..

Installation went ok.. but now this server shows me a Bad Gateway (502) screen after 1 hour of running... and I have to reboot it..

Reading on the web mentioned that this could be a bug..of the mpm_event.. so decide to try with mpm_prefork.. but Im unable to find a way to enable it? how can I do that? Im just trying this only on one server, the one who has ngnix .. and if it fixes the issue then Ill update the others..

Can anyone assit? or give me a hint regarding my issue?


thanks!







if I try to access to the webpage located at the server I just got a blank page..eventually a cloudflares showsme that my server is down...