Apache keeps going down on a Plesk server: server reached MaxRequestWorkers setting

Dmitry Bystrov

10 months ago

Symptoms

Websites on the server are experiencing 504 Gateway Time-out nginx errors consistently after working without issues for a period of time
The Apache web server goes down periodically with the following error message in Apache error log:
- on CentOS/RHEL-based distributions: /var/log/httpd/error_log
- on Debian/Ubuntu-based distributions: /var/log/apache2/error.log
  
  AH00484: server reached MaxRequestWorkers setting, consider raising the MaxRequestWorkers setting
  
  server is within MinSpareThreads of MaxRequestWorkers, consider raising the MaxRequestWorkers setting
The following error message is logged in domain's log at Plesk > Domains > Logs:
[error] 26927#0: *417 connect() failed (110: Connection timed out) while connecting to upstream, client: 203.0.113.2, server: example.com, request: "GET /?amp HTTP/1.1", upstream: "https://203.0.113.2/?amp", host: "example.com"

Cause

The MaxRequestWorkers limit set for the Apache web server has been reached and there are no free workers to serve new requests to Apache. As a result, requests are being served slowly or not served at all and therefore the websites become inaccessible, even though the server's CPU and RAM resources are not exhausted.

Resolution

Adjust the MaxRequestWorkers settings for Apache. The general formula for making the necessary calculation is the following:

# MaxRequestWorkers = (Total RAM - Memory used for Linux, DB, etc.) / average Apache process size

MPM Event: The default ServerLimit value is 16. To increase it, you must also raise MaxRequestWorkers using the following formula: ServerLimit value x 25 = MaxRequestWorkers value. For example, if ServerLimit is set to 20, then MaxRequestWorkers will be 20 x 25 = 500.
MPM Prefork: The default MaxRequestWorkers value is 256. To increase it, you must also raise ServerLimit.

Note: The below values are examples. They must be adjusted according to the server capabilities and load.

For CentOS/RHEL-based distribution

Find which Multi-Processing Module (MPM) is currently in use:

# httpd -V | grep MPM
Open the /etc/httpd/conf.modules.d/01-cgi.conf file in a text editor (for example, vi editor) and increase the values of the following directives (or add them if they are not present in the file):
- For MPM Event:
  
  <IfModule mpm_event_module>
  LoadModule cgid_module modules/mod_cgid.so
  MaxRequestWorkers 400
  ServerLimit 16
  </IfModule>
- For MPM Prefork:
  
  <IfModule mpm_prefork_module>
  LoadModule cgi_module modules/mod_cgi.so
  MaxRequestWorkers 400
  ServerLimit 400
  </IfModule>
Save the changes and close the file.
Restart Apache to apply the changes:

# service httpd restart

For Debian/Ubuntu-based distributions

Find which Multi-Processing Module (MPM) is currently in use:

# apache2ctl -V | grep MPM
Change the MPM configuration:
- For MPM Event:
  
  Open the /etc/apache2/mods-enabled/mpm_event.conf in a text editor (for example, vi editor) and increase the values of the following directives (or add them if they are not present in the file):
  
  MaxRequestWorkers 400
  ServerLimit 16
- For MPM Prefork:
  
  Open the /etc/apache2/mods-enabled/mpm_prefork.conf file in a text editor (for example, vi editor) and increase the values of the following directives (or add them if they are not present in the file):
  
  MaxRequestWorkers 400
  ServerLimit 400
Save the changes and close the file.
Restart Apache to apply the changes:

# service…

0 Shares