Online interviewer: php-fpm optimization summary

Nginx and php-fpm running process

  • Nginx view the nginx.conf configuration file
  • Load nginx's fast-cgi module
  • php-fpm listens on
  • php-fpm receives the request and enables the worker process to process the request
  • php-fpm processes the request and returns it to nginx
  • nginx returns the result to the browser via http

Nginx and php-fpm communication mechanism
route to
load nginx of fast-cgi module
fast-cgi Monitor address
        | request reached
     waiting to be processed...

The combination of Nginx and php-fpm
On Linux, nginx communicates with php-fpm in two ways: tcp socket and unix socket.
The advantage of tcp socket is that it can be used across servers, which can only be used when nginx and php-fpm are not on the same machine.
Unix socket is also called IPC(inter-process communication) socket, which is used to realize inter-process communication on the same host. In this way, you need to fill in the socket file location of php-fpm in the nginx configuration file.

The data transmission process of the two methods is shown in the following figure:

See these two articles for details on the two ways of communication transmission:
TCP protocol send/receive data implementation
UDP protocol send/receive data implementation

php-fpm process management

php-fpm adopts the process mode of master-worker. in,

  • The master is responsible for listening to the port and waiting for the link; secondly, the registration signal can be managed by the master through the information
  • The worker is responsible for handling the specific logic
    View php-fpm log information
  • Default path - /usr/local/php/var/log

php-fpm optimization

php.ini optimization

//By default, the server has a limit on the size of the uploaded file. If you want to modify the limit of the uploaded file, you can modify the php.ini file
file_uploads = On; //Whether to allow uploading of files 
upload_max_filesize = 1024M; //Maximum limit for uploading files

post_max_size = 1024M; //Most data submitted via post

max_execution_time = 300; //The longest script execution time in seconds
max_input_time = 30000; //Time limit for receiving submitted data in seconds
memory_limit = 256M; //Maximum memory used per script; in safe mode, you cannot change this setting at runtime with ini_set().

php-fpm.conf optimization

(1)Process number setting

 pm = dynamic

 pm.max_children = 15  //Number of php-fpm processes started in static mode

 pm.start_servers = 5    //Number of starting php-fpm processes in dynamic mode

 pm.min_spare_servers = 5 //The php-fpm process with the smallest idle time in dynamic mode

 pm.max_spare_servers = 5 //The php-fpm process with the largest idle time in dynamic mode

(2)Maximum number of requests processed

The maximum number of processing requests refers to a php-fpm of worker The process terminates after processing how many requests, master process will restart respawn new.  This configuration can avoid php Caused by the interpreter itself or by the program memory leaks.  
The default value is 500, pm.max_requests = 1024 Such a plan, 1 second 
Maximum number of requests: 15*1024=15360 Minimum number of requests: 5*1024=7120

How to avoid program hang

  • Method 1: Set the timeout period of php-fpm execution to a fixed value
    vi php-fpm.conf change into request_terminate_timeout = 60
  • Method 2: reload php-fpm regularly
    Periodically reload php-fpm on a server with high load
    Reload can restart smoothly without affecting the php script running of the production system, reload every 15 minutes
    0-59/15 * * * * /usr/local/php/sbin/php-fpm reload
  • Method 3: Optimize the process pool configuration
    php-fpm dynamically creates subprocesses to process requests according to the content of the configuration file and the actual situation. It can only block when the maximum value that can be created is reached. Execute one by one.
    • Process optimization
      pm = dynamic
      pm.max_children = 300
      pm.start_servers = 20
      pm.min_spare_servers = 5
      pm.max_spare_servers = 35
    • Maximum number of requests optimization
      pm.max_requests = 10240
      Tip: This is used to deal with memory leaks caused by the PHP parser or referenced third-party libraries.
      Maximum number of requests: Refers to how many requests a php-fpm worker process will terminate after processing.
    • Maximum execution time optimization (php.ini)
      request_terminate_timeout = 20
      Tip: This is used to deal with the 502 error reported because the PHP execution time is too long.
      This duration configuration can be configured in php.ini (max_execution_time) or php-fpm.conf. In order not to affect the global configuration, crontab timing tasks can be implemented in php-fpm.conf to restart php-fpm smoothly. This way It is to use the crontab timing task to regularly query whether the website is 502. If it is 502, restart php-fpm smoothly
      First create a script under the /root/ directory
      Then give execute permission to this script
      chmod +x /root/
      Then it is to write the content of the script, on the code
      RESULT=`curl -I $MY_URL|grep "HTTP/1.1 502"`
      if [ -n "$RESULT" ]; then
       /etc/init.d/php7.2-fpm restart
      Note: I restarted using /etc/init.d/php7.2-fpm restart
      Then write a scheduled task
      crontab -e
      Use crontab -l to view the list of scheduled tasks

Posted by 0p3n_p0rT on Thu, 02 Jun 2022 07:10:31 +0530