Installing and Configuring PHP-FPM¶
PHP-FPM is a service which executes PHP code on behalf of a web server. As NGINX is unable to natively execute PHP code, it can be onfigured to pass all .php files over to PHP-FPM.
Apache is able to execute PHP itself with mod_php, but from version 2.4+ you can now also pass .php files to PHP-FPM.
Installation¶
To install PHP-FPM use the following command.
yum install php-fpm
This will create the file /etc/php-fpm.d/www.conf
Using the enable --now flag, we can start the service and set it to start on boot all at once.
systemctl enable --now php-fpm
Configuration¶
Remote Server¶
By default, this configuration starts a PHP-FPM server listening on port 9000 that binds to 127.0.0.1 (localhost).
If your intention is to run PHP-FPMon a separate server, then you will need to amend the following values in your PHP-FPM domain configuration file.
listen = <add your internal IP address>:9000
listen.allowed_clients = 127.0.0.1,<add your remote server IP address>
You can then pass all your PHP requests over to this port on your dedicated PHP-FPM server.
Warning
Always be aware of the risks of opening additional firewall ports, and restrict to specific IP addresses where possible.
Local Server¶
If you’re running PHP-FPM on the same server as your web server, it’s a much better idea to configure a unix socket to pass requests to, instead of a port. This means the communication is handled by the kernel, instead being tied to any network TCP limits.
listen = '/var/run/php-fcgi-firstdomaincom.sock'
Example¶
Note
Creating a PHP-FPM pool per site allows for greater separation of resource, and makes it simpler to identify problematic sites using up excessive resource
Here is a sample configuration for a domain called domain.com.
[firstdomaincom]
listen = '/var/run/php-fcgi-firstdomaincom.sock'
;listen.backlog = -1
listen.allowed_clients = 127.0.0.1
listen.owner = apache
listen.group = apache
;listen.mode = 0666
user = firstdomainuser
group = firstdomainuser
pm = dynamic
pm.max_children = 300
pm.start_servers = 5
pm.min_spare_servers = 5
pm.max_spare_servers = 20
pm.max_requests = 2000
;pm.status_path = /status
;ping.path = /ping
;ping.response = pong
;request_terminate_timeout = 0
;request_slowlog_timeout = 0
slowlog = /var/www/vhosts/firstdomain.com/phpfpm-slow.log
;rlimit_files = 1024
;rlimit_core = 0
;chroot =
;chdir = /var/www
;catch_workers_output = yes
;env[HOSTNAME] = $HOSTNAME
;env[PATH] = /usr/local/bin:/usr/bin:/bin
;env[TMP] = /tmp
;env[TMPDIR] = /tmp
;env[TEMP] = /tmp
php_admin_value[error_log] = /var/www/vhosts/firstdomain.com/phpfpm-error.log
php_admin_flag[log_errors] = on
This is configured to work with Apache. If you were configuring this with NGINX, you would replace these lines.
listen.owner = nginx
listen.group = nginx
The PHP-FPM pool itself needs to run as its own Linux user.
user = firstdomainuser
group = firstdomainuser
You will then need to run a configuration test.
php-fpm -t
and then reload the configuration.
systemctl reload php-fpm
You should now be able to see the PHP-FPM socket.
stat /var/run/php-fcgi-firstdomaincom.sock
which should show something like.
File: '/var/run/php-fcgi-firstdomaincom.sock'
Size: 0 Blocks: 0 IO Block: 4096 socket
Device: 13h/19d Inode: 3108659 Links: 1
Access: (0666/srw-rw-rw-) Uid: ( 48/ apache) Gid: ( 48/ apache)
Access: 2016-08-10 13:15:26.968222080 +0100
Modify: 2016-08-10 13:15:26.968222080 +0100
Change: 2016-08-10 13:15:26.968222080 +0100
Birth: -
Note
The pm values in the configuration are there mainly to prevent the server from running out of memory. However, on servers with higher specification these values can be increased.