Included features – from original website :
- Adaptive process spawning
- Basic statistics
- Advanced process management with graceful stop/start
- Ability to start workers with different uid/gid/chroot/environment and different php.ini
- Stdout & stderr logging
- Emergency restart in case of accidental opcode cache destruction
- Accelerated upload support
- Support for a “slowlog”
- Enhancements to FastCGI, such as fastcgi_finish_request() – a special function to finish request & flush all data while continuing to do something time-consuming
..and much more..
Notice :
PHP-FPM is not designed with virtual hosting in mind (large amounts of pools) however it can be adapted for any usage model.
Let’s start with installation (Ubuntu/Debian) :
sudo apt-get update
sudo apt-get install php5-fpm
# check if everything is working
php5-fpm -v
You can also install other required PHP packages :sudo apt-get install php5-common
sudo apt-get install php5-curl
sudo apt-get install php5-mysqli
...
Notice :This configuration is based on this Nginx test server.
We can now continue to configure PHP-FPM, but first try to make backup of your original www.conf file :
sudo nano /etc/php5/fpm/pool.d/www.conf
# pool name
[www1]
# user of FPM processess
user = www-data
# group of FPM processess
group = www-data
# address on which FastCGI requests are accepted, this can also be unix socket. Mandatory for each pool
# please take in mind that they say that unix socket is bit faster than tcp but honestly i didn't notice it, what i noticed is that on unix socket i get bunch of errors..
listen = "127.0.0.1:9000"
# how process manager controll number of children. Option is mandatory
# static - child processes are fixed - pm.max_children
# ondemand - processes spawn on demand, opposite to dynamic where started processes on service started
# dynamic - processes are spawned as said depending on directives : max_children, start_servers, min_spare_servers, max_spare_servers
pm = dynamic
pm.max_children = 4096
pm.start_servers = 20
pm.min_spare_servers = 5
pm.max_spare_servers = 128
# number of requests each child should execute before respawning.
pm.max_requests = 1024
# uri on which you can check your fpm status
pm.status_path = /php-status
# log file for slow requests
slowlog = /var/log/php-fpm.slow.log
listen.backlog = -1
After this done you can copy it and paste no more than 4 times, it’s
not good to have more than 4 pools because those configs are made for
heavy load web site (only one) : [www1]
user = www-data
group = www-data
listen = "127.0.0.1:9000"
pm = dynamic
pm.max_children = 4096
pm.start_servers = 20
pm.min_spare_servers = 5
pm.max_spare_servers = 128
pm.max_requests = 1024
pm.status_path = /php-status
slowlog = /var/log/php-fpm.slow.log
listen.backlog = -1
[www2]
user = www-data
group = www-data
listen = "127.0.0.1:9001"
pm = dynamic
pm.max_children = 4096
pm.start_servers = 20
pm.min_spare_servers = 5
pm.max_spare_servers = 128
pm.max_requests = 1024
pm.status_path = /php-status
slowlog = /var/log/php-fpm.slow.log
listen.backlog = -1
[www3]
user = www-data
group = www-data
listen = "127.0.0.1:9002"
pm = dynamic
pm.max_children = 4096
pm.start_servers = 20
pm.min_spare_servers = 5
pm.max_spare_servers = 128
pm.max_requests = 1024
pm.status_path = /php-status
slowlog = /var/log/php-fpm.slow.log
listen.backlog = -1
[www4]
user = www-data
group = www-data
listen = "127.0.0.1:9003"
pm = dynamic
pm.max_children = 4096
pm.start_servers = 20
pm.min_spare_servers = 5
pm.max_spare_servers = 128
pm.max_requests = 1024
pm.status_path = /php-status
slowlog = /var/log/php-fpm.slow.log
listen.backlog = -1
What i noticed, when i have only one pool it always brakes when i
have heavy load at the some moments but after few seconds it’s working
normally again, so i added more than one pool but as i said above, max 4
pools.After we added pools now we need to setup upstreams in Nginx config with sudo nano /etc/nginx/nginx.conf :
upstream php {
server 127.0.0.1:9100;
server 127.0.0.1:9001;
server 127.0.0.1:9002;
server 127.0.0.1:9003;
}
Those are our upstream FastCGI servers which will be used as round robin for processing requests.On Nginx vhost side, we will need change param for passing requests to upstream servers (pools) :
location ~ \.php$ {
fastcgi_pass php;
}
Now restart Nginx and PHP-FPM and you are done.Please take in mind that this guide is made for our test server and you will maybe need to play with those numbers, depending on your CPU and memory you can lower down max_children and max_spare_servers.
Happy hacking!
No comments:
Post a Comment