最後更新: 2018-01-22

介紹

php5-fpm - PHP FastCGI Process Manager

 * process spawning

Pool

不同的 Pool 是不同設定的 php process, 比如 php version, upload_max_filesize

Pool 是用不同的 port 去區分 (Default first pool: 9000/TCP)

Daemon

runs as a daemon and receives Fast/CGI requests.

Client --> Nginx --> php-fpm --> php

* 當 Nginx 連不到 php-fpm proccess 時, 那 client 會收到 "502 Bad Gateway"

Tools

/etc/init.d/php5-fpm

/usr/sbin/php5-fpm

目錄

php5-fpm

Opts:

• -t            Test FPM configuration file
• -i             PHP information and configuration
• -m           Show compiled in modules

php*-fpm 設定

php-fpm 主設定檔

/etc/php5/fpm/php-fpm.conf

    error_log = /var/log/php5-fpm.log
    log_level = warning
    process.max = 128
    # master process
    rlimit_files = 10240
    # child processes to wait for a reaction on signals from master
    process_control_timeout = 0
    include=/etc/php5/fpm/pool.d/*.conf

Pool 的設定

/etc/php5/fpm/pool.d/www.conf

# 這是 pool 的名, variable $pool 相當於這字
[www]

# 這 pool 的 proccess 的執行身分
user = www-data
group = www-data

# TCP
listen = 127.0.0.1:9000
listen.allowed_clients = 127.0.0.1
# backlog: maximum length to which the queue of pending connections
# man 2 listen 有解釋
listen.backlog = 128

# Socket
listen = /var/run/php5-fpm.sock
listen.owner = www-data
listen.group = www-data
listen.mode = 0660


# Proccess 數量設定
pm = dynamic

# alive at the same time (ApacheMaxClients)
# simultaneous requests that will be served.
pm.max_children = 32
# children created on startu
pm.start_servers = 4
# children in 'idle'
pm.min_spare_servers = 2
pm.max_spare_servers = 6
# idle process will be killed
pm.process_idle_timeout = 10s
# each child process should execute before respawning
pm.max_requests = 1024

# real-time FPM status monitoring
# Default Value: not set
# http://127.0.0.1/status Or http://www.foo.bar/status?full
pm.status_path = /status

# The access log file
; Default: not set
; Default Format: "%R - %u %t \"%m %r\" %s"
;  %R: remote IP address
;  %u: remote user
;  %t: server time
;  %m: request method
;  %r: the request URI
;  %s: status
access.log = /var/log/fpm-php/pool/$pool.access.log

; The log file for slow requests
slowlog = /var/log/fpm-php/pool/$pool.log.slow
request_slowlog_timeout = 10
request_terminate_timeout = 10m

# Limits the extensions of the main script FPM will allow to parse.
security.limit_extensions = .php .php3 .php4 .php5


# Additional php.ini defines, specific to this pool of workers.
php_admin_value[sendmail_path] = /usr/sbin/sendmail -t -i -f www@owncloud.datahunter.org
php_flag[display_errors] = off
php_admin_value[error_log] = /var/log/fpm-php/pool/$pool.log
php_admin_flag[log_errors] = on
php_admin_value[memory_limit] = 128M
php_admin_value[max_execution_time] = 180
php_admin_value[date.timezone] = Asia/Hong_Kong
php_admin_value[upload_max_filesize] = 15M
php_admin_value[post_max_size] = 20M
php_admin_value[short_open_tag] = on

LOG

/var/log/php-fpm.log

Nginx 設定

setting:

location ~* \.php$ {

  include fastcgi_params;

  fastcgi_param   SCRIPT_FILENAME    $document_root$fastcgi_script_name;
  fastcgi_param   SCRIPT_NAME        $fastcgi_script_name;

  # disallows the FPM to process files in the /uploads/ directory
  if ($uri !~ "^/uploads/") {
    # fastcgi_pass unix:/var/run/php5-fpm.sock;
    fastcgi_pass 127.0.0.1:9000;
  }
}

Status page config in Nginx

Setting:

location ~ ^/(status|ping)$ {
     access_log off;
     allow 127.0.0.1;
     deny all;
     include fastcgi_params;
     fastcgi_pass 127.0.0.1:9000;
}

查看 status

curl -k  https://owncloud.datahunter.org:8443/status

pool:                 www
process manager:      dynamic
start time:           03/May/2014:09:29:47 +0000   <-- # started or reloaded
start since:          110
accepted conn:        12
listen queue:         0
max listen queue:     0
listen queue len:     128
idle processes:       3
active processes:     1
total processes:      4
max active processes: 1
max children reached: 0                            <-- # 多少次到達了 children limit 

listen queue

The number of request in the queue of pending connections.

If this number is non-zero, then you better increase number of process FPM can spawn.

max listen queue

the maximum number of requests in the queue of pending connections since FPM has started

listen queue len

the size of the socket queue of pending connections

max active processes

the maximum number of active processes since FPM has started