原文:http://liuyu.blog.51cto.com/183345/166381/
前言:
由于服务器apache抗不住目前的并发.加上前端squid配置后,问题依然无法解决.而页面程序大部分是动态.无法使用fastcgi来处理.因此想 使用nginx做为反向代理apache.整个配置安装过程很简单.在考虑高并发的情况下,在安装前就做了些优化.目前配置能抗住3000以上并发.好像 不是特别大哦?呵~~ 但足以~~ 只是还有少量499问题..期待有人跟我讨论解决
第1部分:安装
1 建立用户及组
/usr/sbin/groupadd www
/usr/sbin/useradd -g www www
2 安装pcre 让nginx支持rewrite 方便以后所需
wget ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/pcre-7.8.tar.gz
tar zxvf pcre-7.8.tar.gz
cd pcre-7.8/
./configure
make && make install
3 安装nginx
wget http://sysoev.ru/nginx/nginx-0.7.58.tar.gz
tar zxvf nginx-0.7.58.tar.gz
cd nginx-0.7.58/
./configure --user=www --group=www --prefix=/usr/local/webserver/nginx --with-http_stub_status_module --with-http_ssl_module --with-cc-opt='-O2' --with-cpu-opt=opteron
make && make install
# 注意上文中的--with-cc-opt='-O2' --with-cpu-opt=opteron 这是编译器优化,目前最常用的是-02 而不是3.后面对应CPU的型号,可参照:http://wiki.gentoo.tw/index.php/HOWTO_CFLAG
第2部分:配置及优化配置文件
1 nginx.conf 配置文件:
user www www;
worker_processes 4;
# [ debug | info | notice | warn | error | crit ]
error_log /usr/local/webserver/nginx/logs/nginx_error.log crit;
pid /usr/local/webserver/nginx/nginx.pid;
#Specifies the value for maximum file descriptors that can be opened by this process.
worker_rlimit_nofile 51200;
events
{
use epoll;
worker_connections 51200;
}
http
{
include mime.types;
default_type application/octet-stream;
source_charset GB2312;
server_names_hash_bucket_size 256;
client_header_buffer_size 256k;
large_client_header_buffers 4 256k;
#size limits
client_max_body_size 50m;
client_body_buffer_size 256k;
client_header_timeout 3m;
client_body_timeout 3m;
send_timeout 3m;
#参数都有所调整.目的是解决代理过程中出现的一些502 499错误
sendfile on;
tcp_nopush on;
keepalive_timeout 120; #参数加大,以解决做代理时502错误
tcp_nodelay on;
include vhosts/upstream.conf;
include vhosts/bbs.linuxtone.conf;
}
2 upstream.conf 配置文件(这也是做负载的配置方法)
upstream.conf
upstream bbs.linuxtone.com {
server 192.168.1.4:8099;
}
3 站点配置文件
bbs.linuxtone.conf
server
{
listen 80;
server_name bbs.linuxtone.conf;
charset GB2312;
index index.html index.htm;
root /date/wwwroot/linuxtone/;
location ~ ^/NginxStatus/ {
stub_status on;
access_log off;
}
location / {
root /date/wwwroot/linuxtone/;
proxy_redirect off ;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header REMOTE-HOST $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
client_max_body_size 50m;
client_body_buffer_size 256k;
proxy_connect_timeout 30;
proxy_send_timeout 30;
proxy_read_timeout 60;
proxy_buffer_size 256k;
proxy_buffers 4 256k;
proxy_busy_buffers_size 256k;
proxy_temp_file_write_size 256k;
proxy_next_upstream error timeout invalid_header http_500 http_503 http_404;
proxy_max_temp_file_size 128m;
proxy_pass http://bbs.linuxtone.com;
}
#参数都有所调整.目的是解决代理过程中出现的一些502 499错误
#Add expires header for static content
location ~* \.(jpg|jpeg|gif|png|swf)$ {
if (-f $request_filename) {
root /date/wwwroot/linuxtone/;
expires 1d;
break;
}
}
log_format access '$remote_addr - $remote_user [$time_local] "$request" '
'$status $body_bytes_sent "$http_referer" '
'"$http_user_agent" $http_x_forwarded_for';
access_log /exp/nginxlogs/bbs.linuxtone_access.log access;
}
注:第二种代理方式
nginx 处理下图片,html等静态的东西.其它动态由apache处理.因此apache也需要做一些参数调整.
设置图片等过期时间.缓解请求.
如果源与nginx在同一台机器建议使用如下方法:
location / {
proxy_pass http://192.168.1.4:8099/;
proxy_redirect default ;
}
针对不同的目录进行代理把下面的配置放到根目录代理的上面
location /linuxtone/ {
proxy_pass http://192.168.1.4:8099/linuxtone/;
proxy_redirect default ;
}
4 源配置
<VirtualHost 192.168.1.4:8099>
ServerAdmin liuyu105#gmail.com
DocumentRoot /date/wwwroot/linuxtone
ServerName bbs.linuxtone.com
ErrorLog logs/linuxtone_error_log
CustomLog "|/usr/local/sbin/cronolog logs/linuxtone_access_log.%Y%m%d" combined
</VirtualHost>
第3部分:源的优化
1 apache-mpm.conf
<IfModule mpm_prefork_module>
StartServers 15
MinSpareServers 15
MaxSpareServers 30
ServerLimit 2536
MaxClients 2048
MaxRequestsPerChild 1500
</IfModule>
2 apache-keepalive
Timeout 120 #与nginx的保持一至
KeepAlive On
MaxKeepAliveRequests 400
KeepAliveTimeout 7
第4部分:PHP的优化
优化一:将PHP由之前的xcache换成eaccelerator
1 安装
wget http://bart.eaccelerator.net/source/0.9.5.3/eaccelerator-0.9.5.3.tar.bz2
tar jxvf eaccelerator-0.9.5.3.tar.bz2
cd eaccelerator-0.9.5.3/
/usr/local/webserver/php/bin/phpize
./configure --enable-eaccelerator=shared --with-php-config=/usr/local/php5/bin/php-config
make
make install
注:PHP路径以安装为准!
2 配置
sed -i 's#extension_dir = "./"#extension_dir = "/usr/local/php5/lib/php/extensions/no-debug-non-zts-20060613/"\nextension = "memcache.so"\n#' /etc/php.ini
sed -i 's#output_buffering = Off#output_buffering = On#' /etc/php.ini
sed -i "s#; always_populate_raw_post_data = On#always_populate_raw_post_data = On#g" /etc/php.ini
配置eAccelerator加速PHP:
mkdir -p /usr/local/webserver/eaccelerator_cache
vi /etc/php.ini
按shift+g键跳到配置文件的最末尾,加上以下配置信息:
[eaccelerator]
zend_extension="/usr/local/php5/lib/php/extensions/no-debug-non-zts-20060613/eaccelerator.so"
eaccelerator.shm_size="128"
eaccelerator.cache_dir="/usr/local/webserver/eaccelerator_cache"
eaccelerator.enable="1"
eaccelerator.optimizer="1"
eaccelerator.check_mtime="1"
eaccelerator.debug="0"
eaccelerator.filter=""
eaccelerator.shm_max="0"
eaccelerator.shm_ttl="300"
eaccelerator.shm_prune_period="120"
eaccelerator.shm_only="0"
eaccelerator.compress="1"
eaccelerator.compress_level="9"
优化二:联系开发重新编译php减少php的模块.以减少php进程所占用内存数.这块尽管影响不大,但也有一定的作用.编译前也可以参照nginx的编译器优化方式安装.
第5部分:测试并启动nginx
ulimit -SHn 51200
/usr/local/webserver/nginx/sbin/nginx -t
/usr/local/webserver/nginx/sbin/nginx
第6部分:nginx日志切割脚本
#!/bin/bash
# This script run at 00:00
# The Nginx logs path
logs_path="/exp/nginxlogs/"
mkdir -p ${logs_path}$(date -d "yesterday" +"%Y")/$(date -d "yesterday" +"%m")/
mv ${logs_path}bbs.linuxtone_access.log ${logs_path}$(date -d "yesterday" +"%Y")/$(date -d "yesterday" +"%m")/bbs.linuxtone_access_$(date -d "yesterday" +"%Y%m%d").log
kill -USR1 `cat /usr/local/webserver/nginx/nginx.pid`
crontab -e
00 00 * * * /bin/bash /usr/local/webserver/nginx/sbin/cut_nginx_log.sh