安装配置Varnish3.0手记_perl
 
在内网的vps上安装Varnish的最新版3.0来用做测试

Varnish有centos下的rpm和ubuntu下的deb二进制包下载,但我选择的是源码编译安装

先下载 wget http://repo.varnish-cache.org/source/varnish-3.0.0.tar.gz

tar -xzvf varnish-3.0.0.tar.gz

 

yum install gcc

yum install gcc-c++ libstdc++-devel

yum install -y httpd-devel pcre perl pcre-devel zlib zlib-devel GeoIP GeoIP-devel

 

Varnish需要pcre支持,如果服务器没有安装pcre-devel会提示需要指定prce目录。
Centos服务器上使用yum install pcre-devel安装

 

解压varnish进行编译安装

cd varnish-3.0.0

./autogen.sh

./configure --prefix=/usr/local/varnish PKG_CONFIG_PATH=/usr/lib/pkgconfig
 

make

make install

安装成功后,安装的目录是

cd /usr/local/varnish/sbin

./varnishd -V

查看版本,看是否安装成功

测试varnish

先编辑配置文件

vi /usr/local/varnish/etc/varnish/default.vcl

 

 

# This is a basic VCL configuration file for varnish.  See the vcl(7)
# man page for details on VCL syntax and semantics.
#
# Default backend definition.  Set this to point to your content
# server.
#
 backend default {
     .host = "192.168.88.156";
     .port = "80";
     ###下面三行为新加配
     .connect_timeout = 1s;
     .first_byte_timeout = 5s;
     .between_bytes_timeout = 2s;
 }
#
# Below is a commented-out copy of the default VCL logic.  If you
# redefine any of these subroutines, the built-in logic will be
# appended to your code.
 sub vcl_recv {
     if (req.restarts == 0) {
        if (req.http.x-forwarded-for) {
            set req.http.X-Forwarded-For =
                req.http.X-Forwarded-For + ", " + client.ip;
        } else {
            set req.http.X-Forwarded-For = client.ip;
        }
     }
     if (req.request != "GET" &&
       req.request != "HEAD" &&
       req.request != "PUT" &&
       req.request != "POST" &&
       req.request != "TRACE" &&
       req.request != "OPTIONS" &&
       req.request != "DELETE") {
         /* Non-RFC2616 or CONNECT which is weird. */
         return (pipe);
     }
     if (req.request != "GET" && req.request != "HEAD") {
         /* We only deal with GET and HEAD by default */
         return (pass);
     }
     if (req.http.Authorization || req.http.Cookie) {
         /* Not cacheable by default */
         return (pass);
     }
     return (lookup);
 }
#
 sub vcl_pipe {
#     # Note that only the first request to the backend will have
#     # X-Forwarded-For set.  If you use X-Forwarded-For and want to
#     # have it set for all requests, make sure to have:
#     # set bereq.http.connection = "close";
#     # here.  It is not set by default as it might break some broken web
#     # applications, like IIS with NTLM authentication.
     return (pipe);
 }
#
 sub vcl_pass {
     return (pass);
 }
#
 sub vcl_hash {
     hash_data(req.url);
     if (req.http.host) {
         hash_data(req.http.host);
     } else {
         hash_data(server.ip);
     }
     return (hash);
 }
#
 sub vcl_hit {
     return (deliver);
 }
#
 sub vcl_miss {
     return (fetch);
 }
#
 sub vcl_fetch {
     if (beresp.ttl <= 0s ||
         beresp.http.Set-Cookie ||
         beresp.http.Vary == "*") {
                /*
                 * Mark as "Hit-For-Pass" for the next 2 minutes
                 */
                set beresp.ttl = 120 s;
                return (hit_for_pass);
     }
     return (deliver);
 }
#
 sub vcl_deliver {
     return (deliver);
 }
#
# sub vcl_error {
#     set obj.http.Content-Type = "text/html; charset=utf-8";
#     set obj.http.Retry-After = "5";
#     synthetic {"
# <?xml version="1.0" encoding="utf-8"?>
# <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
#  "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
# <html>
#   <head>
#     <title>"} + obj.status + " " + obj.response + {"</title>
#   </head>
#   <body>
#     <h1>Error "} + obj.status + " " + obj.response + {"</h1>
#     <p>"} + obj.response + {"</p>
#     <h3>Guru Meditation:</h3>
#     <p>XID: "} + req.xid + {"</p>
#     <hr>
#     <p>Varnish cache server</p>
#   </body>
# </html>
# "};
#     return (deliver);
# }
#
 sub vcl_init {
        return (ok);
 }
#
 sub vcl_fini {
        return (ok);
 }

 

/usr/local/varnish/sbin/varnishd -f /usr/local/varnish/etc/varnish/default.vcl -s malloc,1024m -T 127.0.0.1:200 -a 0.0.0.0:80

 

现在可以打开varnish了

http://192.168.88.1/    (开启varnish的端口号)

启动varnishncsa用来将Varnish访问日志写入日志文件:

/usr/local/varnish/bin/varnishncsa -n /var/vcache -w /var/logs/varnish.log &



配置开机自动启动Varnish

vi /etc/rc.local


  在末尾增加以下内容:

引用
ulimit -SHn 51200
/usr/local/varnish/sbin/varnishd -n /var/vcache -f /usr/local/varnish/vcl.conf -a 0.0.0.0:80 -s file,/var/vcache/varnish_cache.data,1G -g www -u www -w 30000,51200,10 -T 127.0.0.1:3500 -p client_http11=on
/usr/local/varnish/bin/varnishncsa -n /var/vcache -w /var/logs/youvideo.log &



 优化Linux内核参数

vi /etc/sysctl.conf


  在末尾增加以下内容:

引用
net.ipv4.tcp_fin_timeout = 30
net.ipv4.tcp_keepalive_time = 300
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_tw_reuse = 1
net.ipv4.tcp_tw_recycle = 1
net.ipv4.ip_local_port_range = 5000    65000


 查看Varnish服务器连接数与命中率
/usr/local/varnish/bin/varnishstat

 

配置启动脚本: 

mkdir /cache/varnish/V -p

vi /usr/local/varnish/sbin/start.sh

 

#!/bin/sh
#file:start.sh
date -u
/usr/local/varnish/sbin/varnishd -a 0.0.0.0:80  -s file,/cache/varnish/V,1024m  -f         /usr/local/varnish/etc/varnish/default.vcl  -p thread_pool_max=1500  -p thread_pools=5  -p listen_depth=512 

chmod 777 start.sh

运行:

/usr/local/varnish/sbin/start.sh