nginx+keepalived实现nginx双主高可用的负载均衡

http://kling.blog.51cto.com/3320545/1253474

一、前言：

在互联网上面，网站为用户提供原始的内容访问，同时为用户提供交互操作。提供稳定可靠的服务，可以给用户带来良好的体验，保证用户的正常访问使用，在网站的可靠性方面，有很多的技术可以来提供，这类技术可以分为：

高可用：保证服务的可靠，稳定，实现故障的屏蔽，避免了单点故障。

高性能：多台服务器连接起来，处理一个复杂的计算问题。

负载均衡：将用户请求引导到后端多台服务器，实现服务器请求的负载。

我们将这类技术称之为集群负载均衡，可以提供负载均衡和高可用的有硬件和软件，软件方面有haproxy，lvs，keepalived，nginx，heartbeat，corosync等等，而这里我们采用的是nginx-keepalived来构建。

Nginx有很强的代理功能，但是一台nginx 就形成了单点，现在使用keepalived来解决这个问题，keepalived可以实现故障转移切换，实现后端的健康检查，前端的高可用，使网站故障记录大大降低，避免了单点故障造成网站无法访问的问题，确保了网站业务的正常运行。

二、Nginx+keepalived有两种配置方案：

2.1、Nginx+keepalived 主从配置

这种方案，使用一个vip地址，前端使用2台机器，一台做主，一台做备，但同时只有一台机器工作，另一台备份机器在主机器不出现故障的时候，永远处于浪费状态，对于服务器不多的网站，该方案不经济实惠，所以本次不予采用。

2.2、Nginx+keepalived 双主配置

这种方案，使用两个vip地址，前端使用2台机器，互为主备，同时有两台机器工作，当其中一台机器出现故障，两台机器的请求转移到一台机器负担，非常适合于当前架构环境，故本次采用此方案对网站进行高可用架构。

三、Nginx+keepalived 主从配置

3.1、Nginx+keepalived 主从配置详情请见http://kling.blog.51cto.com/3320545/1240359

这里不做重点介绍。

四、Ningx+Keepalived 双主配置

4.1、拓扑结构

4.2、测试环境如下：

系统：Ceentos 6.4 64位

前端node1服务器：

DIP: 192.168.122.2

VIP: 192.168.122.22

前端node2服务器：

DIP: 192.168.122.3

VIP：192.168.122.23

后端服务器：

web server01：192.168.122.4

web server02：192.168.122.5

web server03：192.168.122.6

4.3、软件安装

分别在两台前端服务器上安装nginx+keepalived，使用脚本如下：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

#!/bin/bash # author: kuangl # mail: kuangl@orient-media.com # description: The installation of Nginx files. # -------------------------------------------------------- #          ## Nginx_install # -------------------------------------------------------- # # Nginx installation #CURRENT_PATH=$(pwd) for i in $(rpm -q gcc gcc-c++ kernel-devel openssl-devel zlib-devel popt-devel popt-static libnl-devel wget make |grep 'not installed' | awk '{print $2}') do     yum -y install $i done [ -d /root/software ] [ "$?" != 0 ] && mkdir /root/software cd /root/software [ !  -e pcre-8.33.tar.gz ] && wget ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/pcre-8.33.tar.gz tar -zxvf pcre-8.33.tar.gz cd pcre-8.33 ./configure make && make install echo $? || [ $? != 0  ] || echo  " installation pcre  failed" || exit 1 cd /root/software [ ! -e nginx-1.2.9.tar.gz ] && wget http://nginx.org/download/nginx-1.2.9.tar.gz tar -zxvf nginx-1.2.9.tar.gz cd nginx-1.2.9 ./configure  --prefix=/usr/local/nginx --with-http_ssl_module --with-http_sub_module --with-http_stub_status_module  --with-http_gzip_static_module make && make install echo $? || [ $? != 0  ] || echo  " installation  nginx  failed" || exit 1 # -------------------------------------------------------- #             ## Keepalived_intsall # -------------------------------------------------------- # # Keepalived installation cd /root/softwarae [ ! -e keepalived-1.2.4.tar.gz ] &&  wget http://www.keepalived.org/software/keepalived-1.2.4.tar.gz tar -zxvf keepalived-1.2.4.tar.gz cd keepalived-1.2.4 ln -s /usr/src/kernels/$(uname -r) /usr/src/kernels/linux ./configure --prefix=/usr  --bindir=/usr/bin  --sbindir=/usr/bin  --libexecdir=/usr/libexec --localstatedir=/var --libdir=/lib64  --infodir=/usr/share/info  --sysconfdir=/etc --mandir=/usr/local/share/man   --with-kernel-dir=/usr/src/kernels/linux make && make install echo $? || [ $? != 0  ] || print " installation keepalived  failed" || exit 1 chkconfig --add keepalived chkconfig --level 345 keepalived on

4.4、在后端服务器上安装apached

后端node4

1 2 3 4 5

[root@node4 ~]# yum  -y install httpd [root@node4 html]# echo "this is 192.168.122.4" > /var/www/htmlindex.html [root@node4 ~]# service httpd start [root@node4 html]# curl  192.168.122.4 this is 192.168.122.4

后端node5

1 2 3 4 5

[root@node5 ~]# yum  -y install httpd [root@node5 html]# echo "this is 192.168.122.5" > /var/www/htmlindex.html [root@node5 ~]# service httpd start [root@node5 html]# curl  192.168.122.5 this is 192.168.122.5

后端node6

1 2 3 4 5

[root@node6 ~]# yum  -y install httpd [root@node6 html]# echo "this is 192.168.122.6" > /var/www/htmlindex.html [root@node6 ~]# service httpd start [root@node6 html]# curl  192.168.122.6 this is 192.168.122.6

4.5、node2、node3上配置nginx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

[root@node2 ~]# vim /usr/local/nginx/conf/nginx.conf upstream web1       ##定义负载均衡组为web1     {         ip_hash;         server 192.168.122.4:80;         server 192.168.122.5:80;         server 192.168.122.6:80;     }  server {         listen       80;         server_name  dev.test01.com;         location /         {         root /home/kuangl/;         index index.html index.htm;         proxy_set_header Host $host;         proxy_set_header X-Forwarded-For $remote_addr;         proxy_pass http://web1;         }       }

4.6、在node2上配置keepalived

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

[root@node2 conf]# vim /etc/keepalived/keepalived.conf ! Configuration File for keepalived global_defs {    notification_email {      404060945@qq.com    }    notification_email_from root@localhost    smtp_server 127.0.0.1    smtp_connect_timeout 30    router_id LVS_DEVEL } vrrp_script chk_haproxy {     script "/etc/keepalived/chk_nginx.sh"     interval 2     weight 2 } vrrp_instance VI_1 {     state MASTER     interface eth0     virtual_router_id 200     priority 250     advert_int 1     authentication {         auth_type PASS         auth_pass kuanglnginx     }    track_script {         chk_nginx     }     virtual_ipaddress {         192.168.122.22     } } vrrp_instance VI_2 {     state BACKUP     interface eth0     virtual_router_id 251     priority 100     advert_int 1     authentication {         auth_type PASS         auth_pass kuangl     }     track_script {         chk_nginx     }     virtual_ipaddress {         192.168.122.23     } }

4.7、在node3上配置keepalived

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

! Configuration File for keepalived global_defs {    notification_email {      404060945@qq.com    }    notification_email_from root@localhost    smtp_server 127.0.0.1    smtp_connect_timeout 30    router_id LVS_DEVEL } vrrp_script chk_haproxy {     script "/etc/keepalived/chk_nginx.sh"     interval 2     weight 2 } vrrp_instance VI_1 {     state BACKUP     interface eth0     virtual_router_id 200     priority 100     advert_int 1     authentication {         auth_type PASS         auth_pass kuanglnginx     }     track_script {         chk_nginx     }     virtual_ipaddress {         192.168.122.22     } } vrrp_instance VI_2 {     state MASTER     interface eth0     virtual_router_id 251     priority 250     advert_int 1     authentication {         auth_type PASS         auth_pass kuangl     }     track_script {         chk_nginx     }     virtual_ipaddress {         192.168.122.23     } }

4.8、在两台双主服务器上添加自动检测脚本

1 2 3 4 5 6 7 8 9 10 11 12

#!/bin/bash # description: # 定时查看nginx是否存在，如果不存在则启动nginx # 如果启动失败，则停止keepalived status=$(ps -C nginx --no-heading|wc -l) if [ "${status}" = "0" ]; then         /usr/local/nginx/sbin/nginx         status2=$(ps -C nginx --no-heading|wc -l)         if [ "${status2}" = "0"  ]; then                 /etc/init.d/keepalived stop         fi fi

4.9、开启nginx、keepalived服务

1 2 3 4

[root@node2 ~]# service keepalived start [root@node2 ~]# /usr/local/nginx/sbin/nginx [root@node3 ~]# service keepalived start [root@node3 ~]# /usr/local/nginx/sbin/nginx

4.10、用 ip a 查看VIP

4.11、测试访问

1 2 3 4 5 6 7 8 9 10 11 12

[kuangl@node01 ~]$ curl http://192.168.122.22 this is 192.168.122.6 [kuangl@node01 ~]$ curl http://192.168.122.22 this is 192.168.122.4 [kuangl@node01 ~]$ curl http://192.168.122.22 this is 192.168.122.5 [kuangl@node01 ~]$ curl http://192.168.122.23 this is 192.168.122.6 [kuangl@node01 ~]$ curl http://192.168.122.23 this is 192.168.122.4 [kuangl@node01 ~]$ curl http://192.168.122.23 this is 192.168.122.5

五、后端用rsync做数据同步

node5-node6上配置进程模式，以node5为例

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

[root@node5 ~]# yum -y install rsync [root@node5 ~]# vim /etc/rsynsd.conf uid = root gid = root use chroot = no max connections = 5 pid file = /var/run/rsyncd.pid lock file = /var/run/rsync.lock log file = /var/log/rsyncd.log [web01]                       path=/home/kuangl/           comment = update           ignore errors               read only = no              list = no                  hosts allow = 192.168.122.0/24 auth users = root         uid = root gid = root secrets file = /etc/rsyncd.secrets [root@node5 ~]# vim /etc/rsyncd.secrets root:123456 [root@node5 ~]# chmod 0600 /etc/rsyncd.secrets [root@node5 ~]# ll /etc/rsyncd.secrets -rw-------. 1 root root 12 Jul 20 19:41 /etc/rsyncd.secrets [root@node5 ~]# rsync --daemon [root@node5 ~]# echo "rsync --daemon" >> /etc/rc.local

node4上配置命令模式：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

[root@node4 ~]# yum -y install rsync [root@node4 ~]# vim /etc/rsyncd.secrets 123456 [root@node4 ~]# chmod 0600 /etc/rsyncd.secrets root@node4 kuangl]# rsync -vzrtopg --delete --progress --password-file=/etc/rsyncd.secrets  rsync+inotify root@192.168.122.5::web01 sending incremental file list rsync+inotify/ rsync+inotify/inotify-tools-3.14.tar.gz       358772 100%    1.85MB/s    0:00:00 (xfer#1, to-check=2/4) rsync+inotify/rsync+inotify_client.sh          617 100%    3.11kB/s    0:00:00 (xfer#2, to-check=1/4) rsync+inotify/rsync+inotify_server.sh          900 100%    4.03kB/s    0:00:00 (xfer#3, to-check=0/4) sent 360679 bytes  received 69 bytes  240498.67 bytes/sec total size is 360289  speedup is 1.00

查看结果

1 2 3 4 5

[root@node5 ~]# cd /home/kuangl/ [root@node5 kuangl]# ll total 8 -rw-r--r--. 1 root root   22 Jul 20 15:16 index.html drwxr-xr-x. 2 root root 4096 Nov 11  2012 rsync+inotify

本文出自 “&思远晨曦” 博客，请务必保留此出处http://kling.blog.51cto.com/3320545/1253474