一、keepalived的工作原理
keepalived是以VRRP协议为实现基础的，VRRP全称Virtual Router Redundancy Protocol，即虚拟路由冗协议。
虚拟路由冗余协议，可以认为是实现路由器高可用的协议，即将N台提供相同功能的路由器组成一个虚拟路由器组，这个组里面有一个master和多个backup，master上面有一个对外提供服务的vip（该路由器所在局域网内其他机器的默认路由为该vip），master会发组播，当backup收不到vrrp包时就认为master出问题了，这时就需要根据VRRP的优先级从backup中选举出一个master。这样的话就可以保证路由器的高可用了。
keepalived主要有三个模块，分别是core、checkers和vrrp。core模块为keepalived的核心，负责主进程的启动、维护以及全局配置文件的加载和解析。checkers负责健康检查，包括常见的各种检查方式(tcp、http、ssl)。vrrp模块是来实现VRRP协议的。
三、keepalived双主模型高可用nginx服务
3.1、环境设置
keepalived-nginx1 ： 172.16.16.11
keepalived-nginx2 ： 172.16.16.12
vip1 ： 172.16.16.9
vip2: 172.16.16.10
upstream_server1 : 172.16.16.3
upstream_server2: 172.16.16.4
3.2、准备工作
（1）设置nginx1 ，nginx2 关闭selinux和iptables
sed -i ‘s@^SELINUX=.*@SELINUX=permissive@‘ /etc/selinux/config setenforce 0 iptables -F service iptables stop &> /dev/null chkconfig iptables off
（2）配置epel源，查看相关包所在yum源
[root@localhost ~]# vim /etc/resolv.conf //配置域名服务器 nameserver 172.16.0.1 [root@localhost ~]# yum list all keepalived keepalived.x86_64 1.2.13-4.el6 centos6.6 [root@localhost ~]# yum list all nginx nginx.x86_64 1.0.15-12.el6 epel
(3) 修改keepalived主备节点的主机名
sed -i ‘s@HOSTNAME=.*@HOSTNAME=nginx1@‘ /etc/sysconfig/network //设置主机名 sed -i ‘s@HOSTNAME=.*@HOSTNAME=nginx2@‘ /etc/sysconfig/network hostname nginx1 hostname nginx2
（4）主机互信
vim /etc/hosts //添加域名解析 172.16.16.11 nginx1 172.16.16.12 nginx2 172.16.16.3 web1.bengbengtu.com web1 172.16.16.4 web2.bengbengtu.com web2 [root@nginx1 ~]# yum install openssh-clients [root@nginx1 ~]# ssh-keygen -t rsa -P ‘‘ //生成一对密钥 Generating public/private rsa key pair. Enter file in which to save the key (/root/.ssh/id_rsa): Created directory ‘/root/.ssh‘. Your identification has been saved in /root/.ssh/id_rsa. Your public key has been saved in /root/.ssh/id_rsa.pub. The key fingerprint is: fa:14:8d:95:f3:48:bf:d1:b8:13:b4:dd:89:0e:51:97 root@nginx1 The key‘s randomart image is: +--[ RSA 2048]----+ | . ..| | .. .E | | =.. | | = *.=...| | S o.B.o..| | . . o= | | . . +. | | o . | | . | +-----------------+ [root@nginx1 ~]# ssh-copy-id -i .ssh/id_rsa.pub nginx2 //公钥发给nginx2 [root@nginx2 ~]# ssh-keygen -t rsa -P ‘‘ [root@nginx2 ~]# ssh-copy-id -i .ssh/id_rsa.pub nginx1 //公钥发给nginx1
（5）同步时间
# yum install ntpdate [root@nginx1 ~]# ntpdate ntp.sjtu.edu.cn //同步时间 [root@nginx1 ~]# hwclock -w [root@nginx1 ~]# date ; ssh nginx2 ‘date‘ //查看nginx1和nginx2时间是否同步 Thu Oct 15 22:44:40 CST 2015 Thu Oct 15 22:44:41 CST 2015
3.3、 安装并配置nginx
这里的nginx用做反向代理，并检查后端upstream的
[root@nginx1 ~]# yum -y install nginx ; ssh nginx2 ‘yum -y install nginx‘
#vim /etc/nginx/nginx.conf
worker_processes 2; //定义2个worker进程
upstream web { //定义上游服务器
server 172.16.16.3:80 weight=1 max_fails=2 fail_timeout=30s;
server 172.16.16.4:80 weight=2 max_fails=3 fail_timeout=40s;
}
#find /etc/nginx/conf.d/ -name ‘*.conf‘ -exec mv {} {}.bak \;
# vim /etc/nginx/conf.d/webserver.conf // 定义server
server {
listen 80;
server_name nginx1 nginx2;
location / {
proxy_pass //反向代理至后端的上游服务器
}
}3.4、后端的upstream_server安装httpd
web1.bengbengtu.com : 172.16.16.3
web2.bengbengtu.com : 172.16.16.4
只需安装httpd
（1)、设置主机名
# hostname web1.bengbengtu.com # hostname web2.bengbengtu.com
（2)、安装httpd服务
# yum install -y httpd [root@web1 ~]# echo "<h1> web1.bengbengtu.com - 172.16.16.3 </h1>" > /var/www/html/index.html [root@web1 ~]# echo "<h1> web2.bengbengtu.com - 172.16.16.4 </h1>" > /var/www/html/index.html #启动服务出现如下错误 Starting httpd: httpd: apr_sockaddr_info_get() failed for web2.bengbengtu.com httpd: Could not reliably determine the server‘s fully qualified domain name, using 127.0.0.1 for ServerName 解决方法： # vim /etc/httpd/conf/httpd.conf ServerName web1.bengbengtu.com ServerName web2.bengbengtu.com
先测试一下~~~
3.5、安装配置keepalived高可用nginx服务
说明：如果要监控nginx服务是否是在线状态，需要用到监控系统来实现nginx服务的重启操作！！
nginx1服务器上的keepalived配置如下：
[root@nginx1 ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs { //全局定义
notification_email {
root@localhost //给root发邮件
}
notification_email_from keepalive@localhost
smtp_server 127.0.0.1 //定义邮件服务器
smtp_connect_timeout 30 //连接邮件服务器的超时时间
router_id LVS_DEVEL //路由id
}
vrrp_instance VI_1 { //定义虚拟路由实例
state MASTER //主路由
interface eth0
virtual_router_id 235 //虚拟路由id
priority 100 //优先级
advert_int 1
authentication { //明文加密认证
auth_type PASS
auth_pass 2b316a978532
}
virtual_ipaddress { //定义vip1
172.16.16.9/16
}
}
vrrp_instance VI_2 {
state BACKUP //备路由
interface eth0
virtual_router_id 236
priority 99 //优先级
advert_int 1
authentication {
auth_type PASS
auth_pass 24985cea886c
}
virtual_ipaddress {
172.16.16.10/16 //vip2
}
}
nginx2服务器上的keepalived配置如下：
[root@nginx2 ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
notification_email {
root@localhost
}
notification_email_from keepalive@localhost
smtp_server 127.0.0.1 //定义邮件服务器
smtp_connect_timeout 30 //连接邮件服务器的超时时间
router_id LVS_DEVEL //路由id
}
vrrp_instance VI_1 {
state BACKUP //备路由
interface eth0
virtual_router_id 235
priority 99 //优先级
advert_int 1
authentication {
auth_type PASS
auth_pass 2b316a978532
}
virtual_ipaddress { //定义vip1
172.16.16.9/16
}
}
vrrp_instance VI_2 {
state MASTER //主路由
interface eth0
virtual_router_id 236 //虚拟路由id
priority 100 //优先级
advert_int 1
authentication { //明文认证
auth_type PASS
auth_pass 24985cea886c
}
virtual_ipaddress {
172.16.16.10/16 //定义vip2
}
}测试：
（1）停掉nginx1的keepalived，查nginx2，vip1，vip2都在；
[root@nginx1 ~]# service keepalived stop
测试完成！双主模型的高可用nginx介绍到此结束~~
本文出自 “bengbengtu” 博客，请务必保留此出处http://bengbengtu.blog.51cto.com/9505633/1703724
原文地址：http://bengbengtu.blog.51cto.com/9505633/1703724
