Nginx+Keepalived双机热备（主主模式）
Keepalived介绍
Keepalived是一个基于VRRP协议来实现的服务高可用方案，可以利用其来避免IP单点故障，类似的工具还有heartbeat、corosync、pacemaker。但是它一般不会单独出现，而是与其它负载均衡技术（如lvs、haproxy、nginx）一起工作来达到集群的高可用。
Keepalived的 VRRP协议将两台或多台路由器设备虚拟成一个虚拟路由器，对外提供虚拟路由器IP(一个或多个)；
而在路由器组内部，如果实际拥有这个对外IP的路由器如果工作正常的话就是MASTER，或者是通过算法选举产生；
MASTER实现针对虚拟路由器IP的各种网络功能，如ARP请求，ICMP，以及数据的转发等；
其他设备不拥有该IP，状态是BACKUP，除了接收MASTER的VRRP状态通告信息外，不执行对外的网络功能。
当主机失效时，BACKUP将接管原先MASTER的网络功能。
双机高可用有两种：
1）双机主从模式：即前端使用两台服务器，一台主服务器和一台热备服务器，正常情况下，主服务器绑定一个公网虚拟IP，提供负载均衡服务，热备服务器处于空闲状态；当主服务器发生故障时，热备服务器接管主服务器的公网虚拟IP，提供负载均衡服务；但是热备服务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。
2）双机主主模式：即前端使用两台负载均衡服务器，互为主备，且都处于活动状态，同时各自绑定一个公网虚拟IP，提供负载均衡服务；当其中一台发生故障时，另一台接管发生故障服务器的公网虚拟IP（这时由非故障机器一台负担所有的请求）。这种方案，经济实惠，非常适合于当前架构环境。
本文将搭建Nginx+Keepalived实现高可用的双机主主模式
一、环境说明
操作系统：centos7 64位
软件源：阿里云
2台服务器（xhk1、xhk2）安装keepalived和安装nginx来反向代理
2台服务器（nginx1、nginx2）安装nginx来提供服务
服务器的防火墙和selinux全部关闭
xhk1 IP：192.168.163.162
xhk2 IP：192.168.163.163
nginx1 IP：192.168.163.164
nginx2 IP：192.168.163.165
虚拟IP：192.168.163.100和192.168.163.200
拓扑图如下：
二、环境安装
首先安装nginx，每台服务器都要安装，前端2台服务器的nginx用作反向代理，后端2台服务器提供web服务。
安装依赖包：
[root@xhk1 src]# yum install -y gcc make pcre-devel zlib-devel openssl-devel
去官网下载nginx软件包，网址为http:// http://nginx.org/
[root@xhk1 ~]# cd /usr/local/src/
[root@xhk1 ~]# wget http://nginx.org/download/nginx-1.13.6.tar.gz
解压到当前路径：
[root@xhk1 src]# tar xf nginx-1.13.6
进入到解压好的目录，编译安装：
[root@xhk1src]#./configure --user=nginx --group=nginx --prefix=/usr/local/nginx --with-http_stub_status_module --with-http_ssl_module --with-pcre
[root@xhk1 src]# make && make install
创建nginx用户，并启动nginx服务：
[root@xhk1 src]# useradd nginx
[root@xhk1 src]# /usr/local/nginx/sbin/nginx
至此，nginx的安装完成，其余3台服务器也按照同样的操作安装nginx
接下来为后端2台web服务器创建简单的首页
[root@nginx1 ~]# echo “nginx1” > /usr/local/nginx/html/index.html
[root@nginx2 ~]# echo “nginx2” > /usr/local/nginx/html/index.html
进行简单的访问测试：
[root@xhk1 ~]# curl 192.168.163.164
nginx1
[root@xhk1 ~]# curl 192.168.163.165
nginx2
接下来，就是要在前端的2台服务器做nginx的反向代理
编辑nginx的配置文件：
[root@xhk1 ~]#vim /usr/local/nginx/conf/nginx.conf
添加upstream模块：（注意该模块要在http模块里，server模块外）
upstream backend {
server 192.168.163.164 max_fails=3 fail_timeout=10s;
server 192.168.163.165 max_fails=3 fail_timeout=10s;
}
每个设备的状态设置为:
1.down 表示单前的server暂时不参与负载
2.weight 默认为1.weight越大，负载的权重就越大。
3.max_fails ：允许请求失败的次数默认为1.当超过最大次数时，返回proxy_next_upstream 模块定义的错误
4.fail_timeout:max_fails次失败后，暂停的时间。
5.backup：其它所有的非backup机器down或者忙的时候，请求backup机器。所以这台机器压力会最轻。
在server里添加一个location模块：
location / {
proxy_pass http://backend;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
}
proxy_set_header指令就是该模块需要读取的配置文件。在这里，所有设置的值的含义和http请求同中的含义完全相同，除了Host外还有X-Forward-For。
Host的含义是表明请求的主机名，因为nginx作为反向代理使用，而如果后端真是的服务器设置有类似防盗链或者根据http请求头中的host字段来进行路由或判断功能的话，如果反向代理层的nginx不重写请求头中的host字段，将会导致请求失败【默认反向代理服务器会向后端真实服务器发送请求，并且请求头中的host字段应为proxy_pass指令设置的服务器】。
保存退出并确认配置文件没错之后，重启nginx服务
[root@xhk1 ~]# /usr/local/nginx/sbin/nginx -t
nginx: theconfiguration file/usr/local/nginx/conf/nginx.conf syntax is ok
nginx:configuration file/usr/local/nginx/conf/nginx.conf test is successful
[root@xhk1 ~]# /usr/local/nginx/sbin/nginx -s reload
检测是否实现负载均衡：
[root@xhk1 ~]# curl192.168.163.162
nginx1
[root@xhk1 ~]# curl 192.168.163.162
nginx2
[root@xhk1 ~]# curl 192.168.163.162
nginx1
[root@xhk1 ~]# curl 192.168.163.162
nginx2
可以发现，每次访问时，前端服务器会将请求均衡地转到后端服务器
在另一台反向代理服务器也进行同样的操作配置nginx反向代理
接下来为前端的2台服务器安装keepalived
[root@xhk1 ~]# yum install -y keepalived
[root@xhk2 ~]# yum install -y keepalived
编辑keepalived文件：
为了实现双机主主模式，我们需要创建2个VIP
先编辑第一台（xhk1）的keepalived配置文件：
[root@xhk1 ~]#vim /etc/keepalived/keepalived.conf
global_defs {
notification_email { #邮件接收者
root@localhost
}
notification_email_fromkeepalived@localhost #邮件发送者
smtp_server 127.0.0.1 #邮件服务器地址
smtp_connect_timeout 30 #超时时限
router_id xhk #此处注意router_id为负载均衡标识
vrrp_mcast_group4 224.0.100.19 #添加ipv4的多播地址，并确保多播功能启用
}
vrrp_script check_nginx {
script /root/1.sh #检测脚本
interval 2 #每隔2秒钟检测一次
weight -2 #优先级减去2
}
vrrp_instance VI_1 {
state MASTER #状态只有MASTER和BACKUP两种
interface ens32 #网络接口
virtual_router_id 51 #虚拟路由标识，同一个instance的MASTER和BACKUP一致的
priority 100 #优先级，同一个instance的MASTER优先级必须比BACKUP高
advert_int 1 #MASTER 与 BACKUP之间同步检查的时间间隔，单位为秒
authentica0tion {
auth_type PASS #验证authentication。包含验证类型和验证密码
auth_pass xhk #AH使用时有问题。验证密码为明文
}
virtual_ipaddress {
192.168.163.100 dev ens32 #虚拟ip地址,可以有多个地址
}
track_script { #运行上面所写的检测脚本
check_nginx
}
}
vrrp_instance VI_2 {
state BACKUP
interface ens32
virtual_router_id 52
priority 99
nopreempt
advert_int 1
authentication {
auth_type PASS
auth_pass 123456
}
virtual_ipaddress {
192.168.163.200 dev ens32
}
track_script {
check_nginx
}
}
再去编辑第二台服务器（xhk2）的keepalived配置文件：
[root@xhk2 ~]# vim /etc/keepalived/keepalived.conf
global_defs {
notification_email {
root@localhost
}
notification_email_fromkeepalived@localhost
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id xhk
vrrp_mcast_group4 224.0.100.19
}
vrrp_script check_nginx {
script"/root/1.sh"
interval 2
weight -2
}
vrrp_instance VI_1 {
state BACKUP
interface ens32
virtual_router_id 51
priority 99
advert_int 1
authentication {
auth_type PASS
auth_pass xhk
}
virtual_ipaddress {
192.168.163.100 dev ens32
}
track_script {
check_nginx
}
}
vrrp_instance VI_2 {
state MASTER
interface ens32
virtual_router_id 52
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 123456
}
virtual_ipaddress {
192.168.163.200 dev ens32
}
track_script {
check_nginx
}
}
附上检测nginx反向代理的脚本：
[root@xhk1 ~]# vim /root/1.sh
#!/bin/bash ps -C nginx -o pid if [ $? -eq 1 ];then /usr/local/nginx/sbin/nginx sleep 3 ps -C nginx -o pid if [ $? -eq 1 ];then systemctl stop keepalived fi fi
为脚本添加执行权限
[root@xhk1 ~]# chmod a+x /root/1.sh
在xhk2也创建以上同样的脚本
启动keepalived服务并设置开机自启动：
[root@xhk1 ~]# systemctl start keepalived
[root@xhk1 ~]# systemctl enable keepalived
[root@xhk2 ~]# systemctl start keepalived
[root@xhk2 ~]# systemctl enable keepalived
查看keepalived服务状态
[root@xhk1 ~]# systemctl status keepalived
● keepalived.service - LVS and VRRP HighAvailability Monitor
Loaded:loaded(/usr/lib/systemd/system/keepalived.service; disabled; vendor preset:disabled)
Active: active (running) sinceFri 2017-10-20 00:07:48 EDT;28s ago
Process: 2712ExecStart=/usr/sbin/keepalived$KEEPALIVED_OPTIONS (code=exited,status=0/SUCCESS)
Main PID: 2713 (keepalived)
CGroup:/system.slice/keepalived.service
├─2306 nginx: masterprocess /usr/local/nginx/sbin/nginx
├─2308 nginx: workerprocess
├─2713/usr/sbin/keepalived -D
├─2714/usr/sbin/keepalived -D
└─2715/usr/sbin/keepalived -D
[root@xhk2 ~]# systemctl status keepalived
● keepalived.service - LVS and VRRP HighAvailability Monitor
Loaded:loaded(/usr/lib/systemd/system/keepalived.service; disabled; vendor preset:disabled)
Active: active (running) sinceFri 2017-10-20 00:07:59 EDT;29s ago
Process:2763ExecStart=/usr/sbin/keepalived $KEEPALIVED_OPTIONS(code=exited,status=0/SUCCESS)
Main PID: 2764 (keepalived)
CGroup:/system.slice/keepalived.service
├─2353 nginx: masterprocess /usr/local/nginx/sbin/nginx
├─2355 nginx: workerprocess
├─2764/usr/sbin/keepalived -D
├─2765/usr/sbin/keepalived -D
└─2766/usr/sbin/keepalived -D
接下来就是测试环节
由于xhk1设置MASTER的VIP是192.168.163.100，所以可以看到xhk1有一个VIP：192.168.163.100
[root@xhk1 ~]# ip addr sh
1: lo: <LOOPBACK,UP,LOWER_UP> mtu65536 qdisc noqueue stateUNKNOWN qlen 1
link/loopback00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scopehostlo
valid_lft foreverpreferred_lft forever
inet6 ::1/128 scopehost
valid_lft foreverpreferred_lft forever
2: ens32:<BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdiscpfifo_fast state UP qlen1000
link/ether00:0c:29:0e:bc:15brd ff:ff:ff:ff:ff:ff
inet 192.168.163.162/24brd192.168.163.255 scope global dynamic ens32
valid_lft 1586secpreferred_lft 1586sec
inet 192.168.163.100/32 scopeglobal ens32
valid_lft foreverpreferred_lft forever
inet6fe80::20c:29ff:fe0e:bc15/64 scope link
valid_lft foreverpreferred_lft forever
在xhk2上设置MASTER的VIP为192.168.163.200
[root@xhk2 ~]# ip addr sh
1: lo: <LOOPBACK,UP,LOWER_UP> mtu65536 qdisc noqueue stateUNKNOWN qlen 1
link/loopback00:00:00:00:00:00brd 00:00:00:00:00:00
inet 127.0.0.1/8 scopehostlo
valid_lft forever preferred_lft forever
inet6 ::1/128 scopehost
valid_lft foreverpreferred_lft forever
2: ens32:<BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdiscpfifo_fast state UP qlen1000
link/ether00:0c:29:82:ca:01brd ff:ff:ff:ff:ff:ff
inet 192.168.163.163/24brd192.168.163.255 scope global dynamic ens32
valid_lft 1772secpreferred_lft 1772sec
inet 192.168.163.200/32 scopeglobal ens32
valid_lft foreverpreferred_lft forever
inet6fe80::20c:29ff:fe82:ca01/64scope link
valid_lft foreverpreferred_lft forever
进行访问，能够看到前端服务器将请求均衡转到后端服务器
[root@clinet ~]# curl 192.168.163.100
nginx1
[root@clinet ~]# curl 192.168.163.100
nginx2
[root@clinet ~]# curl 192.168.163.200
nginx1
[root@clinet ~]# curl 192.168.163.200
nginx2
接下来测试前端服务器nginx服务停掉的状态，首先会尝试启动nginx，结果启动nginx服务成功，网页正常访问
[root@xhk1 ~]# /usr/local/nginx/sbin/nginx -s stop
[root@xhk1 ~]# systemctl status keepalived
● keepalived.service - LVS and VRRP HighAvailability Monitor
Loaded:loaded(/usr/lib/systemd/system/keepalived.service; disabled; vendor preset:disabled)
Active: active (running) sinceFri 2017-10-20 00:07:48 EDT;18min ago
Process: 2712ExecStart=/usr/sbin/keepalived$KEEPALIVED_OPTIONS (code=exited,status=0/SUCCESS)
Main PID: 2713 (keepalived)
CGroup:/system.slice/keepalived.service
├─2713/usr/sbin/keepalived -D
├─2714/usr/sbin/keepalived -D
├─2715 /usr/sbin/keepalived -D
├─4375 nginx: masterprocess /usr/local/nginx/sbin/nginx
└─4377 nginx: workerprocess
Oct 20 00:08:13 xhk1Keepalived_vrrp[2715]: Sending gratuitous ARP onens32 for 192.168.163.200
Oct 20 00:08:13 xhk1 Keepalived_vrrp[2715]:Sending gratuitous ARP onens32 for 192.168.163.200
Oct 20 00:08:13 xhk1Keepalived_vrrp[2715]: Sending gratuitous ARP onens32 for 192.168.163.200
Oct 20 00:08:13 xhk1Keepalived_vrrp[2715]: Sending gratuitous ARP onens32 for 192.168.163.200
Oct 20 00:08:16 xhk1Keepalived_vrrp[2715]: VRRP_Instance(VI_2)Received advert with higher priority100, ours 99
Oct 20 00:08:16 xhk1Keepalived_vrrp[2715]: VRRP_Instance(VI_2)Entering BACKUP STATE
Oct 20 00:08:16 xhk1Keepalived_vrrp[2715]: VRRP_Instance(VI_2)removing protocol VIPs.
Oct 20 00:26:05 xhk1Keepalived_vrrp[2715]: VRRP_Script(check_nginx)timed out
Oct 20 00:26:05 xhk1Keepalived_vrrp[2715]: /root/1.sh exited due tosignal 15
Oct 20 00:26:05 xhk1Keepalived_vrrp[2715]: VRRP_Script(check_nginx)succeeded
[root@xhk1 ~]# ps -ef | grep nginx
root 4375 1 0 00:26? 00:00:00 nginx:master process/usr/local/nginx/sbin/nginx
nginx 4377 4375 0 00:26 ? 00:00:00 nginx:worker process
root 4431 2222 0 00:26 pts/0 00:00:00grep--color=auto nginx
可以看到，keepalived每隔2秒运行的nginx检测脚本，发现nginx停掉的时候，尝试启动nginx，启动成功！
接下来测试nginx起不来的情况
为了模拟nginx起不来的情况，修改一下脚本，检测nginx不提供服务，就停掉keepalived服务
[root@xhk1 ~]# vim /root/1.sh
#!/bin/bash ps -C nginx -o pid if [ $? -eq 1 ];then systemctl stop keepalived fi
先停掉nginx服务
[root@xhk1 ~]# ps -ef |grep nginx |awk ‘{print $2}‘ |xargs kill -9
[root@xhk1 ~]# ps -ef |grep nginx
root 5390 2222 0 00:51pxs/0 00:00:00 grep --color=auto nginx
发现keepalived自动停掉
[root@xhk1 ~]# systemctl status keepalived
● keepalived.service - LVS and VRRP HighAvailability Monitor
Loaded:loaded(/usr/lib/systemd/system/keepalived.service; disabled; vendor preset:disabled)
Active: inactive (dead)
Oct 20 00:38:50 xhk1 Keepalived_vrrp[5400]:Opening file‘/etc/keepalived/keepalived.conf‘.
Oct 20 00:39:07 xhk1Keepalived_vrrp[5400]: VRRP_Instance(VI_1)removing protocol VIPs.
Oct 20 00:39:07 xhk1Keepalived_vrrp[5400]: VRRP_Instance(VI_2)removing protocol VIPs.
Oct 20 00:39:07 xhk1 Keepalived_vrrp[5400]:SECURITY VIOLATION -scripts are being executed but script_security n...bled.
Oct 20 00:39:07 xhk1Keepalived_vrrp[5400]: Using LinkWatch kernelnetlink reflector...
Oct 20 00:39:07 xhk1Keepalived_vrrp[5400]: VRRP_Instance(VI_2)Entering BACKUP STATE
Oct 20 00:39:07 xhk1Keepalived_vrrp[5400]: VRRP sockpool: [ifindex(2),proto(112), unicast(0),fd(10,11)]
Oct 20 00:39:07 xhk1 Keepalived[5398]:Stopping
Oct 20 00:39:07 xhk1 systemd[1]: StoppingLVS and VRRP HighAvailability Monitor...
Oct 20 00:39:09 xhk1 systemd[1]: Stopped LVSand VRRP High AvailabilityMonitor.
Hint: Some lines were ellipsized, use -lto show in full.
可以发现VIP都自动跳到另一台前端服务器（xhk2）上
[root@xhk2 ~]# ip add sh
1: lo: <LOOPBACK,UP,LOWER_UP> mtu65536 qdisc noqueue stateUNKNOWN qlen 1
link/loopback00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scopehostlo
valid_lft foreverpreferred_lft forever
inet6 ::1/128 scopehost
valid_lft foreverpreferred_lft forever
2: ens32:<BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdiscpfifo_fast state UP qlen1000
link/ether00:0c:29:82:ca:01brd ff:ff:ff:ff:ff:ff
inet 192.168.163.163/24brd192.168.163.255 scope global dynamic ens32
valid_lft 1598secpreferred_lft 1598sec
inet 192.168.163.200/32scopeglobal ens32
valid_lft foreverpreferred_lft forever
inet 192.168.163.100/32scopeglobal ens32
valid_lft foreverpreferred_lft forever
inet6fe80::20c:29ff:fe82:ca01/64 scope link
valid_lft foreverpreferred_lft forever
网页正常访问
[root@clinet ~]#curl 192.168.163.200
nginx1
[root@clinet ~]#curl 192.168.163.200
nginx2
[root@clinet ~]#curl 192.168.163.100
nginx1
[root@clinet ~]#curl 192.168.163.100
nginx2
本文出自 “xhk777” 博客，请务必保留此出处http://xhk777.blog.51cto.com/13405744/1974591
原文地址：http://xhk777.blog.51cto.com/13405744/1974591
