使用 做双机热备非常简单，经常和 搭配来实现高可用负载平衡方案。
1. Master / Slave
首先准备两台测试服务器和一个虚拟IP。
Server A: 192.168.1.10 (主服务器)
Server B: 192.168.1.20
Virtual IP: 192.168.1.100
测试服务: 在两台服务器上分别安装 Nginx，并修改默认的 index.html 文件，显示当前服务器 IP 以便识别。
1. 在两台服务器上分别安装 keepalived。
$ sudo apt-get install keepalived
2. 添加配置文件。
Server A$ sudo vim /etc/keepalived/keepalived.conf
global_defs {
router_id LVS_DEVEL
}
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 51 # 保持主从服务器一致
priority 100 # 优先级 (主服务器较高)
advert_int 1 # 心跳广播间隔(秒)
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.100 # 虚拟IP地址，可以多个。
}
}
Server B$ sudo vim /etc/keepalived/keepalived.conf
global_defs {
router_id LVS_DEVEL
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
virtual_router_id 51
priority 99
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.100
}
}
注意：备份服务器 Server B 配置中 state 要改成 BACKUP，同时调低 priority。
3. 启动两台服务器上的 keepalived 服务。
$ sudo service keepalived start
重启后可以使用 "ip a" 查看虚拟 IP 信息。
Server A$ ip a
1: lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:4c:e7:e7 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.10/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.100/24 scope global secondary eth0
inet6 fe80::20c:29ff:fe4c:e7e7/64 scope link
valid_lft forever preferred_lft forever
Server B$ ip a
1: lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:01:d8:16 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.20/24 brd 192.168.1.255 scope global eth0
inet6 fe80::20c:29ff:fe01:d816/64 scope link
valid_lft forever preferred_lft forever
4. 在第三台机器上进行访问测试。
$ curl
Welcome to nginx!
Welcome to nginx! 192.168.1.10
$ curl
Welcome to nginx!
Welcome to nginx! 192.168.1.20
$ curl
Welcome to nginx!
Welcome to nginx! 192.168.1.10
我们关掉主服务器 192.168.1.10，再访问 就会自动切换成备份服务器 (Server B: 192.168.1.20)。
$ curl
Welcome to nginx!
Welcome to nginx! 192.168.1.20
同时 Server B 绑定了虚拟 IP。
Server B$ ip a
1: lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:01:d8:16 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.20/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.100/24 scope global secondary eth0
inet6 fe80::20c:29ff:fe01:d816/64 scope link
valid_lft forever preferred_lft forever
重新打开主服务器(Server A: 192.168.1.10)，访问恢复。
2. Master / Master
Master / Slave 方案中备份服务器(Server B)平时就是个摆设，有点浪费。我们完全可以用来跑其他服务，让两台主机形成相互热备。
Server A: 192.168.1.10, Virtual IP: 192.168.1.100
Server B: 192.168.1.20, Virtual IP: 192.168.1.200
修改配置文件。
Server Aglobal_defs {
router_id LVS_DEVEL
}
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.100
}
}
vrrp_instance VI_2 {
state BACKUP
interface eth0
virtual_router_id 52
priority 99
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.200
}
}
Server B:global_defs {
router_id LVS_DEVEL
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
virtual_router_id 51
priority 99
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.100
}
}
vrrp_instance VI_2 {
state MASTER
interface eth0
virtual_router_id 52
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.200
}
}
其实很简单，我们增加了一个新的配置 VI_2 (注意 virtual_router_id 不同)。不过这回用 Server B 做主服务器，如此 Server A、Server B 各自拥有主虚拟IP，同时备份对方的虚拟 IP。重启两台服务器的 keepalived 服务后，查看虚拟 IP 绑定信息。
Server A$ ip a
1: lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:4c:e7:e7 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.10/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.100/24 scope global secondary eth0
inet6 fe80::20c:29ff:fe4c:e7e7/64 scope link
valid_lft forever preferred_lft forever
Server B$ ip a
1: lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:01:d8:16 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.20/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.200/24 scope global secondary eth0
inet6 fe80::20c:29ff:fe01:d816/64 scope link
valid_lft forever preferred_lft forever
正常情况下，会使用各自的主服务器。
$ curl
Welcome to nginx!
Welcome to nginx! 192.168.1.10
$ curl
Welcome to nginx!
Welcome to nginx! 192.168.1.20
一旦任何一台服务器当机，另一台就会自动接管。我们停掉 192.168.1.20，看看访问 是不是切换到 192.168.1.10 上。
$ curl
Welcome to nginx!
Welcome to nginx! 192.168.1.10
同时 Server A 绑定虚拟 IP 192.168.1.200。
$ ip a
1: lo: mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
link/ether 00:0c:29:4c:e7:e7 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.10/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.100/24 scope global secondary eth0
inet 192.168.1.200/24 scope global secondary eth0
inet6 fe80::20c:29ff:fe4c:e7e7/64 scope link
valid_lft forever preferred_lft forever
Server B 重启后，一切恢复正常。
这个方案可以是不同的服务，或者是同一服务的访问分流(配合 DNS 使用)。
更详细的信息请参考官方网站。
