一、原理介绍

keepalived的原理可以这样简单理解:

keepalived安装在两台物理服务器上，并相互监控对方是否在正常运行。当节点A正常的时候，节点A上的keepalived会将下面的信息广播出去，192.168.8.100 这个IP对应的MAC地址为节点A网卡的MAC地址当节点A发生故障的时候，节点B上的keepalived会检测到，并且将下面的信息广播出去，192.168.8.100 这个IP对应的MAC地址为节点B网卡的MAC地址

高可用负载平衡配置

环境：

两台物理服务器和一个虚拟服务器(vip):

master：192.168.1.103
backup：192.168.1.104
vip：192.168.1.110

在主备机器上安装keepalived

tar -zxvf keepalived-1.2.20.tar.gz
cd keepalived-1.2.20
./configure --prefix=/usr/local/keepalived1.2
make && make install
ln -s /usr/local/keepalived1.2 /usr/local/keepalived
ln -s /usr/local/keepalived/sbin/* /usr/sbin/

配置keepalived

mkdir -p ~/etc/keepalivedcp /usr/local/keepalived/etc/keepalived/keepalived.conf ~/etc/keepalived/cd ~/etc/keepalived/

keepalived.conf配置

主机器（192.168.1.103）配置

global_defs {
  router_id NodeA
}
vrrp_instance VI_1 {
  state MASTER    #设置为主服务器，该配置决定了节点为主节点
  interface eth0    #监测网络接口
  virtual_router_id 51    #主、备必须一样
  priority 100    #(主、备机取不同的优先级，主机值较大，备份机值较小,值越大优先级越高)
  advert_int 1   #VRRP Multicast广播周期秒数
  authentication {
    auth_type PASS    #VRRP认证方式，主备必须一致
    auth_pass 1111    #(密码)
  }
  virtual_ipaddress {
    192.168.1.110/24 #VRRP HA虚拟地址
  }
}

备用机器（192.168.1.104）配置

global_defs {
  router_id NodeB
}
vrrp_instance VI_1 {
  state BACKUP    #设置为备服务器
  interface eth0    #监测网络接口
  virtual_router_id 51    #主、备必须一样
  priority 90    #(主、备机取不同的优先级，主机值较大，备份机值较小,值越大优先级越高)
  advert_int 1   #VRRP Multicast广播周期秒数
  authentication {
    auth_type PASS    #VRRP认证方式，主备必须一致
    auth_pass 1111    #(密码)
  }
  virtual_ipaddress {
      192.168.1.110/24  #VRRP HA虚拟地址
  }
}

启动keepalived:

keepalived -D -f ~/etc/keepalived/keepalived.conf

查看log消息:

tail -f /var/log/messages

关闭keepalived

killall keepalived

启动主节点A后的日志为：会广播ARP消息通过 ip addr show 命令可以看到主服务器192.168.1.110/24绑定到了eth0上

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 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: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 08:00:27:57:e2:8d brd ff:ff:ff:ff:ff:ff
inet 192.168.1.103/24 brd 192.168.1.255 scope global eth0
inet 192.168.1.110/32 scope global eth0
inet6 fe80::a00:27ff:fe57:e28d/64 scope link
   valid_lft forever preferred_lft forever

部署nginx（略）

虚拟IP：http://192.168.1.110/
主服务器：http://192.168.1.103/
备服务器：http://192.168.1.104/

模拟故障自动切换

将主服务器（192.168.1.103）的keepalived停止: killall keepalived或者init 0关闭服务器通过下面的url访问 http://192.168.1.110/，返回的内容应该为备服务器上的内容同时查看备服务器的日志：发现备服务器转为主节点并且会广播ARP消息

Apr 14 00:25:39 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) Transition to MASTER STATE
Apr 14 00:25:40 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) Entering MASTER STATE
Apr 14 00:25:40 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) setting protocol VIPs.
Apr 14 00:25:40 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.1.110
Apr 14 00:25:40 fdfs2 Keepalived_healthcheckers[4323]: Netlink reflector reports IP 192.168.1.110 added
Apr 14 00:25:45 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.1.110

主服务器故障恢复

通过下面的url访问 http://192.168.1.110/，返回的内容应该为主服务器上的内容主服务器日志

Apr 14 00:27:48 fdfs1 Keepalived_vrrp[8651]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.1.110
Apr 14 00:27:49 fdfs1 Keepalived_vrrp[8651]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 192.168.1.110

备服务器日志

Apr 14 00:27:44 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) Received higher prio advert
Apr 14 00:27:44 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) Entering BACKUP STATE
Apr 14 00:27:44 fdfs2 Keepalived_vrrp[4324]: VRRP_Instance(VI_1) removing protocol VIPs.
Apr 14 00:27:44 fdfs2 Keepalived_healthcheckers[4323]: Netlink reflector reports IP 192.168.1.110 removed

常见问题：

处理vip无法被应用绑定

解决：net.ipv4.ip_nonlocal_bind=1

CloudX

一、原理介绍

高可用负载平衡配置

环境：

常见问题：

参考链接