keepalived高可用的简单配置

搭建mongodb时出现的一个问题，防止自己忘记

该错误只针对阿里的云服务器

其他厂商的服务器未出现该问题

主备同时出现vip的情况

报错日志

May 24 18:31:43 pro-002 Keepalived_vrrp[1313]: VRRP_Instance(MG_1) removing protocol VIPs.
May 24 18:31:43 pro-002 Keepalived_vrrp[1313]: Using LinkWatch kernel netlink reflector...
May 24 18:31:43 pro-002 Keepalived_vrrp[1313]: VRRP_Instance(MG_1) Entering BACKUP STATE
May 24 18:31:43 pro-002 Keepalived_vrrp[1313]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]
May 24 18:31:47 pro-002 Keepalived_vrrp[1313]: VRRP_Instance(MG_1) Transition to MASTER STATE
May 24 18:31:48 pro-002 Keepalived_vrrp[1313]: VRRP_Instance(MG_1) Entering MASTER STATE
May 24 18:31:48 pro-002 Keepalived_vrrp[1313]: VRRP_Instance(MG_1) setting protocol VIPs.
May 24 18:31:48 pro-002 Keepalived_vrrp[1313]: Sending gratuitous ARP on eth0 for 192.168.100.1

keepalived只要启动，过几秒就抢占为master状态

主的配置

   !Configuration File for keepalived
         vrrp_instance MG_1 {
                state MASTER            
                interface eth0
                virtual_router_id 55
                priority 100           
                advert_int 1
                #nopreempt            
                authentication {
                    auth_type PASS
                    auth_pass 1111
                }     
    unicast_src_ip  192.168.0.1 //（主IP地址）
    unicast_peer {
                  192.168.0.2 //（备IP地址)
                       }
                virtual_ipaddress {
            192.168.100.1             
             }
         }
            virtual_server 192.168.100.1 80 {
            delay_loop 6
            lb_algo rr
            persistence_timeout 5
            protocol TCP
            real_server 0.0.0.0 80 {
            notify_down /root/ng.sh
            TCP_CHECK {
            connect_timeout 3
            nb_get_retry 3
            delay_before_retry 3
            connect_port 80
            }
            }
            }

备的配置

   !Configuration File for keepalived
         vrrp_instance MG_1 {
                state BACKUP
                interface eth0
                virtual_router_id 55
                priority 10           
                advert_int 1
                #nopreempt            
                authentication {
                    auth_type PASS
                    auth_pass 1111
                }     
                virtual_ipaddress {
            192.168.100.1             
             }
         }

这两段配置，直接粘贴拿过去用即可

阿里的云服务器默认将vrrp的组播，在物理层面的交换机和路由网中禁用掉的

所以不管是防火墙还是selinux、iptables都解决不了vrrp发不出去组播的情况，所以备服务器认为主是死掉了，将自己提升为master了

将组播改正单播尝试一下，指定备的eth0的IP地址，然后重新启动主的keepalived

    unicast_src_ip  192.168.0.1 //（主IP地址）
    unicast_peer {
                  192.168.0.2 //（备IP地址)
                       }

备服务器能够接收到vrrp报文后，自动切换为backup状态