目录

问题背景：

日志分析如下：

原因和解决方案总结：

问题背景：

keepalived-master和keepalived-slave同时出现了VIP，出现了非对称路由和双主现象

日志分析如下：

• master能够接受到来自slave的通告消息，并且master优先级100高于slave80，这是正常的。
• slave日志没有与master替换相关的信息记录，smtp持续报错，并且没有接受到来自master的通告消息，无法确认master状态即slave认为master宕机，所以就up上vip给自己。

[root@test1 ~]# journalctl -fu keepalived  
-- Logs begin at Wed 2024-10-30 18:47:13 CST. --
May 28 21:46:28 test1 Keepalived_vrrp[28060]: (VI_1) Received advert from 192.168.2.191 with lower priority 80, ours 100
May 28 21:46:29 test1 Keepalived_vrrp[28060]: (VI_1) Received advert from 192.168.2.191 with lower priority 80, ours 100
May 28 21:46:30 test1 Keepalived_vrrp[28060]: (VI_1) Received advert from 192.168.2.191 with lower priority 80, ours 100
May 28 21:46:31 test1 Keepalived_vrrp[28060]: (VI_1) Received advert from 192.168.2.191 with lower priority 80, ours 100
May 28 21:46:32 test1 Keepalived_vrrp[28060]: (VI_1) Received advert from 192.168.2.191 with lower priority 80, ours 100
May 28 21:46:33 test1 Keepalived_vrrp[28060]: (VI_1) Received advert from 192.168.2.191 with lower priority 80, ours 100
May 28 21:46:34 test1 Keepalived_vrrp[28060]: (VI_1) Received advert from 192.168.2.191 with lower priority 80, ours 100[root@test2 ~]# journalctl -fu keepalived
-- Logs begin at Wed 2024-10-30 18:43:15 CST. --
May 28 21:39:30 test2 Keepalived_healthcheckers[22487]: Removing service [192.168.2.193]:tcp:80 from VS [192.168.2.100]:tcp:80
May 28 21:39:30 test2 Keepalived_healthcheckers[22487]: Lost quorum 1-0=1 > 0 for VS [192.168.2.100]:tcp:80
May 28 21:39:30 test2 Keepalived_healthcheckers[22487]: smtp fd 10 returned write error
May 28 21:43:48 test2 Keepalived_healthcheckers[22487]: TCP connection to [192.168.2.192]:tcp:80 success.
May 28 21:43:48 test2 Keepalived_healthcheckers[22487]: Adding service [192.168.2.192]:tcp:80 to VS [192.168.2.100]:tcp:80
May 28 21:43:48 test2 Keepalived_healthcheckers[22487]: Gained quorum 1+0=1 <= 1 for VS [192.168.2.100]:tcp:80
May 28 21:43:48 test2 Keepalived_healthcheckers[22487]: smtp fd 10 returned write error
May 28 21:43:56 test2 Keepalived_healthcheckers[22487]: TCP connection to [192.168.2.193]:tcp:80 success.
May 28 21:43:56 test2 Keepalived_healthcheckers[22487]: Adding service [192.168.2.193]:tcp:80 to VS [192.168.2.100]:tcp:80
May 28 21:43:56 test2 Keepalived_healthcheckers[22487]: smtp fd 10 returned write error

原因和解决方案总结：

• 查看防火墙状态，按以上日志的话主要查看slave设备的防火墙，将其关闭。（本人所遇问题为防火墙问题）。
• 在产生切换master行为后，网络故障或延迟导致通告消息不及时，产生双主，建议手动停止keepalived，删除一侧VIP地址，再按照优先级配置从大到小顺序进行启动master和slave。
• 主从keepalived配置文件中的vrrp_instance配置错误，着重关注网卡名称interface是否配置正确。
• 查看主从服务的时间是否同步，若不同步可能需要使用chrony进行统一时间。