Active / Passive HA Proxy: IP flutuante não é “flutuante”

Question

Estou configurando o HA Proxy no modo Ativo / Passivo.

haproxy-a : 172.29.240.172
haproxy-b : 172.29.240.173
Floating IP (VIP) : 172.29.240.188

Antes de qualquer configuração:

[root@haproxy-a/b ~]# cat /etc/redhat-release 
Red Hat Enterprise Linux Server release 7.3 (Maipo)

[root@haproxy-a ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000
    link/ether 00:50:56:9b:22:86 brd ff:ff:ff:ff:ff:ff
    inet 172.29.240.172/26 brd 172.29.240.191 scope global ens160
       valid_lft forever preferred_lft forever

[root@haproxy-b keepalived]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000
    link/ether 00:50:56:9b:2b:a6 brd ff:ff:ff:ff:ff:ff
    inet 172.29.240.173/26 brd 172.29.240.191 scope global ens160
       valid_lft forever preferred_lft forever

Passos que eu fiz.

Configurar o keepalived em ambos os servidores

haproxy-a/b # yum install -y keepalived
[root@haproxy-a ~]# cat /etc/keepalived/keepalived.conf 
vrrp_script chk_haproxy {
      script "killall -0 haproxy"
      interval 1
            weight -90
      }

vrrp_instance VI_1 {
    interface ens160 #interface to monitor
        state MASTER
        virtual_router_id 51
        priority 100  # highest priority wins the election of master

    virtual_ipaddress {
      172.29.240.188
    }

    track_script {
      chk_haproxy
    }
}

[root@haproxy-b ~]# cat /etc/keepalived/keepalived.conf 
vrrp_script chk_haproxy {
      script "killall -0 haproxy"
      interval 1
            weight -10
      }

vrrp_instance VI_1 {
    interface ens160 #interface to monitor
        state BACKUP
        virtual_router_id 51
        priority 50  # highest priority wins the election of master

    virtual_ipaddress {
      172.29.240.188
    }

    track_script {
      chk_haproxy
    }
}

Inicio então o serviço keepalived nos dois nós. firewalld & iptables está parado e nenhuma outra configuração foi alterada no nível do SO.

Quando o keepalived está em alta, não vejo os IPs flutuantes atribuídos a nenhum dos sistemas. Registros do sistema nos dois nós dizem:

Jul  6 13:26:51 haproxy-a Keepalived_vrrp[1862]: ip address associated with VRID not present in received packet : 172.29.240.188
Jul  6 13:26:51 haproxy-a Keepalived_vrrp[1862]: one or more VIP associated with VRID mismatch actual MASTER advert
Jul  6 13:26:51 haproxy-a Keepalived_vrrp[1862]: bogus VRRP packet received on ens160 !!!
Jul  6 13:26:51 haproxy-a Keepalived_vrrp[1862]: VRRP_Instance(VI_1) Dropping received VRRP packet...

    

Answer 1

Cada nó lógico no keepalived deve ter seu próprio virtual_router_id exclusivo no domínio de broadcast.