Comment 3 for bug 1668489

the log show "eth2 is down", the issue is resolved after I config eth2.

moreover, we can better kolla code.

[root@kola templates]# pwd
/opt/kolla-ansible-4.0.0.0rc1/ansible/roles/haproxy/templates
[root@kola templates]# vim keepalived.conf.j2

  track_interface {
        {{ kolla_external_vip_interface }}
    }

eth2 is the kolla_external_vip_interface, so it requires eth2 is up and has been configured before keepalived process starting.

but prechecks hasn't this check, I think we should add a checking in precheck code.