The option ha_vrrp_health_check_interval [1] can re-trigger
the election process until a master is re-elected when multiple
masters problem appear. This is an important feature that enables
the system to recover automatically, we should enable it.
[1] https://docs.openstack.org/ocata/networking-guide/ \
deploy-ovs-ha-vrrp.html#keepalived-vrrp-health-check
Change-Id: Iaf15ac77e249d1fe4a5101068761302e53385642
Closes-Bug: 1732154
7b60534ce8