What is wrong here is specifying the same interface in both 'interface' and in the 'track_interface' block. By virtue of the 'interface' specification, the configured interface will be tracked, and so keepalived issues a warning when the same interface is specified in the 'track_interface' block.
In the same appendix, it has
% if TYPE == 'MASTER':
state MASTER
% else:
state SLAVE
% endif
It is in fact completely unnecessary to specify the initial state; unless the priority of the VRRP instance is 255 the VRRP instance will always start in backup state since it has to ensure that there is no other instance in master state with the same or a higher priority.
Rudolfo,
The keepalived configuration that I was referring to is in the Appendix of https:/ /wiki.openstack .org/wiki/ Neutron/ L3_High_ Availability_ VRRP
It has: get_ha_ device_ name(TRACK_ PORT_ID) } router_ id ${VR_ID}
${L3_AGENT. get_ha_ device_ name(TRACK_ PORT_ID) }
interface ${L3_AGENT.
virtual_
priority ${PRIORITY}
track_interface {
}
What is wrong here is specifying the same interface in both 'interface' and in the 'track_interface' block. By virtue of the 'interface' specification, the configured interface will be tracked, and so keepalived issues a warning when the same interface is specified in the 'track_interface' block.
In the same appendix, it has
% if TYPE == 'MASTER':
state MASTER
% else:
state SLAVE
% endif
It is in fact completely unnecessary to specify the initial state; unless the priority of the VRRP instance is 255 the VRRP instance will always start in backup state since it has to ensure that there is no other instance in master state with the same or a higher priority.