I am having what appears to be the exact same issue, although for me in addition to the floating IP address interface having status DOWN the external router's gateway also always has a DOWN status and I cannot get any external access from my instances. I've installed Juno using the OpenStack Juju charms with neutron networking.
I've followed the network troubleshooting guide ( http://docs.openstack.org/openstack-ops/content/network_troubleshooting.html (http://docs.openstack.org/openstack-o...) ) and find that my instances can ping each other, but not external hosts. The outbound external ping packets do appear on the integration bridge on the network node, so the internal connectivity between compute and network node seems fine, but the problem appears to be with the external router itself.
I'm not sure how to debug this but I suspect it may have something to do with the l3 agent? I have checked and the agent is running and has a smiley face in its status - the only indication of a problem is the interface status being DOWN for all the ports on the external network. Can anyone tell us what process/service/agent is responsible for issuing the "DOWN" status on interface ports?
I am having same issue too, If you have found any solution please post it over here.