bgpd - carp demotion


other way around: bgpd influences carp master/backup state
carp interface group can be "demoted" (demotion counter)
demoted carp always loses against non-demoted
when both (all) are demoted, the one with the lowest demotion counter becomes master
we also use that to make sure a host doesn't become master while /etc/rc is running and starting services (unless there is no other, of course)
bgpd can demote the group based on session state, so your router does not become master unless some important sessions are up