When I failover two openbsd 4.3 firewalls running bgp with the depend on
carp directive, there are certain times where the bgp state seems to get
stuck in an Active state and stays in that state in what seems an indefinate
amount of time, although I have only waited up to about 5 minutes in one
case for any timers to expire but they never did. Other times the failover
is clean and the bgp session is established almost instantaneously and
routes are propogated both ways almost immediately which is what I need to
happen whenever I decide to failover the firewalls. Logs dont reveal much
insight and packet captures have revealed some RST packets being sent to the
OPEN messages but Im not sure exactly why that is? Any help resolving this
would be appreciated?
