In order to keep the
failover pair operating in spite of a network partition, in which both can
communicate with clients but not with each other, you must make available more
addresses than the addresses needed to run a single server. Configure the main
server to allocate a percentage of the currently available (unassigned)
addresses in each scope or prefix delegation address pool to its partner. These
addresses become unavailable to the main server. The partner uses them when it
cannot talk to the main server and does not know if it is down. However, when
the failover partners are in communication, they periodically rebalance these
The backup server
needs enough addresses from each scope or prefix to satisfy the requests of all
new DHCP clients that arrive during the period in which the backup does not
know if the main server is down. In Cisco Prime IP Express 8.2 or later,
the default backup percentage for a failover pair is 50%. This ensures that
during the failover the other partner has equal number of addresses.
PARTNER-DOWN state, the backup server waits for the lease expiration and the
maximum client lead time (MCLT), a small additional time buffer, before
reallocating any leases. When these times expire, the backup server offers:
- Leases from its private pool
- Leases from the main server
pool of addresses.
- Expired leases to new
During the working
hours, if the administrative staff can respond within two hours to a
COMMUNICATIONS INTERRUPTED state to determine if the main server is working,
the backup server needs enough addresses to support a reasonable upper bound on
the number of new DHCP clients that might arrive during those two hours.
During off-hours, if
the administrative staff can respond within 12 hours to the same situation, and
considering that the arrival rate of previously unheard from DHCP clients is
also less, the backup server then needs enough addresses to support a
reasonable upper bound on the number of DHCP clients that might arrive during
those 12 hours.
number of addresses over which the backup server requires sole control would be
the greater of the numbers of addresses given out during peak and non-peak
times, expressed as a percentage of the currently available (unassigned)
addresses in each scope or prefix.
Starting in 8.2,
the default use-safe-period is enabled for the DHCP failover pair and the
default safe period is 4 hours. This ensures that if the failover partner is in
COMMUNICATIONS-INTERRUPTED state for 4 hours, it will enter PARTNER-DOWN state
automatically after the safe period elapses.