Failover Without Address Ranges

[Norman Elton]

It seems that configuring failover requires a pool declaration, which
requires an address range. All this makes sense for most of our
subnets. But we have a few subnets where hosts are only given
addresses manually assigned via a "host" declaration. That is, there
is no address range for "anonymous" clients.

How does such a subnet interact with failover? Do the servers act
independently of each other? Would this cause a problem?

Thanks

Norman