migration to failover setup (3.1.0)

[Carlos Vicente]

I'm about to migrate a fairly large standalone server setup to a
failover/redundant pair, but first I decided to test on two lab machines. 

I'm watching the logs and I'm curious as to what this means exactly:

    dhcpd: bind update on x.x.y.y from dhcp-peer rejected: x.x.y.y:
invalid state transition: active to free

Looking at that particular IP in both leases files I see:

Primary:

lease x.x.y.y {
  starts 2 2008/03/25 14:29:59;
  ends 3 2008/03/26 02:29:59;
  tstp 3 2008/03/26 02:29:59;
  tsfp 3 2008/03/26 02:29:59;
  cltt 2 2008/03/25 14:29:59;
  binding state free;
  hardware ethernet 00:16:cb:c0:bf:ec;
  uid "\001\000\026\313\300\277\354";
  client-hostname "xxxx";
}

Secondary:

lease x.x.y.y {
  starts 2 2008/03/25 14:29:59;
  ends 3 2008/03/26 02:29:59;
  tsfp 3 2008/03/26 02:29:59;
  atsfp 3 2008/03/26 02:29:59;
  cltt 2 2008/03/25 14:29:59;
  binding state active;
  next binding state expired;
  hardware ethernet 00:16:cb:c0:bf:ec;
  uid "\001\000\026\313\300\277\354";
}

which doesn't look very good.  Won't the primary assign that address to
some other machine, even though it's active in the secondary?

The assignment of addresses seems to be working OK, at least from my
test laptop.

My procedure was:

1) Copy exact same dhcpd.conf on both machines
2) Include individual /etc/dhcpd.peer files with their respective custom
values
3) Copy old /etc/dhcpd.leases file in primary
4) Create empty leases file in secondary
5) Start primary
6) Start secondary

Is this the most recommended procedure?  Should I worry about the
situation above?

Thank you.

cv