Proper procedure for making config changes and restarting servers in failover relationship?

[Nomad]

I would appreciate any advice/suggestions/recommendations on whether I'm
doing this correctly after making a configuration change.  I have two
servers in a failover relationship and after I make a change to the primary
(ex. adding a new subnet), here's my procedure:

1.  I down the primary gracefully using this script:

#!/bin/sh
omshell << EOF
port 7911
connect
new control
open
set state = 2
update
EOF

2.  I delete /var/run/dhcpd.pid and make sure it's gone

3.  I wait for two seconds (have noticed if I don't do this, dhcpd fails to
come up)

4.  Restart dhcpd

5.  Copy common conf file to secondary server

6.  Wait for the MCLT to expire

7.  Restart dhcpd on secondary (using same procedure as in steps 1-3)


I guess my real concern is what happens after I restart dhcpd on the primary
... what happens to client requests?  I see in the logs that the server goes
into the RECOVER state but the "not responding" messages worry me.  Is it
possible that I restart the server in mid-conversation
with a client so it appears to the client that the server is down?  Or
are the clients resilient enough that they can restart the
conversation with the secondary answering?

Sorry for all the questions ... I am reading along in the DHCP Handbook but
some things aren't clear to me ... and to top things off, these servers are
in operation so some items seem like emergencies and incomprehensible  when
under normal circumstances they would be obvious.


Thank you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20080110/d3781e78/attachment.html>