FW: Config help!

[Bruce Hudson]

> I hate to see when this happens but the server is back up assigning
> addresses to my subnets and I don't know what made it to work.
>  From working with a couple of you guys (thanks a lot for your time put
> aside) I made these changes as suggested:
> 
> - took the shared-network statement

    Helpful but irrelevant to the "no interface" problem.

> - instead of using the script in /etc/rc.d/init.d/ I used:
>  /usr/sbin/dhcp -cf /etc/dhcpd.conf -lf /var/lib/dhcp/dhcpd.leases start |
> stop (didn't work)

    Adding "start" or "stop" to the command line would certainly cause the
"no interface" problem. This is an parameter to the script to tell it what
to do. The only parameter (as opposed to options) the binary takes is a 
list of interfaces it should listen to. You do not have an interface named
"start" or "stop". Leaving that off, the rest of the command should work.
 
> - I put the authoritative statement in the global section, I had it inside
> each subnet.
 
    Again, this is largely irrelevant to the problem at hand. You also have
host declarations inside your subnet blocks. Host declarations are global in
scope but including them inside a subnet has caused weird problems in some
cases. I was going to wait to mention them. :)

> After making these changes I saw in the log that it wasn't working.
> Sometime while I was making the changes, it decided to work.  I am still
> not able to start/stop the server with the script.

    This is mostly grasping at straws but, is it possible that you have some
non-printable character embedded into either the script or the DHCPDOPTS
definition? That might be seen by the daemon as an interface name but not
have gotten (visibly) included in the trace output you sent us. 

> One other change I made this morning that I know was after the fact, was to
> define a subnet for my ip phones (they run on a vlan at the Corporate
> network and use a separate MS dhcp with their own scope) like this:
> 
> shared-network INITIAL {
> # Subnet: Corporate
>         subnet 192.9.200.0 netmask 255.255.255.0 {
>                 option routers 192.9.200.241;
>                 range 192.9.200.50 192.9.200.97;
>                 range 192.9.200.116 192.9.200.199;
>                 option broadcast-address 192.9.200.255;
>                 }
> # Subnet: IPPhonesCorporate
>         subnet 172.16.99.0 netmask 255.255.255.0 {
>                 }
> 
> }

    If I understand you correctly, you may want to include a "not
authoritative;" statement in the IPPhonesCorporate block. Otherwise
your server may NAK requests for addresses from that subnet.

> After all this headache, I decided to move the dhcp services down to the
> router at each location and only keep my corporate subnet.
> Sorry I have no solution or a concrete fix.

    I am sorry we could not track this down.
--
Bruce A. Hudson				| Bruce.Hudson at Dal.CA
UCIS, Networks and Systems		|
Dalhousie University			|
Halifax, Nova Scotia, Canada		| (902) 494-3405