will dynamic dhcpd.conf be supported in future releases

David W. Hankins David_Hankins at isc.org
Fri Aug 11 20:13:51 UTC 2006

On Fri, Aug 11, 2006 at 11:45:16AM -0700, Vithal Shirodkar wrote:
> The use will be seen in virtualization worlds, where you would want to dynamically add subnets and define ip ranges for the subnets.  Currently there is no way to do this without stopping dhcpd and restarting it.  Ideally it would be good to extend OMAPI to define a SUBNET object and be able to create it.  Any reason why the omapi support did not include subnets in the first place ?Vithal Shirodkar

It's not the easiest thing to make ISC DHCP's memory objects
OMAPI-manageable, so I'm certain the original author, Ted, did
what he was able to at the time and intended to add more pieces
over future releases.

In particular with subnets, it gets terribly hazy if the subnets
are tied to shared-networks that are present on local interfaces
(since DHCP uses the address advertised from the system on the
interfaces to find subnets and therefore shared networks, the
DHCP package would now have to 're-probe' existing interfaces
during runtime - something it is not nimble enough to do
currently).  So these complexities may or may not have contributed
to its lack of presence in the software, but I suspect the main
contributor was a lack of time.

In the HEAD branch when I came to ISC, there were a few OMAPI
features there already.  Namely, features to supply and manage
class and subclass statements.  These don't appear to be unwise,
and don't impact the server's ability to run if you don't use
this feature, so I'm leaving them 'in' for 3.1.0 after doing
what I hope was a careful and effective source level review of
the changes.

If I haven't said it clearly before, let me say it now: our long
term plan is to remove OMAPI from the software.  So this is
probably the last specifically OMAPI feature addition you'll
see from us, unless you count the 'feature' of removing it

But before we do that, we will provide a not-quite-drop-in
replacement, which I don't know enough about (still in the
vaguest of designs) to speak more of.

But...patches go to dhcp-suggest at isc.org.  If we can review it
cleanly, it will go in.

David W. Hankins	"If you don't do it right the first time,
Software Engineer		you'll just have to do it again."
Internet Systems Consortium, Inc.	-- Jack T. Hankins

More information about the dhcp-users mailing list