reserved leases

[David W. Hankins]

On Thu, Sep 21, 2006 at 06:59:01PM +1000, Ray Phillips wrote:
> >It might be easier to use host records, which are 'failover compatible'
> >(in that both servers will happily answer).
> 
> Yes, it would be easier.  I just like the DNS records to be updated 
> automatically.

You mean torn down automatically.  If you wanted updates, you could
just use 'update-static-leases'.  It's teardowns you miss (static
leases have no expiration event).

> >Reserved leases will only work in the 1:1 case.
> 
> OK, there's only one client per one-number pool.

Then it should work a lot better than trying to limit a class
of one client to one-address pools.

But it also hasn't been widely used or tested so thanks in
advance for testing our software with your forehead...

If only there weren't this pesky problem of not having a good way
to set the lease reserved.

> How would you use the client identifier, for example, to make a 
> device lease the number you want it to?  Is it a matter of inserting 
> a lease with that client-id and IP number into dhcpd.leases manually?

Precisely.  This is described as a 'UID' in dhcpd.leases.

> By the way, in case it matters, when I tried using 3.1.0a1 briefly on 
> a machine running NetBSD/i386 3.0 "non-null pointer" appeared in 
> /var/log/dhcpd a couple of times:

Woot!

> Sep 21 01:03:11 ns dhcpd: dhcp.c(3928): non-null pointer

Cool, thanks.  This is an easy bug (stack variable uninitialized),
we'll fix it in 3.1.0a2.

In the meantime it looks, due to this bug, like configuring a
server-identifier on your system will fail to work (it will be
ignored).

Thanks man.

-- 
ISC Training!  October 16-20, 2006, in the San Francisco Bay Area,
covering topics from DNS to DDNS & DHCP.  Email training at isc.org.
-- 
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