Biggest installation of ISC DHCP?

Nick Urbanik nicku at
Sat Jan 19 01:54:08 UTC 2008

Dear Folks,

On 18/01/08 15:00 -0800, David W. Hankins wrote:
>On Sat, Jan 19, 2008 at 07:02:35AM +1100, Nick Urbanik wrote:
>> ISC dhcp with our own LDAP patch, and the "next file" patch (see
>> so that there is no need to restart the
>you know, i did a lot of talking 2 years ago that we wanted to do SQL
>and LDAP and "others", and that i wanted to integrate this sort of
>work into one event rather than create some sort of frankenstein
>monster.  make sure we got the API right so we didn't have to do a
>lot of rewriting later.

One thing that I would do now instead of the "next file" solution to
the hourly restarting problem is to simply have the server write its
leases to a DB file.  That way it would remain at a constant size and
wouldn't need aggregating or restarting.  Startup time is a
consideration when the log-structured leases file is more than 300MB
in size, and hourly restarting just becomes a problem.

>essentially putting this off into some nebulous "future time" that
>never seems to come, but always sat in the 1-2 year plan in my head.
>but that was when i was under the impression that DHCP as it exists
>today would continue forward with very subtle change, and i would be
>the only one working on it (and thus _very_ concerned about my budget
>for maintenance-related coding hours).
>recently i've not been alone, and we're thinking more and more that
>DHCP 3 and DHCP 4 are more like (hopefully useful) iterations in
>development leading to an even grander change in architecture.
>stepping stones.
>essentially putting this at the 5 year evolutionary scale when
>today's single-celled DHCP amoeba would be unrecognizable in the
>multi-celled fully ambulatory DHCP dragon that drags itself up onto
>the beach with its litter of bird-like children.
>maybe i'm stretching the metaphor a bit far; the point is this is an
>even more nebulous and distant hour.
>so these days i'm a lot more interested in pulling up feature patches
>for features we want to have in the multi-celled organism, so that we
>can have examples and some kind of experience with the features when
>we come around to fitting them into the new design.
>i'm a lot less attached to our even mid-term architecture
>if you, or anyone else is willing to work with us to get patches
>integrated (is there more than one LDAP patch floating around now?),
>please port them to DHCP-4 and submit them to dhcp-suggest.  it
>wouldn't hurt to join dhcp-workers either.

This LDAP patch simply extends the expression language to look up
things from LDAP.  We use it to look up customers' MAC addresses.  It
does not read the configuration from LDAP.  We read the configuration
from /etc/dhcpd.conf.

We are also using a DHCPLEASEQUERY patch.

>if there are any licensing considerations or carry-along libraries,
>please bring them up on dhcp-workers first before you spend a lot of
>time coding.

I need to determine the origin of the LDAP expression patch.  We use
no additional proprietary libraries.
Nick Urbanik   RHCE        nicku at
GPG: 7FFA CDC7 5A77 0558 DC7A 790A 16DF EC5B BB9D 2C24 ID: BB9D2C24
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <>

More information about the dhcp-users mailing list