[Kea-users] Kea-DHCPv6: No static reservations available
Darren Ankney
darren.ankney at gmail.com
Thu Dec 4 01:40:05 UTC 2025
Hi Christoph,
On Tue, Dec 2, 2025 at 9:07 PM Christoph Markert <maggiv8 at gmail.com> wrote:
>
> Hi Darren,
>
> please find the log attached. You can see that after around 6min KEA uses the defined hostname.
> Could you please elaborate a bit more on the prefix delegation part. Why would I need it?
You probably don't need it unless this is some kind of router with a
network behind it that needs subnet assignment. Prefix delegation
(IA_PD) is meant to distribute one or more subnets for use as an
"inside" network since NAT, generally speaking, is not used in IPv6
(at least not in the same way it was in IPv4). You can read all of
the details about it here:
https://datatracker.ietf.org/doc/html/rfc8415#section-6.3 It is safe
to not configure a prefix delegation subnet. You would know if you
needed to.
I see what happened here though I don't know why. Initially when your
client obtained the ip address (IA_NA) via a full SARR
(solicit->advertise->request->reply). It did ask for an IA_PD during
this time as well. During the advertise, Kea sent back a blank FQDN.
During the Reply, Kea generated a dynamic hostname and returned this
in option 39 after first logging about a partial FQDN in incoming
option 39.
Sometime later, the client asks for only an IA_PD and does not include
option 39 (FQDN). Therefore, Kea does not complain about the FQDN
being partial. Kea sends back the correct hostname as created with
the reservation + the qualifying suffix. A subsequent full SARR
sometime later results in Kea again sending the correct FQDN again.
I do not know the reason for this behavior; why it failed on the
initial SARR. I had a look at
https://www.rfc-editor.org/rfc/rfc4704.html which describes option 39,
and expected server and client behavior but nothing jumped out at me.
It may be correct behavior based on a subtle reading of the RFC, or it
may be a bug.
In any case, I guess it works after a short time... If you think it
is a bug perhaps you could open an issue in Gitlab:
https://gitlab.isc.org/isc-projects/kea/-/issues
Thank you,
Darren Ankney
More information about the Kea-users
mailing list