isc-dhcp: should we start phasing it out?

[Mark Shuttleworth]

On 16/05/2022 18:34, Andreas Hasenack wrote:
> Could we perhaps start with phasing out the client, get its rdeps to
> use alternatives, and then stop building it, and eventually get to the
> server? This could be a lot of work, as I said, isc-dhcp is a classic,
> but if upstream is shifting its focus elsewhere, soon we will be
> alone.


Thanks for raising the topic. Upstream is clearly signalling their 
preference for a Kea-only future, so one way or another folks who are on 
ISC dhcpd are going to need to pick a new offering. I moved a network to 
Kea last year because ISC dhcpd HA had a number of unexpected 
behaviours, Kea seems much more rational in that regard and a good fit 
for our own offerings and services such as MAAS. I also think we might 
be able to contribute dqlite support for Kea upstream, which would make 
HA even easier.

Mark