IPv6 failover with isc-dhcpd-4.3.5?

Paul B. Henson henson at acm.org
Mon Feb 13 23:56:51 UTC 2017

> From: Tomek Mrugalski
> Sent: Wednesday, February 08, 2017 8:35 AM
> https://datatracker.ietf.org/doc/draft-ietf-dhc-dhcpv6-failover-protocol/.

Cool; nice to see there is progress being made on that.

> There's no shortage of IPv6 addresses, that's true, but there are other
> reasons why failover could be useful. IPv6 prefixes is the primary one.

We are not currently delegating prefixes in our environment.

> Another is the avoidance of changing addresses for clients, so long

For IPv6, we are giving out a default two-week and maximum four-week lease.
If one of my two servers is down for a month, I've got bigger problems than
a random client getting a different IPv6 address 8-/.

I'm definitely in favor of seeing IPv6 failover get implemented, but I think
in the average deployment it just doesn't have quite the same impetus as
IPv6 failover.

> changes. When ISC decides to implement it, Kea will most likely be used
> as a base. No firm decisions has been made yet, though.

We looked at Kea back when it was still part of bind 10, it seemed overly
complicated. We are still pretty happy with good old dhcpd for now.

Thanks for the info.

More information about the dhcp-workers mailing list