[Kea-users] Failure Scenario Evaluation

[Darren Ankney]

Hi Diogo,

You haven't missed anything.  This is the current intended behavior.
There is a new Gitlab issue about this here:
https://gitlab.isc.org/isc-projects/kea/-/issues/3790 that may
interest you.

Thank you,
Darren Ankney

On Mon, Mar 10, 2025 at 8:46 PM Diogo André de Assumpção via Kea-users
<kea-users at lists.isc.org> wrote:
>
> Hi All,
>
> I’m evaluating possible failure scenarios with Kea (currently testing with version 2.2.1). One specific scenario I’m going through is with two servers set up as load-balancing HA pair. Some setup details:
>
> - Servers are hosted at different locations
> - Clients are in a third location, using relays to interact with DHCP
> - Using MT and dedicated listener
> - Servers1 and Server2 are able to communicate and exchanged  heartbeats without interruptions
> - Clients can obtain v4 and v6 addresses and classification is even
>
> What I’m testing:
>
> Network failure scenario where Server2 is no longer receiving DHCP traffic from relays. Server 1 continues to receive traffic from relays. Servers 1 and 2 continue to receive heartbeats from peer.
>
> In a scenario like this, I expected that server1 would detect a partner down due `secs` or  `elapsed-time` exceeding `max-ack-delay`, and `max-unacked-clients` also exceeding the max value set. But I’m not seeing that and clients that were previously classified and allotted to server2 do not receive any DHCP response after the simulated network failure.
>
> Is it a wrong assumption? Am I missing something in the configuration?
>
> Sample config attached.
>
>
> Thanks,
> – Diogo
> --
> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
>
> To unsubscribe visit https://lists.isc.org/mailman/listinfo/kea-users.
>
> Kea-users mailing list
> Kea-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/kea-users