Message-ID: <rt-4.4.1-2462-1501213157-1586.44790-0-0@isc.org>
References: <RT-Ticket-44790@isc.org>
 <rt-4.2.8-78137-1488400667-1945.44790-7-0@isc.org>
 <rt-4.2.8-35554-1494912921-1112.44790-7-0@isc.org>
 <20170621174717.GA29267@jurassic>
 <rt-4.4.1-2689-1498067250-1790.44790-0-0@isc.org>
Content-Type: text/plain; charset="utf-8"
X-RT-Interface: Web
X-RT-Original-Encoding: utf-8
X-Mailer: MIME-tools 5.508 (Entity 5.508)
Content-Transfer-Encoding: binary
Content-Disposition: inline
MIME-Version: 1.0
In-Reply-To: <rt-4.4.1-2689-1498067250-1790.44790-0-0@isc.org>
RT-Send-CC:
Content-Length: 472

On Wed Jun 21 07:47:30 2017, muks wrote:
> * Why is stale-answer-ttl clamped to a minimum of 1? Should 0 not be
>   allowed? TTL=0 seems more correct when serving stale answers as they
>   will not be cached, but maybe a non-zero TTL would be required for
>   some cases. (This was pointed out to DCL too as a a review comment on
>   the draft.)

When servers are daisy chained we want some caching to occur and
when the answer is direct to client 0/1 makes no difference.