To:
Pekka Savola <pekkas@netcore.fi>
Cc:
Brad Knowles <brad.knowles@skynet.be>, Jun-ichiro itojun Hagino <itojun@itojun.org>, Mohsen.Souissi@nic.fr, dnsop@cafax.se, namedroppers@ops.ietf.org, ngtrans@sunroof.eng.sun.com, ipng@sunroof.eng.sun.com, vladimir.ksinant@6wind.com, rfc1886@nic.fr, g6@g6.asso.fr
From:
David Terrell <dbt@meat.net>
Date:
Sun, 14 Jul 2002 23:55:48 -0700
Content-Disposition:
inline
In-Reply-To:
<Pine.LNX.4.44.0207150848470.31868-100000@netcore.fi>
Reply-To:
David Terrell <dbt@meat.net>
Sender:
owner-dnsop@cafax.se
User-Agent:
Mutt/1.4i
Subject:
Re: RFC 1886 Interop Tests & Results
On Mon, Jul 15, 2002 at 08:49:22AM +0300, Pekka Savola wrote: > On Mon, 15 Jul 2002, Brad Knowles wrote: > > At 6:35 AM +0900 2002/07/15, Jun-ichiro itojun Hagino wrote: > > > > > the co-existence of ip6.int and ip6.arpa tree will require us to: > > > query ip6.arpa; > > > if (no record) > > > query ip6.int; > > > for backward compatibility. was it taken into account, or did you > > > test just "ip6.arpa" lookups? > > > > I checked the source code for BIND 9.2.1, and IIRC it checks > > ip6.int first and then ip6.arpa second. This allows us to stand up > > ip6.arpa whenever, and then once that is set, we can tear down > > ip6.int. > > FWIW, e.g. Linux glibc resolver only checks ip6.arpa now, so you'd better > start standing up.. 2.0.0.2.ip6.arpa: NXDOMAIN e.f.f.3.ip6.arpa: NXDOMAIN That's probably 70-80% of all IP6 deployments reachable via the public ipv6 internet (granted, especially most 2002:: folks don't have reverse DNS set up yet, but plenty do...). It would be reasonable for glibc to at least make fallback to ip6.int an option... -- David Terrell | "Any sufficiently advanced technology Prime Minister, Nebcorp | is indistinguishable from a rigged demo." dbt@meat.net | - Brian Swetland http://wwn.nebcorp.com/