[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: "DNS-4to6 automatic mapping" - No more DNS-ALGs & 6then4 DNSqueries
fails to query AAAA records unless the application explicitly requests it?
How hard are we trying to make the transition?
*sigh*
--
Todd Fries .. todd@fries.net
_____________________________________________
| \ 1.636.410.0632 (voice)
| Free Daemon Consulting, LLC \ 1.405.227.9094 (voice)
| http://FreeDaemonConsulting.com \ 1.866.792.3418 (FAX)
| "..in support of free software solutions." \ 1.700.227.9094 (IAXTEL)
| \ 250797 (FWD)
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
37E7 D3EB 74D0 8D66 A68D B866 0326 204E 3F42 004A
http://todd.fries.net/pgp.txt
Penned by Remi Denis-Courmont on 20071129 12:17.15, we have:
|
| On Wed, 28 Nov 2007 22:26:21 -0600, "Todd T. Fries" <todd@fries.net> wrote:
| > I'm sorry, turning off IPv6 routing does not imply the userland should
| not
| > and cannot query for v6 addresses. That's just bogus. The operating
| > system
| > can inform the application that the v6 address has no route, and the
| > application can try the next address, just as if it were a v4 host with
| > multiple v4 addresses, and the first one is not reachable, only faster,
| > because no default route means the application will get notified much
| > sooner.
|
| Like it or not, just about every major dual-stackable operating system
| "fail" to query AAAA RRs if the host has no IPv6 address except for
| loopback and link-local ones.
| There are simply way too many broken DNS servers and somewhat-DNS-aware
| middleboxes in the field that screw up on AAAA queries.
|
| Besides, if an application really wants to resolve a name to an IPv6
| address, it can probably request it explicitly (e.g.
| hints.ai_family = PF_INET6 with getaddrinfo()). Still, from a
| connectivity perspective, you would normally not be using DNS for
| loopback (::1). And you simply cannot use DNS for link-local
| addressing, since link-local addresses require a host-specific
| scope, which DNS cannot provide.
|
|
| Since this part of the problem is essentially fixed in the field,
| I think the main remaining issue is what Nathan describes - lame
| automatic 6to4 deployment.
|
| --
| R??mi