On Tue, 29 May 2012 21:33:13 +0200
Per Jessen
Carl Hartung wrote:
On Tue, 29 May 2012 19:33:59 +0200 Per Jessen
wrote: [snip]
We're not talking about responses going over IPv6, but about responses with AAAA DNS records (which happen to be IPv6 addresses). The actual communication might happen over IPv4, -5 or -6 or even carrier pigeon (RFC1149). :-)
I get this part, Per. I really do - at least I /thought/ I did. This explains why, when I run a query via cli from my IPv4 network, I get back a listing containing both addresses (if the IPv6 address exists.)
Exactly.
But that's a diagnostic procedure, right?
Nope.
It isn't an exact replacement for the 'live' transaction that occurs between, say, a browser on an IPv4 network and a DNS server configured for, and residing on, an IPv6 network, right? IOW, in everyday operation, if a request arrives at an IPv6 enabled server on IPv4 port 53, the fact that it arrives at the IPv4 address results in the IPv4 address, alone, being returned - the rational being it is presumed that the client is unable to use the IPv6 address.
Am I way out past left field here today? :-)
Yeah ... how a DNS query arrives will at most determine how the response will be sent back. Whether a DNS query arrives via IPv4 or IPv6 does not in any way determine what the response contains.
So the server sends back two records instead of one, and the client just discards the one it can't use. That's not very efficient (stupid server!) Thanks for the clarification ;-) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org