[dns-operations] glitch on [ip6|in-addr].arpa?

[Viktor Dukhovni]

> On Oct 11, 2019, at 2:21 PM, Paul Vixie <paul at redbarn.org> wrote:
> 
> i think there are 13 names each having an A and an AAAA. so, 26 candidate addresses. most resolvers will try them all and home in on the one with the lowest RTT. if one of the 13 it tries via IPv6 doesn't answer, it won't affect operations. in fact, one or more are unreachable from random places almost always, and the system is designed with that in mind. (for example, the use of UDP means unreliability is in-scope.)

Fair enough, and I don't have much "standing" in this case,
my server prefers IPv4, and I have a slave copy of both the
root and arpa zones...

I still feel (on behalf of others) that the situation is
unfortunate, though of course by no means dire.

There are likely more noticeable consequences for some users
than C-root unreachability, where the work-around isn't just
an occasional rare short lookup delay.

-- 
	Viktor.