Joe Abley wrote: > On 30 Sep 2015, at 12:53, Paul Hoffman wrote: > > >I'll add the v4/v6 wording to the post-IESG-review draft unless there is > >objection in the WG. > > I like the v4/v6 wording, for what that's worth. > > >John Levine just answered your question about why the address might > >already be in use, which was something that was brought up in the early > >discussion of this draft in the WG. It means that you can't run both this > >and some other DNS-listening task on ::1, whereas you can run both on > >different addresses in 127/8. We'll cover that in the new wording. > > Since a single operator controls both ends, there's no need to use a > well-known port. If you can't bind to 127.0.0.1:53 because something else is > already listening there, use 127.0.0.1:12345.
Hi, Joe: If I'm not mistaken, this would depend on the support in the recursive implementation for sending queries to non- well-known ports. Appendix B gives an example Unbound configuration which supports this (you append @<port> to the IP address), but AFAIK the example BIND configuration only supports querying the "static-stub" servers on the well-known port. -- Robert Edmonds _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop