[dns-operations] Questions on DNS Flag day 2020 proposal

[Paul Vixie]

On Thursday, 4 July 2019 23:16:59 UTC Mark Andrews wrote:
> > ...
> 
> Well RFC 1123 really fails to enumerate all the cases where TCP is
> necessary.
> 
> While it mentions RR too big it fails to mention referrals that are too big
> but they are just as important. Glue records are not optional.  Referrals
> should be setting TC=1 if glue does not fit.  The same applies to priming
> query responses.

i think there's a distinction between nec'y glue (owner names at or below this 
delegation point) which as you say should cause TC=1 if it won't fit, vs. 
unnec'y glue (which can be retrieved elsewhere if nec'y) where TC should not 
be affected if it won't fit.

in the binary orbit case where A.X is served by NS.B.Y and B.Y is served by 
NS.A.X, the result should be eventual SERVFAIL by the iterator, rather than 
either delegation including an AAAA or A RRset for names not at-or-below the 
authority who is issuing the delegation.

-- 
Paul