Sender domain must exist

[Joe Pruett]

a lot of dns setups are broken nowadays.  onelist.com has been acting up
lately.  intel.com also breaks frequently.  there is some situation where
bind will end up with ns records for a domain, but no matching a records
for those ns records, so it can't do any resolving for that domain until
the ns records timeout.  i've been thinking of trying to hack into the
code and recognize that case and trash the ns records and ask the root
servers again.  it always seems to be caused by a mismatch between what
the root servers have and what the supposed authoritative servers return,
but i haven't figured out the exact mechanism that causes it to lose the a
records, but not the ns records.

see, here is intel.com being broken:
; <<>> DiG 8.2 <<>> ns intel.com 
;; res options: init recurs defnam dnsrch
;; got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 4
;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 0
;; QUERY SECTION:
;;      intel.com, type = NS, class = IN

;; ANSWER SECTION:
intel.com.              3h6m56s IN NS   cps.ns.intel.com.
intel.com.              3h6m56s IN NS   ns.intel.com.

;; Total query time: 5 msec
;; FROM: q7.q7.com to SERVER: default -- 206.58.126.2
;; WHEN: Mon Nov 29 07:57:59 1999
;; MSG SIZE  sent: 27  rcvd: 62

and ucdavis.edu:
; <<>> DiG 8.2 <<>> ns ucdavis.edu 
;; res options: init recurs defnam dnsrch
;; got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 4
;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 0
;; QUERY SECTION:
;;      ucdavis.edu, type = NS, class = IN

;; ANSWER SECTION:
ucdavis.edu.            8h35m18s IN NS  teddy.ucdavis.edu.
ucdavis.edu.            8h35m18s IN NS  tyler.ucdavis.edu.

;; Total query time: 4 msec
;; FROM: q7.q7.com to SERVER: default -- 206.58.126.2
;; WHEN: Mon Nov 29 07:58:35 1999
;; MSG SIZE  sent: 29  rcvd: 69