[dns-operations] Why would a recusrive caching server not resolve a CNAME?

[Mohamed Lrhazi]

We had a little mail outage which turned out to be caused by one of our
caching DNS servers returning the bellow incomplete reply.

Clearing the cache on the problematic server fixed the issue....

Am thinking it is now impossible for me to find the root cause in this
instance... but wondering if you guys could hint at what could cause such a
problem... bugs in the DNS servers involved? temporary misconfig at
Google's servers? network issue?

The setup is a bit convoluted:

cache server --> resolver cache server --> Internet

The fix was clearing at the first server. so I am guessing at some point
the resolver gave the incomplete answer.

Thanks a lot,
Mohamed.

➜  ~  dig mail.google.com @141.161.100.201

; <<>> DiG 9.9.5-3-Ubuntu <<>> mail.google.com @141.161.100.201
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 20414
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 4, ADDITIONAL: 5

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;mail.google.com. IN A

;; ANSWER SECTION:
mail.google.com. 10213 IN CNAME googlemail.l.google.com.

;; AUTHORITY SECTION:
google.com. 96485 IN NS ns2.google.com.
google.com. 96485 IN NS ns3.google.com.
google.com. 96485 IN NS ns4.google.com.
google.com. 96485 IN NS ns1.google.com.

;; ADDITIONAL SECTION:
ns3.google.com. 108462 IN A 216.239.36.10
ns4.google.com. 108462 IN A 216.239.38.10
ns1.google.com. 108462 IN A 216.239.32.10
ns2.google.com. 108462 IN A 216.239.34.10

;; Query time: 22 msec
;; SERVER: 141.161.100.201#53(141.161.100.201)
;; WHEN: Sun Jul 06 12:42:09 EDT 2014
;; MSG SIZE  rcvd: 207
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dns-oarc.net/pipermail/dns-operations/attachments/20140706/35d953f9/attachment.html>