IPv6 site snapshot

Doug Barton dougb at dougbarton.us
Thu Apr 28 02:57:02 CEST 2011


On 04/27/2011 17:27, Jared Mauch wrote:
>
> On Apr 27, 2011, at 8:16 PM, Doug Barton wrote:
>
>> On 04/27/2011 13:25, Jared Mauch wrote:
>>> So, some further research here.
>>>
>>> 1) Appears there is a FreeBSD bug with getaddrinfo() when it
>>> queries for AAAA when it points to a cname.
>>>
>>> http://www.freebsd.org/cgi/query-pr.cgi?pr=156684&cat=
>>
>> It's not clear to me what you think the bug is. I looked at the
>> www.apple.com example in your PR, the ones you listed below, and
>> spot-checked a bunch of the hostnames in your freebsd.errors list.
>> They fall broadly into 2 groups, either an empty CNAME chain or
>> ANSWER=0. In either case, what do you think getaddrinfo()'s
>> appropriate course of action should be given that there is no AAAA
>> record to return? Please understand, this is not a snark, if there
>> is a bug we want to fix it, but I'm genuinely confused.
>>
>
> It should return "hostname nor servname provided, or not known" vs
> "Non-recoverable failure in name resolution".  The second error
> indicates there is some other upstream error, eg SERVFAIL with the
> resolution.

Ah, well I think that depends on what name server you're querying too. 
I'm using BIND 9.8.0 locally, and quite a few of the hostnames you 
provided actually do return SERVFAIL for me. Meanwhile I'm not sure I 
agree with your assertion that an empty CNAME chain should fail with 
EAI_NONAME vs. EAI_FAIL. I'm going to follow up on bind-users at isc.org 
regarding that question. I am however prepared to be wrong. :)


Doug

-- 

	Nothin' ever doesn't change, but nothin' changes much.
			-- OK Go

	Breadth of IT experience, and depth of knowledge in the DNS.
	Yours for the right price.  :)  http://SupersetSolutions.com/



More information about the ipv6-ops mailing list