question about how a particular dig works ...

Kevin Darcy kcd at chrysler.com
Tue Sep 18 17:19:39 UTC 2012


On 9/18/2012 12:59 PM, M. Meadows wrote:
>
> Thanks Kevin. I understand how the chained alias works. Sorry, I 
> didn't explain my question very well.
>
> I can see that the 8.8.8.8 google public dns server gets an answer.
>
> I know that this domain has a cname coexisting with an SOA record and 
> NS records ... both of which I have read are a bad thing. And I've 
> seen the other reply that indicates that this combination of records 
> in a zone file wouldn't even load in BIND ... so it's done with some 
> other more forgiving DNS app.
>
> What I also see (but failed to explain) is that we have a local 
> nameserver that can't find an answer to the dig www.careerone.com.au 
> query. Gets no record back. Our local nameserver is an AD server that 
> just throws up its imaginary hands in despair. So is this what we 
> should expect from this problematic DNS setup in the 
> www.careerone.com.au zone file? Erratic or somewhat erratic results? 
> Just curious why google and some other public facing dns servers get 
> an answer when our own local nameserver can't figure it out.
>
>
>
> ------------------------------------------------------------------------
> Date: Tue, 18 Sep 2012 11:18:58 -0400
> From: kcd at chrysler.com
> To: bind-users at lists.isc.org
> Subject: Re: question about how a particular dig works ...
>
> On 9/18/2012 9:45 AM, M. Meadows wrote:
>
>
>     dig www.careerone.com.au <http://www.careerone.com.au> +short @8.8.8.8
>     www.careerone.com.au.edgesuite.net
>     <http://www.careerone.com.au.edgesuite.net>.
>     a903.g.akamai.net.
>     208.44.23.99
>     208.44.23.121
>
>     Why does the above dig work when
>
>     dig careerone.com.au +nssearch @8.8.8.8
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server usw1.akam.net in 106 ms.
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server usw4.akam.net in 136 ms.
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server usc4.akam.net in 124 ms.
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server usc1.akam.net in 40 ms.
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server usw5.akam.net in 190 ms.
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server ns1-24.akam.net in 171 ms.
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server asia1.akam.net in 161 ms.
>     SOA dns0.news.com.au. hostmaster.news.com.au. 2012082200 3600 1200
>     86400 1200 from server ns1-50.akam.net in 161 ms.
>
>     shows 8 auth nameservers for careerone.com.au
>
>     and if you use
>
>     dig www.careerone.com.au <http://www.careerone.com.au> +short
>     @<any of the 8 auth nameservers>
>
>     you get no answer.
>
>     How does that work? Where does the 8.8.8.8 google public dns
>     server get its answer from?
>
> www.careerone.com.au <http://www.careerone.com.au> is an alias 
> (through chained aliasing) ultimately to a903.g.akamai.net. To get an 
> authoritative answer for a903g.akamai.net you'd need to ask one of the 
> g.akamai.net nameservers. Which is presumably what Google's public 
> resolver did to get the answers it returned to your query.
>
It's possible that the CNAME-and-other error is causing Microsoft DNS to 
choke. Can you dump the cache on that box and see if it has records 
other than the CNAME?

Another possibility is that the CNAME chain is giving Microsoft DNS 
fits, although Akamai has been doing that for years and it seems to 
mostly work, despite being technically a violation of standards.

I don't know much about troubleshooting Microsoft DNS resolution 
problems, and I doubt many others on this list can (or would be willing 
to) help either. Maybe try a Microsoft list?

                                                     - Kevin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/bind-users/attachments/20120918/9b08b60c/attachment.html>


More information about the bind-users mailing list