A record of domain name must be name server ?

Kevin Darcy kcd at chrysler.com
Wed Sep 10 21:13:59 UTC 2014

On 9/10/2014 11:58 AM, Alan Clegg wrote:
> On 9/10/14, 8:42 AM, Sam Wilson wrote:
>> And you could reduce maintenance very slightly by replacing
>> www             in      A
>> with
>> www             in      CNAME   @
> And now you have an MX record, 3 NS records and a bunch of other crap
> associated with the WWW address.
And why is that a _bad_ thing?

If I ever change that IP, I want to change it in *one*place*. The CNAME 
allows everything to automatically follow that change. Why necessitate 
multiple updates when a single update will do? If TTL-manipulation is 
necessary in order to minimize caching complications, the number of 
RRset updates is magnified, of course.

MXes and NSes are a non-issue, IMO, since the contexts in which people 
look up a "www" name (usually end-users trying to access a website) are 
usually quite disjoint from the use cases of MXes (automated systems 
delivering mail) or NSes (nameserver-to-nameserver traffic). I see 
little or no risk of confusion or misdirection.

I suppose it's _possible_ that some day a mail sender might mistype a 
recipient as user at www.example.com instead of (as they should have) 
user at example.com, and maybe in that scenario the CNAME will cause the 
recipient address to show up in the headers of the received message in 
an unexpected way. But, to me, this falls under the generic category of 
GIGO (garbage in, garbage out) -- you type something wrong into a 
computer system, you might not get the results you expected...

                             - Kevin
> _______________________________________________
> Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe from this list
> bind-users mailing list
> bind-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/bind-users/attachments/20140910/d8d658cd/attachment.html>

More information about the bind-users mailing list