Quick question about Host records

[Bill Moseley]

At 10:09 AM 08/09/00 +0100, Jim Reid wrote:

Hi Jim,

Maybe it's too early for me as I'm not following your point below.  Could
you clarify?  I'm not clear if the "potential problem" you mention is a
result of having two name servers names share the same IP, or by having NSI
automatically change the name of name servers so there's only one unique
name per IP number.


>>>>>> "Don" == Don Stokes <don at news.daedalus.co.nz> writes:
>
>    Don> Last time I looked, which was some time ago, NSI (and I'm not
>    Don> sure about other registrars) key name server addresses on IP
>    Don> address, and don't allow duplicates.
...
>    Don> Annoying.
>
>Maybe. But if they do this - I don't know or care - it prevents an
>even more annoying problem. Suppose ns1.example.com was renumbered.
>In the scenario you describe all that means is one glue record in the
>.com zone is changed: the new A record for ns1.example.com. All the
>other .com, .net and .org delegations that used ns1.example.com will
>automatically go to the new IP address for that server. This is far
>preferable to going through every delegation and changing each glue
>record that might or might not be affected. It also saves subtle cache
>pollution problems when glue for ns1.example.com in the parent zone,
>.com, differs from what's in the child zone, example.com. Multiply
>that by every delegation that could have its own NS and glue record
>for the IP address of ns1.example.com and you should see the scale of
>that potential problem.
>
>
>

Bill Moseley
mailto:moseley at hank.org