One Domain; Multiple IPs.

[Simon Waters]

Kevin Darcy wrote:
> 
> BTW, try querying www.jeepunpaved.com on the Internet and tell me if your client
> blows up. Sure, neither of its nameservers can answer non-A queries intelligently,
> but why do they *need* to? (Amusingly, Rob Austein's insistence about serial-number
> increment tends to become rather moot when there *isn't* an SOA record, at least
> not any that anyone can successfully query...)

I find that looking up www.jeepunpaved.com is a tad
unreliable. At least the first look up - but I presume this
is something to do with reaching the right place in the
hierarchy, rather than the record itself, as later lookups
seem okay, and with such a low TTL the record itself can't
be the problem.

Hmm - looks like something is odd with your use of the C&W
DNS servers - seems that the domain is delegated to ns, but
is on ns4. Might be worth just running through with doc or
similar (Doesn't DJB do one).

But I suspect maybe one of the links between my server and
the Chrysler servers is overworked or some such - I means
I'm still on 64Kbps *8-(, and it takes 13 hops till it hits
a firewall, and goes from London to New York, via Amsterdam,
I guess it is amazing the Internet ever works at all. I'm
guessing we have DNS servers only on CHRYSLER-NET at some
point - so maybe a bottleneck there depending on
CHRYSLER-NET bandwidth and routing. I guess the C&W DNS
server should sort (well work around) that...

Of course setting such a low TTL, encourages traffic and
this leads to bottlenecks, which was Brad's point. Although
personally I see nothing wrong with using DNS to load
balance, I suspect much beyond round robin leads to small
returns for increased complexity. Our efforts might be
better focused on making sure clients fail-over gracefully
to extra records, so we get fail-over by just having
multiple A records.