reverse resolution failing

[Mark Andrews]

In message <B64A959A-A3CC-494D-A262-D739163BEA7C at kumari.net>, Warren Kumari wri
tes:
> Hmmm
> 
> So, for many years I've wondered this. I've looked a little myself, and 
> spoken to a few folk, but never gotten a really satisfactory answer -- 
> maybe there just isn't one
> 
> You often see freaky things like the above (actually, this one is only 
> somewhat freaky), where nameserver implementation behave in bizarre / 
> insane manners. For many of these it seems as through the NS is simply 
> bored, and misbehaving for entertainment purposes ("Whhheee! I'll reply 
> to all A queries with TXT answers, lets see what they make of that!", or, 
> my favorite, "Whatever folk ask, I'll reply with a cname containing the 
> query name, but with alternate labels swapped...").
> For example, even if I tried (well, without much hacking) I cannot figure 
> out how to duplicate Saturn's behavior
> What is it about the DNS protocol that leads to this? Do folk look at all 
> the existing auth server offerings and decide "No, don't like those, I'll 
> just write my own in haskel!"?
> Or is is simply that nameservers themselves are malicious bastards? Are 
> they all conspiring against us? Is there some very subtle joke that we 
> are not seeing?
> 
> W

They want a load balancer but don't want to pay the money to get a
product that does it correctly so they look around for cheaper
alternatives which often don't cover all of the protocol.

The other thing that is common is the operator can't configure the
load balancer correctly.

Mark
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org