[rbldnsd] Question on requirements for inclusion of 127.0.0.2 in dnsbl list

[David Landgren]

William Leibzon wrote:
> 
> On Mon, 8 Nov 2004, Jon Lewis wrote:
> 
>> Exactly.  The point of having a standard entry such as 127.0.0.2 is that
>> users can test for that entry in a dnsbl to make sure the dnsbl is
>> working.  If that entry is missing, it may be a sign the zone is
>> corrupt/incomplete.
> 
> 
> Ok, I get the point ...
> 
> Now I still would like to find out if this (having 127.0.0.2) is considered
> a normal practice for dnsbl operations or not?

Who's your audience ? It all boils down to a question of shell access to 
the machine.

If you can look at a zone file, you can run queries against any number 
of samples to confirm that rbldnsd is working correctly.

People who don't have shell access have no idea what is in rbldnsd's 
zones. They could try any number of random queries and always receive 
negative replies. They might never know if it would return positive for 
anything.

To confirm that it (or any other dnsbl software) will return positive 
for *something*, 127.0.0.2 has been chosen by some sort of concensus to 
act as a NOP address. You'll always get a positive response back, so you 
can infer that the software is working.

Sort of like the chargen/echo TCP services of the goode olde days.

So, the net at large may use your dnsbl, then yes, 127.0.0.2 is 
considered good etiquette. If it's just for you then it is not strictly 
required.

David