Different TTL Behavior from Microsoft DNS

[phn at icke-reklam.ipsec.nu]

Martin McCormick <martin at dc.cis.okstate.edu> wrote:
> 	When one queries a BIND  DNS, the TTL in the response is
> whatever TTL that particular zone has in it for that particular
> record.  Yesterday, I was trouble-shooting some weirdness related to a
> Microsoft DNS which is part of an Active Directory setup.  The record
> was an MX record which the MS DNS had received from our master bind
> DNS with a TTL of 600 seconds.

> 	When I used dig to query the MS box, I got a correct answer
> but the TTL was less than 600 seconds.  I queried again and the number
> was even less than before.  I then got it through my skull that
> every query to the MS DNS returned a TTL based upon the number of
> seconds left in the MS box's cache of the record.

> 	I wouldn't exactly call this wrong, but it is sure different
> than what bind does.  It would appear that if 100 clients requested
> that record from the MS system, all their TTL timers should time out
> at the same second as the TTL timer on the Microsoft DNS.

> 	Is this behavior significant other than the fact it is different
> than what I expected to see?

What you see is expected. Your M$-DNS was caching a RR and should as
a non-authorative dns-server reduce the TTL When TTL becomes zero
it should throw away the cached data (and get a fresh copy if needed)


Asking an authorative NS will always give the "original TTL" 


> Martin McCormick WB5AGZ  Stillwater, OK 
> OSU Information Technology Division Network Operations Group


-- 
Peter Håkanson         
        IPSec  Sverige      ( At Gothenburg Riverside )
           Sorry about my e-mail address, but i'm trying to keep spam out,
	   remove "icke-reklam" if you feel for mailing me. Thanx.