Is this standard behavior?

[Barry Margolin]

In article <ab9lup$82a3$1 at isrv4.isc.org>,
Fred Viles <fv+abuse at nospam.epitools.com> wrote:
>Rasmus Aaen <ra at back-bone.dk> wrote in
><ab8oko$7bau$1 at isrv4.isc.org>: 
>
>> We are running Bind 8.3.1 on a couple of Win2K servers for about
>> 200 domains. All domains have the same soa values:
>> 
>> Refresh: 7200, retry: 3600, expire: 864000, min. TTL: 43200
>>
>> I noticed that the slave server updates the "modified" time on
>> the zone files only after the zone has expired and been
>> transfered again. It is my understanding that the modified time
>> on the files should be updated every two hours (7200 sec.) in our
>> case. 
>
>I would not expect that.  I would expect the file time to change only 
>when the zone is updated, which happens only when the SOA serial 
>number is updated on the master or, as you say, after it expires.

On Unix the file time changes every time the slave performs a successful
refresh.  This includes times when it queries the SOA and finds that the
serial number is still the same as the one it has.  The file timestamp is
used to determine when the expiration time has been reached (this
information is kept in named's memory while it's running, but the timestamp
is used to keep track of it across restarts of named).

If the Windows port of BIND doesn't update the timestamp every time it does
a successful SOA query, how does it know when to expire the zone?

-- 
Barry Margolin, barmar at genuity.net
Genuity, Woburn, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.