Thanks to all!
The suggestion to copy the whole ampr.org from ucsd.edu and set this
zone as master was the key to success!
Robert
-----Ursprüngliche Nachricht-----
Von: David J Brown <[EMAIL PROTECTED]>
An: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Datum: Dienstag, 28. September 1999 22:48
Betreff: RE: Still need DNS help! (was: HAM related DNS problem)
>
>> -----Original Message-----
>> From: [EMAIL PROTECTED]
>> [mailto:[EMAIL PROTECTED]]On Behalf Of Robert Schelander
>> Sent: Monday, September 27, 1999 11:11 PM
>
>> I would like to have a configuration where I can list the names for local
>> ham stations here (within ampr.org) and forward only if requests for not
>> listed stations should be answered.
>> Setting up my DNS server as master for ampr.org is impossible, because
>> requests for names which are not listed in my ampr.org zone file are not
>> forwarded. Setting up as a slave is impossible too, because zonetransfers
>> between the ampr.org master are out of question.
>
>Why not do what I used to do, which is to FTP the ampr.org zone and reverse
>files from ucsd.edu and install them in my own DNS server. I used to hack
>the SOA etc to point to myself to try and keep things simple. I know this
is
>not a clean and pure way of running DNS, but as most of ampr.org is slow
>changing it worked well with a weekly update.
>
>Another alternative some stations used was to grab the entire ampr.org
files
>and strip out all but local (reachable stations). This was done when 8 megs
>of RAM was considered a lot and the ampr.org took up valuable swap space
:-)
>I must point out that this only really scales when there is no local
>wormhole to allow worldwide connectivity.
>
>If you've a reasonably modern machine you'll have no real trouble having
all
>of ampr.org loaded in your own DNS server. And, should you use the same
>machine for internet stuff too, you'll find that you can retain your own
>copy of ampr.org and still allow your ISP to be your forwarder for
>everything else when online.
>
>(ofcourse, the best solution for slow half duplex networks and memory
>starved machines would be, A. a resolver library with longer timeouts and
B.
>a crippled bind that could look up from a disk based database (as nos
>does/did) ).
>
>David G7PIT
>
>
>
>
