Redirect DNS Requests

Genco YILMAZ gyilmaz at mw.net.tr
Thu Jan 15 23:38:20 UTC 2004


hi,
redirection dns request to another dns server  doesn't cause a problem ?
What happens, when destionation address in the udp packet and the source 
address in the reply packet differ ?
Doesn't it cause a problem ?
just a curiosity ..

thanks


Kirk Strauser wrote:

>--=-=-=
>Content-Transfer-Encoding: quoted-printable
>At 2004-01-15T05:50:54Z, "Matt" <nospam.hciss at yahoo.com> writes:
>
>  
>
>>Would it be possible to use NAT to redirect all UDP port 53 packets to
>>AT&T's DNS servers?
>>    
>>
>
>Yes.
>
>  
>
>>Would this cause total chaos?
>>    
>>
>
>No, just some random, irreproducible chaos.  The correct solution is to get
>the customers to fix their IP addresses.  If they're using dialup, then they
>should be getting those IPs from the PPP server.  If they're using static
>DNS settings, then they should be technical enough to update the settings
>themselves.  If you *have* to, though, couldn't you selectively redirect
>queries to the known-bad servers to the new servers and let everything else
>go as expected?
>
>  
>
>>We have one user that is running a W2K box that they use as there own DNS
>>server and it goes directly to the root servers.
>>    
>>
>
>There was a thread about this on Slashdot last weekend.  Here's that thread,
>starting with my comments:
>
>    http://slashdot.org/comments.pl?sid=3D92245&cid=3D7937746
>
>I *think* I was correct in my assertion, at least to a first approximation
>(I didn't want to muddy the waters with outlier situations).  Is there any
>reason for end users not to use their ISP's DNS as forwarders other than
>when the ISP's servers provide bad service, or for testing?
>=2D-=20
>Kirk Strauser
>In Googlis non est, ergo non est.
>
>--=-=-=
>Content-Type: application/pgp-signature
>
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.2.4 (GNU/Linux)
>
>iD8DBQBABtMN5sRg+Y0CpvERAsUMAJ9r+zsy7CocNt5jUKN0TlMgNNq/NACgicOc
>pCkLCxQ+GLSwX69+IDWd46c=
>=aqFE
>-----END PGP SIGNATURE-----
>--=-=-=--
>
>  
>




More information about the bind-users mailing list