Fwd: Re: Server issues

Neal McBurnett neal at bcn.boulder.co.us
Wed Nov 21 07:37:25 UTC 2007


On Wed, Nov 21, 2007 at 12:04:55AM -0700, Neal McBurnett wrote:
> On Wed, Nov 21, 2007 at 07:28:01AM +0100, Ante Karamati? wrote:
> > On Tue, 20 Nov 2007 19:15:20 -0700
> > Neal McBurnett <neal at bcn.boulder.co.us> wrote:
> > 
> > > I don't really have a well informed opinion on the topic of zeroconf
> > > and/or LLMNR, despite having paid some attention to it.
> > 
> > It's very simple. Both technologies claim one undefined domain. And
> > this discussion went in wrong direction. It's not about LLMNR vs
> > Zeroconf. I'm arguing that *both* of them brake lots of existing
> > networks. .local is undefined domain and thus it is used all around the
> > world on real DNS (like Bind) for small-medium sized local networks.
> 
> Again, I haven't studied the details, but I have read enough to know
> that, despite some claims here to the contrary, this is not true for
> LLMNR.  ".local" appears nowhere in RFC 4795
> http://tools.ietf.org/html/rfc4795

And also note that I know this is still mostly besides the point :-)
And that avahi does have .local problems, at least some of which are
documented by Avahi here:

 http://avahi.org/wiki/AvahiAndUnicastDotLocal

 If Avahi and nss-mdns is installed properly a machine does not
 contact a unicast DNS server when resolving names from the .local
 domain, thus the unicast DNS domain .local becomes unreachable.

Also I've seen, as reported here, that Microsoft is giving bogus
advice on using .local.  And of course all of that is a problem for
avahi and for Microsoft customers....

-Neal




More information about the ubuntu-server mailing list