forwarding queries to another server (huh?)

Tomas B. Winkler tomasw at cs.huji.ac.il
Wed Mar 21 10:49:47 UTC 2001



I was thought that there can be only one canonical name for an
interface, but I'll give it a try.
Do I have to provide also reverse mapping or that will create the 
collision ?

Thanks a lot.

Tomas Winkler
The System Group
CSE HUJI

On Tue, 20 Mar 2001, Kevin Darcy wrote:

> 
> Tomas B. Winkler wrote:
> 
> > Thanks a lot Kevin,
> > I still don't get it (sometimes I'm slow). I think I have to add more
> > details:
> > The mycomp.foo.bar.org alread exists in foo.bar.org zone, so Iguess I
> > cannot set test.foo.bar.org to be an A entry (mycomp.foo.bar.orgis
> > supposed to be the dns server for test.foo.bar.org)
> 
> There's no reason why test.foo.bar.org cannot own an A record. That A record would
> just coincidentally have the same address as the mycomp.foo.bar.org A record. Why
> would you think this causes problems? Note that the name you use on the right-hand
> side of an NS record must be a "canonical" name, not an alias. So once you created
> the test.foo.bar.org A record, you could actually use test.foo.bar.org as the
> right-hand side of the NS record for the zone. But if you do this, you'll need to
> create a "glue record" copy of the A record in the parent zone, i.e. in foo.bar.org.
> 
> > I've delegated test.foo.bar.org from the foo.bar.org (probably not
> > correctly)
> 
> Hmmm... Ithought you said you couldn't change foo.bar.org...
> 
> > So I would rerephraze the question: Is there any way I can run the dns
> > server on a computer which cannonical name is not in the domain of which
> > it is master.
> 
> Sure. This is done all of the time. One of the slaves for chrysler.com (and actually
> for *all* of our external domains), for instance, isn't master for the zone which
> contains its own name. This doesn't cause any problems whatsoever.
> 
> > And more then that, each entry is actuall supposed to be
> > a CNAME to foo.bar.org names.
> 
> You can choose any canonical name you want as the target of a CNAME. The main
> restriction on CNAMEs you need to watch out for though is that the name of the zone
> itself (test.foo.bar.org) cannot own a CNAME record, because of the infamous
> "CNAME and other data" rule. Look at the recent archives of this list to find more
> information about this rule (and the arguments for and against it) than you probably
> ever wanted to know. Since you cannot create a CNAME for test.foo.bar.org, you have
> to create an A record instead, and update it whenever it changes.
> 
> 
> - Kevin
> 
> > huh
> >
> > Tomas Winkler
> > The System Group
> > CSE HUJI
> >
> > On Tue, 20 Mar 2001, Kevin Darcy wrote:
> >
> > >
> > > Okay then, change that test.foo.bar.org CNAME to an A record, and get rid of
> > > the "out of zone data" mycomp.foo.bar.org. That should at least allow the
> > > zonefile to load properly as test.foo.bar.org. Note that if mycomp.foo.bar.org
> > > doesn't already exist, and youcannot add it to the foo.bar.org zone, then your
> > > only other option is to define mycomp.foo.bar.org as a zone by itself, and then
> > > explicitly define that zone (as master/slave/stub) on all nameservers which
> > > need to know about it. Ditto for the test.foo.bar.org zone, if you can't
> > > delegate it from the foo.bar.org zone -- you'll have to define it explicitly on
> > > all of the nameservers which need to know about it.
> > >
> > > Maybe it would be easier just to define a "shadow" foo.bar.org zone on all of
> > > those nameservers, and not bother delegating test.foo.bar.org at all. The only
> > > possible drawback is that you might need to keep this "shadow" foo.bar.org zone
> > > in synch with the real foo.bar.org for the duration of the test.
> >>
> > >
> > > - Kevin
> > >
> > > Tomas B. Winkler wrote:
> > >
> > > > Your remark is true, but what I've probably havn't stressed enough that
> > > > what I need is a SEPARATE named to be in charge of testdomain. I cannot
> > > > temper with the main domain name servers.
> > > >
> > > > Tomas Winkler
> > > > The System Group
> > > > CSEHUJI
> > > >
> > > > On Mon, 19 Mar 2001, Kevin Darcy wrote:
> > > >
> > > > >
> > > > > Tomas B.Winkler wrote:
> > > > >
> > > > > > Hi
> > > > > > I have a similar problem I have an existing domain, let say foo.bar.org
> > > > > > and I would like to span on it a sub-domain let say test.foo.bar.org
> > > > > > It wouldn't be so hard but I need a separated named to be a master dns.
> > > > >
> > > > > Huh? Are you under the impression that a nameserver can be master for only
> > > > > one zone? Untrue: a single nameserver can be master for large numbers of
> > > > > zones-- some folks are running thousands of zones from a single
> > > > > nameserver instance. So there's no reason why your nameserver couldn't be
> > > > > master for both "foo.bar.org" and "test.foo.bar.org".
> > > > >
> > > > > But, before you go ahead and configure that, ask yourself whether it's
> > > > > necessary to delegate a subzone for the test.foo.bar.org subdomain at
> > > > >all. You could just add the test.foo.bar.org names to the foo.bar.org
> > > > > zone. Zones and subdomains are not synonymous, although it is true all
> > > > > zones -- other than the root zone, of course -- are distinct subdomains of
> > > > > some parentdomain.
> > > > >
> > > > > > The problem is the computer which runs named has not canonical name in
> > > > > > the test.foo.bar.org domain so the entry is ignored
> > > > > > The dbfile looks
> > > > > >
> > > > > > $TTL86400
> > > > > > @  IN SOA test.foo.bar.org. tomasw.foo.bar.org. (
> > > > > >                                   42            ; serial
> > > > > > (d. adams)
> > > > > >                                   3H              ; refresh
> > > > > >                                   15M           ; retry
> > > > > >                                   1W              ; expiry
> > > > > >                                 1D )            ; minimum
> > > > > >
> > > > > >   1D IN NS        mycomp.foo.bar.org.
> > > > > >
> > > > > > localhost       1D IN A         127.0.0.1
> > > > > > mycomp.foo.bar.org. 1D IN A       xx.xx.xx.xx
> > > > > > test.foo.bar.org. 1D IN CNAME     mycomp.foo.bar.org.
> > > > > > zzz.test.foo.bar.org. 1D IN CNAME mycomp.foo.bar.org.
> > > > >
> > > > > Okay, now you've really lost me. Is this the zone file for foo.bar.org or
> > > > > test.foo.bar.org? You didn't say. If it's the foo.bar.org zone file, then
> > > > > the "test.foo.bar.org" entries in it should work fine *without* the need
> > > > > for any subzone delegation, as discussed above. On the other hand, if it's
> > > > > the test.foo.bar.org zonefile, then it's pretty hopeless: not only does
> > > > > the "test.foo.bar.org" CNAME conflict with the "test.foo.bar.org" SOA and
> > > > > NS records, but the "mycomp.foo.bar.org" entry doesn't belong in the zone
> > > > > file at all.
> > > > >
> > > > > By the way, regardless of how the zone is loaded, you shouldn't be using
> > > > > an alias in the MNAME field ("test.foo.bar.org") of your SOA RR. You
> > > > > should beusing the canonical name ("mycomp.foo.bar.org") instead.
> > > > >
> > > > >
> > > > > - Kevin
> > > > >
> > > > > > On Mon, 19 Mar 2001, Pumpkinhead wrote:
> > > > > >
> > > > > > > Hi,
> > > > > > >
> > > > > > > I have an internal domain, say foo.bar.com.I have a name server
> > > > > > > behaving properly for it.Call it nsfoo.
> > > > > > >
> > > > > > > Now I have set up a test environment domain, say test.bar.com.Note
> > > > > > > that it's not a subdomain of the first.
> > > > > > >
> > > > > > > I want a name server (nstest) on test.bar.com to handle its own
> > > > > > > queries.That is, when nsfoo receives those queries it makes nstest
> > > > > > > deal with them.
> > > > > > >
> > > > > > > In named.conf on nsfoo I have the following entry:
> > > > > > >
> > > >> > > zone "test.bar.com" {
> > > > > > > type forward;
> > > > > > > forward only;
> > > > > > > forwarders{
> > > > > > >         172.xxx.xxx.xxx;
> > > > > > > };
> > > > > > > };
> > > > > > >
> > > > > > > But alas it doesn't work.Do I need to configure my forward RR's ?
> > > > > > >
> > > > > > > Thanks in advance,
> > > > > > >
> > > > > > > Peter
> > > > > > >
> > > > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > >
> > >
> > >
> > >
> > >
> 
> 
> 
> 
> 



More information about the bind-users mailing list