class B delegation

Andris Kalnozols andris at hpl.hp.com
Tue Dec 18 01:08:27 UTC 2001


> In article <9vla47$n2t at pub3.rc.vix.com>,
> Michael Kjorling  <michael at kjorling.com> wrote:
> >Remember the alias law: a CNAME can not point to a CNAME.
> 
> RFC 1034 seems to describe this as a good idea, but not quite a "law":
> 
>     Domain names in RRs which point at another name should always point at
>     the primary name and not the alias.  This avoids extra indirections in
>     accessing information.  For example, the address to name RR for the
>     above host should be:
> 
> 	52.0.0.10.IN-ADDR.ARPA  IN      PTR     C.ISI.EDU
> 
>     rather than pointing at USC-ISIC.ARPA.  Of course, by the robustness
>     principle, domain software should not fail when presented with CNAME
>     chains or loops; CNAME chains should be followed and CNAME loops
>     signalled as an error.

FWIW, here are BIND's observed limits of chasing the following CNAME chain:

$ORIGIN example.com.
cname-1	   CNAME    cname-2
cname-2    CNAME    cname-3
...
cname-20   CNAME    foo
foo        CNAME    127.0.0.1

BIND 8 will chase 8 CNAMEs - if #8 points to a non-CNAME RR, you'll
get the answer.  If #8 points to yet another CNAME, you'll get the
9 CNAMEs and the response code will be set to SERVFAIL.

BIND 9 will chase 16 CNAMEs - if #16 points to a non-CNAME RR, you'll
get the answer.  If #16 points to yet another CNAME, you'll get the
17 CNAMEs with a NOERROR response code.

Andris



More information about the bind-users mailing list