On Mon, Oct 04, 1999 at 07:09:40PM -0600, Mark Collette wrote:
[I wrote this bit:]
> > Ugh. Assuming the last one maps to the machines real name, keep that and
> > ditch the rest. Then make {,ns,www,ftp}.h.a.c CNAMEs for othello in the fwd
> > zone file. I suspect a glance through the BIND documentation and the RFCs
> > might be instructive.
OK - time for me to grovel: I should have checked that the RFCs I was
thinking of said what I thought they said (they don't) _before_ I shot
my mouth off, rather than after (I've just done it). RFC 2181 (sec 10.2)
says:
"Confusion about canonical names has lead to a belief that a PTR
record should have exactly one RR in its RRSet. This is incorrect,
the relevant section of RFC1034 (section 3.6.2) indicates that the
value of a PTR record should be a canonical name. That is, it should
not be an alias. There is no implication in that section that only
one PTR record is permitted for a name. No such restriction should
be inferred."
Having said that, I can't believe there are many situations where you would
need or want to do that. The only one I can think of is if you have, say
asdf.com. SOA ..
IN NS blah
IN A 1.2.3.4
www IN CNAME blah
blah IN A 1.2.3.4
which you have to do if you want the domain name to also be a host - e.g.
for "Information Superhighway surfers" (pass the barf bucket) who can't cope
with actually typing in a domain name fully. Anyway, to be strictly accurate
you'd want the 4.3.2.1.in-addr.arpa PTR's to list blah.asdf.com. /and/
asdf.com. You might also need it if you use the schemes laid out in RFC1101
-- does anyone do this?
> I read the DNS Howto and DNS Docs that kernelnotes.org links to, and they
> said not to use CNAMEs, and alluded to some long running arguement. Before
> reading them I was using CNAMEs. Hmmm... your way does sound much simpler.
It does, doesn't it? <g> And I can't imagine that there a significant no.
of resolvers out there that are still broken w.r.t. CNAME records. In fact I
hope and pray that there aren't, 'cos working CNAME following behaviour is
needed for sane management of CIDR subnet's *.in-addr.arpa. info, I believe[0].
RFC 2219 describes using CNAME records to alias well known service names to
machines; it does say (sec. 5)
"It isn't a simple matter of recommending CNAMEs over A records. Each
site has it's own set of requirements that may make one approach
better than the other. RFC 1912 [RFC-1912] discusses some of the
configuration issues involved in using CNAMEs."
So I'm going to go and re-read 1912 <g>
Once again, sorry for going off half-cocked first time.
[0] There's an RFC for this too, I believe. 2317?
--
Fortune: You will be attacked next Wednesday at 3:15 p.m. by six samuri
sword wielding purple fish glued to Harley-Davidson motorcycles.
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
