Implementing a RFC 2317 subnet zone with Digital ULTRIX 4.5

[Barry Margolin]

In article <a99vda$pea at pub3.rc.vix.com>,
Andreas Meile <andreas at hofen.ch> wrote:
>Dear ULTRIX and DNS users
>
>I'm using a broadway connection with 8 fixed IP addresses as my future
>Internet connection solution in the next time. To be able to maintain the
>reverse DNS entries by myself, I intend to use RFC 2317. I also intend to
>use an old DECstaton 5000/25 running ULTRIX 4.5 (RISC) with applied Y2K
>patches as my secondary DNS server.
>
>Today, I set up a small RFC 2317 test environment where I simulated this
>situation. I used a SuSE Linux 6.4 (i386) system (BINDv4) as the ISP's DNS
>replacement, i.e. this box simulated the ISP's DNS server. When I set both
>the DNS zones to the Linux, it worked fine. But when I tried to move the
>test DNS zone (for example "56/29.77.168.192.in-addr.arpa") to the ULTRIX
>machine, I got "*** No address (A) records available for ..." errors,

It's hard to tell, since you didn't post your zone files or the complete
message, but it sounds like the NS records in the subdomain delegation are
pointing to names that don't exist.

>although every NS record is set accordingly, i.e. the "56/29 IN NS" entry
>explicitely points to the ULTRIX machine. I applied a "ls -d
>56/29.77.168.192.in-addr.arpa." against both DNS servers but the output
>didn't show any significant difference, except that ULTRIX shows additional
>A entries for the authoritative DNS server's IP addresses.
>
>Questions:
>1.) RFC 2317 mentionnes that some older DNS server implementations don't
>support the CNAME delegation trick. Is ULTRIX 4.5's "named" daemon affected
>by this issue or not?

Maybe, but the error message you mentioned above doesn't sound like it has
anything to do with whether the server understands RFC 2317.  It's
complaining about a nonexistent A record, not a nonexistent PTR record.

>2.) The ULTRIX Y2K update kit contains an updated /usr/etc/named file (size
>524'936 bytes, date November 5th, 1996. Original file: 327'680 bytes,
>October 18th, 1995). When I try to use the newer one, named immediately
>crashes after the first name resolution query while the original version
>works fine. Anybody who knows that problem, too?

Those are both pretty old versions of BIND.  I suggest you abandon the
vendor-supplied version of BIND and download the latest from www.isc.org.

-- 
Barry Margolin, barmar at genuity.net
Genuity, Woburn, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.