Problems With IXFR

[Smith, William E. (Bill), Jr.]

I already have an options statement with the transfer-format statement there
to apply globally.  I verified that in fact the named-xfer is the one we
build with 8.2.3. I verified this using strings command and grepping for
8.2.3 which returned

@(#)named 8.2.3-REL Tue Feb 13 13:44:09 EST 2001
humesdg1 at wishbone.jhuapl.edu:/u
sr/local/sparc-sun-solaris2.6/obj/opt/bind/bind823/bin/named
named 8.2.3-REL Tue Feb 13 13:44:09 EST 2001
8.2.3-REL


I've gotten ixfr to work pretty much w/out problem under BIND 9 but we're
still a little ways from that. I was hoping to get things working under
8.2.X in the interim

Any other suggestions would be appreciated.  Otherwise, I guess I'll just
have to hold off until we're to BIND 9

Bill

-----Original Message-----
From: Cricket Liu [mailto:cricket at nxdomain.com] 
Sent: Friday, October 26, 2001 12:10 PM
To: comp-protocols-dns-bind at moderators.isc.org
Subject: Re: Problems With IXFR


> Oct 25 12:28:16 apldns1.jhuapl.edu named-xfer[679]: [128.244.197.22] 
> bad respons e for _sites.jhuapl.edu, SOA query got rcode 0, aa 1, 
> ancount 13, aucount
0
> Oct 25 12:28:18 apldns1.jhuapl.edu named-xfer[680]: [128.244.197.22] 
> bad respons e for ssd.jhuapl.edu, SOA query got rcode 0, aa 1, ancount 
> 400, aucount 0

Looking at these errors again, it appears that the master is sending IXFRs
(hence the ancount is larger than expected) but the slave isn't interpreting
them correctly.  Are you sure this is the named-xfer that you built with
BIND 8.2.3?  Do you have

server <slave's IP address> {
    transfer-format many-answer;
};

in the master's named.conf?  If not, from the CHANGES file, it looks like
you'd better upgrade to at least 8.2.5.

cricket

Men & Mice
DNS Software & Services
www.menandmice.com