can't sync two slaves to master

[Kevin Darcy]

Len Conrad wrote:

> >In article <4.2.2.20000120163917.0423cf00 at mail.Go2France.com>,
> >Len Conrad  <lconrad at Go2France.com> wrote:
> > >additional info:
> > >
> > >turned on logging and in both the FreeBSD non-syncing slaves, I get:
> > >
> > >"Err/TO getting serial# for ....." for every slave zone.
> > >
> > >What's that and how to fix it?
> >
> >It means that the slave got an error or timed out while trying to get the
> >SOA record from the primary.  Make sure that you can connect from the
> >slaves to the primary on TCP port 53.
>
> ============
>
> I can traceroute, ping, and telnet from both non-syncing slaves to the
> master (one is on same enet switch as the master and the NT BIND machine
> that slaves perfectly, the other BIN is one hop removed), and from a 3rd
> machine I've verified the there is a tcp service behind port 53 on the master.

What happens if you do a TCP AXFR query on the slave, using the master as its
server, in some tool like nslookup or dig? *That* is what needs to work for zone
transfers to happen. Is the master limiting zone transfers, by any chance?

> also, what does a BIND log line mean in this format:
>
> XX /156.21.1.2/as2.meiway.com/A

It's just a query log entry for 156.21.1.2 doing an "A" record query to your
nameserver for "as2.meiway.com". Depending on what version of BIND you're using,
the absence of a "+" after "XX" may indicate that it was an iterative as opposed
to a recursive query.


- Kevin