Secondary DNS servers don't appear to work in Windows clients
kcd at daimlerchrysler.com
Fri Jan 10 18:18:48 UTC 2003
Thomas B wrote:
> Kevin Darcy <kcd at daimlerchrysler.com> wrote in message news:<avkc1l$c89a$1 at isrv4.isc.org>...
> > Thomas B wrote:
> > > I use DHCP and Dynamic DNS to supply IP addresses, default gateway and
> > > DNS servers. We have 2 DNS servers which both appear to work fine.
> > >
> > > But if I disable the first DNS server in the list and try to an
> > > nslookup, the PC responds with:
> > >
> > > C:\>nslookup mq
> > > DNS request timed out.
> > > timeout was 2 seconds.
> > > *** Can't find server name for address 10.1.0.20: Timed out
> > > *** Default servers are not available
> > > Server: UnKnown
> > > Address: 10.1.0.20
> > >
> > Looks like 10.1.0.20 cannot do a reverse-lookup of its own address.
> > nslookup stupidly does a reverse lookup of the nameserver's address
> > before actually looking up what you requested, and will fall on its face
> > if the reverse lookup fails. Dig and other DNS troubleshooting tools
> > don't have this misfeature.
> > Can you ping "mq"? Maybe the nameserver is only hosed with respect to
> > reverse DNS. Maybe it can resolve forward lookups, like "mq", just fine.
> > By the way, you should try to wean yourself from the use of short names.
> > They are evil.
> > - Kevin
> I'm sorry - it looks like I'm not being clear.
> I'm using a user PC to access various resources (Linux boxes,
> internet, ssh service). My IP address is 10.1.5.34 (DHCP).
> The DNS name servers supplied by DHCP are 10.1.0.20 and 10.1.0.91. I
> also set the router, domain name, Wins server, etc... with DHCP.
> The problem occurs when I try to ping or use nslookup or ssh of the MQ
> server (10.1.0.15) from the user PC after I disable 10.1.0.20.
> I'm trying to simulate the automatic use of a slave nameserver after
> the master goes down.
> The order of name servers that the client uses is 10.1.0.20, and then
> 10.1.0.91. But what I'm finding out is that the secondary dns server
> entry is irrelevent, because the client PC fails with error messages
> if the primary name server goes down.
> 1) There is no value to specifying secondary and tertiary name servers
> 2) My configuration is incorrect
> 3) Everything is working fine, but client redirection to secondary
> name servers takes a very long time to switch over (20+seconds).
That would appear to be a client resolver problem rather than a BIND problem. The query returns
immediately if you query 10.1.0.91 directly for "mq", right?
More information about the bind-users