[Dnsmasq-discuss] dnsmasq does not respond correctly on loopback(lo)

[richardvoigt at gmail.com]

On Wed, Oct 13, 2010 at 11:21 PM, Frederick C. Damen <fred at damen.org> wrote:

>  On 10/13/2010 04:16 PM, richardvoigt at gmail.com wrote:
>
>
>
> On Wed, Oct 13, 2010 at 2:20 PM, <fred at damen.org> wrote:
>
>> I now see the option in the installed man page.
>> There is no mention of this option in my /etc/dnsmasq.conf file.
>>
>> I do not suspect that this is the issue though:
>> Each computer, the server in question included, all have the same name(s),
>> i.e. interfaces are not named separately.
>>
>
>  localise-queries couldn't care less about the interface names.  It tries
> to return an address in the same subnet as the address of the interface
> where the query came in.  So for queries coming in on lo0, it would try to
> return addresses in the 127.x.x.x block.
>
> Assuming localise-queries is turned on:
> If I read the man page correctly, if I had the same DNS name associated
> with multiple ip addresses and one of these ip address
> is on the same subnet as the subnet of the interface that the request came
> in on then dnsmasq would return the address on this
> subnet,  But if there are no single name to multiple ip address
> associations then return the ip address associate with the name
> even if the subnets do not match.
> This is the situation that I am having trouble with.  DNS name printer1 is
> assigned one and only one address, 192.168.0.x.
> Thus dnsmasq should return this address no matter what subnet the request
> came in on.
>
> Besides the /etc/dnsmasq.conf does not enable localise-queries and the man
> page seems to suggest that this feature is disabled by default.
>
> Anyhow, how do I get printer1 to be resolved on SERVER without having to
> specify the interface by which to do the DNS name resolution?
>

Initially, use either netstat or the dnsmasq logs (check after you query,
not just the startup information) to verify that your request actually is
going to dnsmasq and not some other dns daemon.  The dnsmasq logs might also
contain other clues as to what is happening.



>
>
>
>
>> Computers severed by dnsmasq both for DHCP and DNS work as expected,
>>  i.e. the network printer 192.168.0.x, DNS name 'printer1' gets its name
>> through dnsmasq DHCP server on 192.168.0.1,
>>       computer nick 192.168.0.y who has its DNS server set in
>> /etc/resolv.conf to 192.168.0.1 can ping printer1, printer1.damen.org.
>> The SERVER 192.168.0.1(and external interface) with its first nameserver
>> set to 127.0.0.1 and following nameservers set to external name
>> servers in its /etc/resolv.conf can not ping printer1, printer1.damen.org.
>>  On the SERVER if you direct the DNS queries
>> to go directly to 192.168.0.1 using dig's @ option the DNS name printer1
>> can be resolved correctly.  Going directly to
>> DNS server at 127.0.0.1 using dig does not resolve the DNS name.
>>
>> Before I added 127.0.0.1 as the first nameserver in /etc/resolv.conf the
>> name resolution for DNS names external to my LAN was slow and after
>> adding it external name resolution when fast,  and the fact that 'dig @
>> 127.0.0.1 printer1' does not indicate inability to contact the
>> DNS server, I suspect that dnsmasq is acting as a DNS server on 127.0.0.1.
>>  But why does dnsmasq not resolve the names in its DHCP
>> configuration on the 127.0.0.1 interface?
>>
>
>  You might use netstat to find out for sure what process is listening on
> 127.0.0.1:53
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/attachments/20101014/a2c35e49/attachment.htm