I have eth0 and using pppd, a sort of virtual node ppp0 is generated, I believe
that something is going wrong in the sense, by default, nsd is going to eth0
when it should go to ppp0.
Can that be the issue?
Would I have to add some routing rules there?
On 4 October 2023 10:38:51 pm IST, Stuart Henderson <stu at
spacehopper.org> wrote:>On 2023/10/04 20:55, Sagar Acharya via nsd-users wrote:
>> Tried with ip-address option, it is not working.
>>
>> I think ip-address binds to local ip address. Anyways localhost binds
to the same ip as that of
>> http, 0.0.0.0 , so everything is right from the system.
>
>If your host has multiple addresses, binding to 0.0.0.0 will often
>not do what you want for UDP. The source address of responses will
>not necessarily be the same address as the query was sent to.
>This is also described in the nsd.conf(5) manual in the description
>of the "ip-address" option.
>
>I suggest listing the specific IP addresses on which NSD will be
>sent DNS queries.
>
>Also use e.g. tcpdump to check that the queries are actually received
>at your machine and not filtered upstream.
>
>> How do I test udp from other network as dig tests over udp. Or can I
dig over tcp?
>
>"dig +tcp" will use TCP for the test, but TCP does not have the
above
>problem on multi-homed hosts.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.nlnetlabs.nl/pipermail/nsd-users/attachments/20231005/87a3d3f6/attachment.htm>