[Dnsmasq-discuss] Static IP client question

[Albert ARIBAUD]

Hi Chris,

Le Sun, 7 Aug 2016 16:36:11 +0100
Chris Green <cl at isbd.net> a écrit:

> On Sun, Aug 07, 2016 at 05:25:24PM +0200, Albert ARIBAUD wrote:
> > Chris Green <cl at isbd.net> a écrit:  
> > > On Sat, Aug 06, 2016 at 08:57:43PM -0400, Edward Crosby wrote:  
> > > >    I've implemented Dnsmasq on a Raspberry Pi 3 running Ubuntu
> > > > Mate 16.04 on my home LAN. I have configured it as a DHCP
> > > > server also. I have quite a few clients on my LAN, most of them
> > > > are DHCP clients. I have one PC, my personal PC, that has a
> > > > static IP address. This PC does not resolve host names of other
> > > > host on my LAN, it doesn't even resolve the hostname of the
> > > > Dnsmasq DNS server, even though I have the Dnsmasq server IP as
> > > > my DNS server.    
> > > 
> > > I'm doing almost exactly the same as you.
> > > 
> > > What I do is fix the IP address of my desktop machine by getting
> > > dnsmasq to always give it the same address.  So leave your desktop
> > > with a dynamic IP in its configuration and have something like the
> > > following to your /etc/hosts file on the pi:-
> > > 
> > >     127.0.0.1       localhost
> > >     #
> > >     #
> > >     # These have fixed IP for various reasons, so dnsmasq serves
> > > their IP from here       
> > >     #
> > >     192.168.1.1     vigor
> > >     192.168.1.2     pi.zbmc.eu raspberrypi  pi
> > >     192.168.1.3     esprimo.zbmc.eu zbmc.eu
> > >     192.168.1.5     maxinexp
> > >     192.168.1.6     ben
> > >     192.168.1.40    mikrotik
> > >     192.168.1.60    fonera
> > > 
> > > My desktop machine is esprimo.  
> > 
> > Hi Chris and Edward,
> > 
> > I don't have many Windows machines around my dnsmasq managed,
> > statically attributed LAN, but at least one is used daily, and I
> > just ran a test on it, which makes me ask a question to Edward:
> > 
> > How did you (both) test that the machine does not resolve?
> > 
> > Here's why I'm asking:
> > 
> > I've tested the following with a Windows machine (let's call it
> > romulus) and my Linux machine (let's call it remus). From the
> > command prompt on romulus, I ran the following:
> > 
> > 	ping remus
> >   
> > 	=> romulus complains that it could not find host remus.  
> > 
> > 	nslookup remus
> >   
> > 	=> This resolves immediately to remus' fixed IP address.  
> > 
> > 	ping remus.  (note the dot at the end!)
> >   
> > 	=> This works.  
> > 
> > If I'd believed the first ping, I might have concluded that the
> > resolution did not work, while it actually does; the issue is with
> > how romulus handles domainless names.
> > 
> > Hence my question re: how exactly the issue was tested.
> >   
> I've not used nslookup in years, it's deprecated now.  I use 'host' or
> 'dig' if something else doesn't work (e.g. an ssh to somewhere).

I'd use dig if that were provided on Windows machines. :)

> I'm not sure why you get the symptoms you're seeing though Albert, I
> don't think I've ever noticed anything like that.  If I try 'ssh
> something' and it doesn't work then 'host something' doesn't work
> either! :-)

I'm not sure why you think I am seeing symptoms here :) -- as far as
I'm concerned, the Windows machines on my network work fine [enough for
their users]. I am just giving an example to show that depending on how
one tests things, one may get to a wrong conclusion, and therefore,
that explaining how one runs a test is as important as running it.

Amicalement,
-- 
Albert.