[Dnsmasq-discuss] client using bootp request gets address from dhcpd, but not dnsmasq

[Alec Habig]

Hi folks,

Been using dnsmasq for years, thank you!  First time I have a question
that I haven't been able to figure out from the docs.

I have a client trying to boot like so:

# tshark -Y "eth.addr == 08:00:30:4c:58:47"
Capturing on 'eno8303' 
 1205 9.134265135 10.204.88.71 → 255.255.255.255 BOOTP 342 Boot Request from 08:00:30:4c:58:47 (NetworkR_4c:58:47)

I haven't been able to get an otherwise perfectly functional dnsmasq
instance to respond.

However, ISC dhcpd works: responds, gives an address, the client moves
on and is happy.  The client is not asking for tftp info, it's just
using bootp for getting an address.  Thus, if I understand things
correctly, listing a "dhcp-host" for dnsmasq should work, where
"dhcp-boot" is for tftp info?

My initial solution of confining dhcpd to one NIC where the troublesome
client lives (and is about to be joined by 89 similar ones) and letting
dnsmasq continue to run the rest of the cluster doesn't work, as both
daemons insist on binding to 0.0.0.0:67 regardless of config file scope.

Replacing dnsmasq with the set of individual {dhcpd, dns, tftp} daemons
would work but would be a PITA: I'm using dnsmasq for a reason.

So: suggestions for figuring out how to make dnsmasq pay attention to
this bootp request?  Is there debugging info I could turn on to
understand why it's ignoring the broadcast?  At this point I'm not above
recompiling things with a really specific "if" statement :)

	   thanks,
	   Alec

PS: alma9.8 running dnsmasq 2.85-18.el9_8.1 and dhcp-server 4.4.2-20.b1.el9
if that matters, but I suspect this is a more general question.

-- 
			       Alec Habig
		     University of Minnesota Duluth
		     Dept. of Physics and Astronomy
			     ahabig at umn.edu
		   http://neutrino.d.umn.edu/~habig/