Multiple rarpd on same ethernet?

[mcvax!solaris!cyliax at uunet.uu.net]

It seems that something gets confused when several machines are running
rarpd on the same net. When a client trys to boot, it will frequently hang
just after it discovers its own IP address and it trys to tftp the boot
program.  After a while something will time-out and the client continues
to boot, but this may take several minutes. If you kill the running
rarpd's on all the machines except one (it doesn't matter which one), the
client will reboot normally. All the servers are in the same YP domain,
and run 4.0 with selected 4.0.1 updates.  In addition all of them have
identical /tftpboot dirs, so it shouldn't matter which server it gets the
boot code from.  Has anyone seen this, and know why it's doing it ?

	Thanks in advance,
	Ingo Cyliax

                    Ingo Cyliax    Stabstelle Software
cyliax%inf.ethz.ch at relay.cs.net    Department fuer Informatik
               cyliax at ethz.uucp    ETH-Zentrum
           Tel: +41 1 256 35 31    CH-8092 Zuerich, Switzerland