On Thu Apr 12 Kevin wrote:
>
> In message <200104122046.NAA00368@dakkon.darykon.cet.com>,
> "Gerald C. Simmons" writes:
> >I have a feeling Layer 2 of the OSI model bit me.  I was unconnecting
> >and re-connecting var ious 10Base-T connections to my NetBSD box,
> >when I ran into the situation that eventually the NetBSD box stopped
> >communi cating with several of these links. What was wierd was I
> >could ping these devices, but complex IP protocols would fail.
> 
> >Is there some unwritten IP/Ethernet rule that says you shouldn't
> >be connecting things with non-unique IP addresses, but with unique
> >Ethernet addresses back and forth??? I never had two devices with
> >the same IP connected to the NetBSD box at the same time, but it
> >didn't seem to help.
> 
> Hmmm.  Sometimes you have to flush out the ARP tables in cases like
> that, although that doesn't explain ping-able systems that can't
> be reached via TCP.  Weird.  You might try flushing the route table 
> next time ("route -f") and reinstalling whatever routes seem appropriate.
> 
> Good luck,
> 
> Kevin
> kml@dotrocket.com
> 

Ya Know, I thought of flushing the route tables with route -f. But, that had no effect.

How do you flush the ARP tables???

Gerry
simmons@darykon.cet.com