two questions

[sharkey@superk.physics.sunysb.edu]

> I understand. However, I've tried three different (identical chipset) cards.

FWIW, the Realtek 8139 is the most problematical chip I have ever used.  I
also have three of these cards, and swapping cards never solves a problem.
I've never observed the type of problem you're suggesting, but I'm almost
ready to trash all of mine, too.  My cards only have problems once or twice
a month, but, that's still too often for me.

> It is completely symmetric:
> 
> A->B ==> A.TX++, B.RX++
> B->A ==> B.TX++, A.RX++

Hmm.  Not likely cabling then.  Have you tried using "tcpdump" to view
traffic on the line?  These machines must have some sort of two way
communication before a ping can actually be sent.  At a minimum you would
need the sender to first issue an arp request and receive an arp reply before
a ping could be sent.  You should be able to see all of that with tcpdump.

Does the hardware address of B get into A's arp cache when you issue a ping?

I don't really have a clue what your problem is, I'm just poking for more
info that might possibly shed some light on the issue.  Donald might just
know everything.  He frequently seems to.

Eric
 | To unsubscribe, send mail to Majordomo@cesdis.gsfc.nasa.gov, and within the
 |  body of the mail, include only the text:
 |   unsubscribe this-list-name youraddress@wherever.org
 | You will be unsubscribed as speedily as possible.