EtherSwitch & ARP

EtherSwitch & ARP

Post by haim [howard] rom » Mon, 08 Jul 1996 04:00:00

Here is our network.  Host names are in quotes.

        -----------     -----------     ---------------
        "albert"        "adam"          "ts"
        Sun clone       Sun clone       Terminal Server
        (Solaris 2)     (Solaris 2)     (Bridge Co.)
        -----------     -----------     ---------------
            |               |               |
            |               |               |
                        Kalpana EtherSwitch Pro16
                        Fan-out Repeater
                    |               |               |
                    |               |               |
                -----------     -----------     ----------
                "bertha"        "betty"        "cisco"
                Sun             Sun             Cisco 2514
                SunOS 4.1.x     SunOS 4.1.x     Router
                -----------     -----------     ----------

The problem is that the hosts on top (the Sun clones & the terminal
server) cannot connect to the other hosts, even by ping.  However, the
other hosts CAN connect to the hosts on top!  What happens is that the
first ping causes the bottom host to add the top host to its arp
table.  I'll illustrate with an example:

        pinging albert --> bertha does NOT work

        bertha does NOT appear in albert's arp table

        pinging bertha --> albert DOES work

        bertha now appears in albert's arp table

        now pinging albert --> bertha WORKS

The same thing happens between albert & the Kalpana, & between albert & adam.

Here is the configuration of the Kalpana:

        all ports are in the default domain.

        IP information:

                the Kalpana, & all the hosts shown are on the same
                subnet.  All have the same subnet address (

                the Kalpana's default gateway is the Cisco

                the IP state is "BootP When Needed"

                Telnetting to the Kalpana works (from the bottom hosts)

        There are NO filters:

                no filters are defined for any MAC addresses

                the Port Security Mode for all ports is "Normal"

        The switching mode for all ports is default:

                Auto Switching Mode

                50% Error Water Mark

                Runt-free Mode off

At first, the Kalpana did NOT participate in the Spanning Tree
Algorithm.  At someone's suggestion, I had it participate.  It did not
help.  Here is the configuration for that:

        Participate in Spanning Tree                 Yes
        Switch Priority                              0
        Switch Hello Time (in Seconds)               2    (default)
        Switch Maximum Message Age (in Seconds)      20   (default)
        Switch Forward Delay (in Seconds)            15   (default)

        the cost for all of the ports is the same (they are all the
        same speed)

        the priority for the port connected to the repeater is 10.
        The priority for the others is default: 128

I would prefer responses by e-mail, since our site does not
necessarily receive articles from all of these newsgroups reliably.  Thanks.
Haim (Howard) Roman     |  Computer Center


EtherSwitch & ARP

Post by Christopher S. Youn » Tue, 09 Jul 1996 04:00:00

One thing I would check would be the state of the port on the switch
that the hub is connected to.. This would been to be in 'backbone mode'
unless that switch supports multiple MAC addresses per port.  Anytime
you have a switch connecting a hub instead of an individual NIC, you
need to either have a segment switch (which can support multiple MACs on
a single port) or a workgroup switch (which usually can only support one
or two MACs per port) that has that port configured to flood all data to
it.  Ie. backbone mode, or un-switched.

Hope this helps...

Christopher Young
Sr. Network Analyst
I-NET, Inc.