[c-nsp] 2950-48-SI arp problems ?

[McLean Pickett]

Sounds like another ARP timeout problem. Try setting the ARP cache/CAM
timers to the same time on the gateway and the switches. You could also
set a static cam entry or run a script to generate traffic from the host
every few minutes. 

You might also check spanning tree to make sure it is behaving properly.

McLean

-----Original Message-----
From: cisco-nsp-bounces at puck.nether.net
[mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of matthew zeier
Sent: Tuesday, December 21, 2004 12:09 PM
To: cisco-nsp at puck.nether.net
Subject: [c-nsp] 2950-48-SI arp problems ?

I have two 2950-SIs running Version 12.1(22)EA2.  Switch config is
pretty
simple:

core1/core2 (6905) -- dot1q trunk -- access4 (6905) --- trunk --- 2950 .


Hosts on the 2950 are able to connect to the Internet, however, if idle
for enough time (something between minutes and less than an hour), the
host's ARP table becomes empty and the gateway is unable to ping/reach
the host (so inbound connections are lost).

Running 'tcpdump' on one of the Linux hosts shows that there aren't any
packets from the gateway reaching the host.  If the host pings the
gateway, everything works - inbound connectivity is restored.

I've used three different computers (PC laptop & windows, supermicro &
redhat, mac) with the same problem.  If I replace the 2950 with a 2948,
the problem appears to go away.

Did I miss the "ip work-all-the-time enable" option?  The 2950's trunk
looks
like:

interface GigabitEthernet0/1
 switchport trunk native vlan 2
 switchport mode trunk
!




--
matthew zeier, Sr. Network Engineer  | "Nothing in life is to be feared.
InteleNet Communications, Inc.       |  It is only to be understood."
(949) 784-7904                       |       - Marie Curie 

_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/