HP3000-L Archives

November 2013, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jeff Kell <[log in to unmask]>
Reply To:
Jeff Kell <[log in to unmask]>
Date:
Tue, 12 Nov 2013 21:43:43 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (33 lines)
On 11/12/2013 9:18 PM, Howard Hoxsie wrote:
> Thank you, Gilles, that's the one.  Apparently memory does not serve very
> well.  :-)

If you have access to your routers/switches, you can also attack this
from the other end :)

Cisco switches/routers (layer-3) have an ARP table that holds the MAC
addresses of the hosts on the subnet.  They are subject to a timeout. 
However, they will generate a "gratuitous unicast ARP" request to any
host in the table 30 seconds before it is scheduled to expire.  If the
host is up, it will respond, and reset the learning timer.  It will ALSO
push the MAC address and IP of the gateway (the router that just
performed the unicast ARP) into the host.

Unfortunately the default ARP timeout on a Cisco device is 4 hours :( 
You may tweak this as desired on a per-interface basis via the "arp
timeout xxx" configuration directive.

If you are doing network monitoring by any SNMP tools, they work best
when the mac-address tables (CAMs/TCAMs/etc) are loaded with the host
values, and the ARP tables are populated with the current hosts.  The
mac-address table timeout default is only 300 seconds (I'd suggest
moving that up, depending on how volatile your connections may be; we
use 600 seconds).  This will cause the "gratuitous ARP" to occur at 530
seconds, and if the host answers, it perfectly repopulates the tables
for your management tools to read.

Jeff

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

ATOM RSS1 RSS2