[nsp] High-availabity GigE connection to two switches

Church, Chuck cchurch at wamnetgov.com
Tue Jun 1 15:45:48 EDT 2004


It's no standard, so I'm sure every company does it a little bit
different.  But the 3Com NICs I worked with about 4 years ago running on
NW 5.1 did send keepalives from one NIC to another (don't remember the
protocol it used) every second.  So the loss of primary NIC ethernet
link status is a no brainer - bring up the standby.  But like you said,
if a SUP dies and link stays up, then it's a tough decision who should
be active.  Maybe it looks at traffic on the primary NIC, and then
decides that even though no heartbeats are being received, there is user
traffic, so no failover occurs.  Just a guess though...


Chuck Church
Lead Design Engineer
CCIE #8776, MCNE, MCSE
Wam!Net Government Services - Design & Implementation Team
13665 Dulles Technology Dr. Ste 250
Herndon, VA 20171
Office: 703-480-2569
Cell: 703-819-3495
cchurch at wamnetgov.com
PGP key:
http://pgp.mit.edu:11371/pks/lookup?op=index&search=cchurch%40wamnetgov.
com

-----Original Message-----
From: Gert Doering [mailto:gert at greenie.muc.de] 
Sent: Tuesday, June 01, 2004 3:23 PM
To: Church, Chuck
Cc: Rubens Kuhl Jr.; cisco-nsp at puck.nether.net
Subject: Re: [nsp] High-availabity GigE connection to two switches

Hi,

On Tue, Jun 01, 2004 at 01:07:19PM -0500, Church, Chuck wrote:
> Depending on the NICs and the OS, you may be able to do a fail-over 
> teaming setup.  I know 3Com, Intel, and Compaq/HP support this.  There

> are variations in design, but they work similar to HSRP, where one is 
> active and the others lay and wait for missing keepalives from the 
> active NIC.  Once it appears to be down, a standby NIC takes over and 
> assumes the old NIC's MAC address, so that no ARP re-learning has to 
> occur.

How do these solution detect a "down" situation on the primary link?
Only by monitoring for the ethernet heartbeat, or by actively monitoring
for packets and/or sending null loopback packets?

I'm curious, because "just monitoring for ethernet link" isn't overly
helpful in the case of a switch supervisor breakdown - the modules are
likely to still send a link, while the actual frame forwarding will
break...

gert
--
USENET is *not* the non-clickable part of WWW!
 
//www.muc.de/~gert/
Gert Doering - Munich, Germany
gert at greenie.muc.de
fax: +49-89-35655025
gert at net.informatik.tu-muenchen.de



More information about the cisco-nsp mailing list