[c-nsp] How to bring one link down if another related link goes down

[Oliver Gorwits]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 28/09/2010 12:38, Alan Buxey wrote:
> its trunk failover of link-state tracking.  dont seem to have solidified the
> name  - appeared in IOS 12.2(25) - but has been around in the blade switches
> for a little longer...and is in Nexus platform too for added bonus
> 
> I'm thinking of a scenarios...a nice real scenario...where this would be useful
> rather than using spanning-tree and normal backup links..

I think we had one a while ago with the Microsoft NLB (Load
Balancing). From what I recall heartbeats are not checked on all
interfaces (in Server 2008).

So if an upstream link on a top of rack switch goes down NLB will
keep the local servers in the hash-pool causing dropped packets for
connections hashing to those servers.

This kind of link state tracking might improve on that by taking
down the links to the servers, too. I didn't test it, though. I
think Server 2010 might be smarter so it was either a self shaving
Yak, or SEP to fix ;-)

regards,

- -- 
Oliver Gorwits, Network and Telecommunications Group,
Oxford University Computing Services
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkyjb18ACgkQ2NPq7pwWBt6QmwCeL5eYACPhUa0TGo061OwBwG1z
aHYAoMI56i39MN2d27Iqwen5U2u3XSxd
=8olD
-----END PGP SIGNATURE-----