[c-nsp] Catalyst 6500 Supervisor Engine Redundancy

[Dale W. Carder]

On Oct 30, 2006, at 11:11 AM, Jared Mauch wrote:
> On Mon, Oct 30, 2006 at 05:05:06PM +0000, Sam Stickland wrote:
>> Lasher, Donn wrote:
>>> I would only offer one caveat as to Redundancy mode discussions.
>>>
>>> For software-related failures, SSO may actually hurt you more  
>>> than it
>>> helps.

The party line is to utilize redundant gear instead of redundant sups
if you (can, and if you) had to pick one:

"When designing a network for optimum high availability, it is
tempting to add redundant supervisors to the redundant topology
in an attempt to achieve even higher availability. However,
adding redundant supervisors to redundant core and distribution
layers of the network can increase the convergence time in the
event of a supervisor failure."
From:  http://www.cisco.com/application/pdf/en/us/guest/netsol/ns432/ 
c649/cdccont_0900aecd801a8a2d.pdf

>>>
>> Can anyone else offer any thoughts on this subject? From our  
>> prespective
>> we hardly ever see any supervisors fail
We found that very early generation hardware tended to have more  
problems
than their hardware version 1.1 counterparts, but even some of those
problems were often fixed in software.

> you should really
> be running one of the modular versions of software on the
> 65xx/76xx (aka "76k").

> This will help mitigate the risk from this
> in providing you 1) core files and 2) significantly reduced risk of
> a catastrophic SW failure causing you to POST all your hardware again.

I feel that any gain is currently offset by the additional risks
from new code paths and new bugs in the modular image.  Hopefully in
a year or so this won't be the case (and we could run v6 & mpls).

Dale

----------------------------------
Dale W. Carder - Network Engineer
University of Wisconsin at Madison
http://net.doit.wisc.edu/~dwcarder