[c-nsp] 6500 stable code

[Jon Lewis]

On Wed, 28 Jan 2009, Derick Winkworth wrote:

> You need to reduce the number of features you are actively using in the 
> switch, pick an IOS, put in the lab, fully regression-test it with 
> load-testing (and test all possible functions and failures of all 
> protocols and features) then deploy.
>
> Oh wait... thats not possible is it?  You need to do this soon right?  
> You don't have a year to develop a test plan and scripts and fully test 
> it?  By the time you do that, the software you are testing will have 
> been deferred or have an advisory on it for major defects.

I assume the above is largely sarcasm.  Actually though, I've had this 
switch on the network for several months, participating in OSPF (about 
1k routes), and getting full BGP routes (about 275k), but it hasn't been 
passing any traffic.  Additionally, when I first started setting it up, 
SXF12a was current.

> Well spin the barrel, close your eyes, and pull the trigger.

Why do you think I coined the phrase "ios roulette" years ago?

I've gotten a number of helpful replies (mostly private...guess people are 
hesitant to announce what versions they're running).

I should have mentioned, we also export netflow (sampled), utilize RADIUS, 
and I'll likely occasionally have monitor sessions to sniff traffic for 
troubleshooting purposes.  We're not doing it yet, but at some point I'd 
like (i.e. need) to start doing IPv6.

----------------------------------------------------------------------
  Jon Lewis                   |  I route
  Senior Network Engineer     |  therefore you are
  Atlantic Net                |
_________ http://www.lewis.org/~jlewis/pgp for PGP public key_________