[c-nsp] Cosmetic bug or unsupported NPE?

[Pete Templin]

Ted Mittelstaedt wrote:

> The devil is in the details here.  You wern't the only one to make the
> assumption
> that everyone is running an S release.

This is the cisco-nsp mailing list, not the cisco-user mailing list. 
There would be a tendency for readers to use software geared for NSPs, 
though whatever code works is apparently fair game.

> I've never run an S release very long on my bgp routers.  I found years ago
> that Cisco regarded service provider feature sets as licenses to experiment
> because they figured that anyone running an S release must know what they
> are doing.

Any router with a console port provides enough rope to hang one's self. 
  Whether Cisco is playing with you or not, ymmv.

> Now, I'll admit that things might be different now but I've preferred my bgp
> routers to have the capability for uptimes in the years, not in the months.  Not
> that any of them ever reached that long before having to be rebooted to load some
> security update, but we can always hope the crackers might get tired one day.

I've had plenty of routers on S versions with uptimes north of a year, 
including some 7507s with RSP4s (i.e. 256MB).

All that said, please, take a deep breath and relax.  We all make 
assumptions, and perhaps they got in the way.  In a previous thread 
here, I said to someone "why on earth would you need >128MB on your VIPs 
anyway?", since mine are only using 67MB with the relevant CEF tables 
for full-table BGP.  Turns out the other person had a very large 
internal network, and sure enough their internal routes were plentiful 
enough to push the limits of 128MB on VIPs.  Oh well, oops, sorry.

pt