[c-nsp] Large GLBP installations

[Ross Vandegrift]

On Tue, Jun 10, 2008 at 06:27:03PM +0200, Jerome Covini wrote:
> One of my customers has been running into a very similar problem after 
> 1.5 year of GLBP stability.
> Running Quad SUP720-3B's + PFC 3B's onto 12.2.18SXD=>8 IIRC, GLBP 
> running in round-robin mode for server LB'ing.

This sounds exactly like the situation we ran into.  Was there a
catalyst (pun kind of intended :) for the beginning of the issue?  In
our situation, nothing spurned it.  No config changes happened at the
same times, nothing was added, just *poof*, and we had a thousand GLBP
groups constantly changing state.

Have you tried adjusting any of the GLBP timers?  Another person
suggested doubling the hello interval, which was something I was going
to try a while back.

I think we're going to move back toward HSRP anyway - we use MST to
split traffic across uplinks, so GLBP actually makes half of our
traffic switched non-optimally.  We've wanted to move away for a
while and Cisco's inability to help with this wasn't 

> Can you let me know if a Cisco bugid has been already created following 
> your TAC request ?

No - Cisco has no idea what's going on, no Bug ID.  Only got a few
wild-ass suggestions.


-- 
Ross Vandegrift
ross at kallisti.us

"The good Christian should beware of mathematicians, and all those who
make empty prophecies. The danger already exists that the mathematicians
have made a covenant with the devil to darken the spirit and to confine
man in the bonds of Hell."
	--St. Augustine, De Genesi ad Litteram, Book II, xviii, 37