In my environment we have a 4500X as a core switch and 3650's as our distribution/access switches. There are approximately 30 switches total on the campus network. These switches are all in several "loops" that provide failover via STP. Most of these loops have a trunk link connecting the floors to each other and then the top and bottom floors connect to the core.
Every so often we have had issues where a flood of MAC Flap errors and Host Flap errors will overload the processors of several switches, which will eventually cascade back to the core and take it down as well. The last time this occurred we isolated it to a small Netgear switch that was installed to provide an additional port for a printer. Our access ports have portfast enabled and we are using RSTP.
We've been in conversations with Cisco TAC for a while, and they have suggested enabling the storm protection feature for broadcast and multicast traffic. We are beginning to implement this now, but I was curious as to if anyone on this fine forum had encountered any issues like this before and how they solved them.
As a temporary fix we have disabled all redundant links in our network and that has temporarily kept us running. Unfortunately, we are a 24/7/365 healthcare facility and cannot shut down the network for testing and troubleshooting.
