We found that we have MAC flooding as a result of running HSRP and doing equal cost routing with OSPF. This is a known issue with the default ARP timeout of 4 hours and the MAC agingtime of 5 minutes. We have looked into removing asymmetric routing and have implemented ARP and MAC times of 1 hour for both. This has helped the situation, but the MAC is becoming "unlearned" in less than an hour (see attachment). Does a change in MAC agingtime require a reboot on specific or all switch platforms?
Doing unequal cost routing to the active HSRP does not seem like a good solution since it does not do load sharing and would require an OSPF convergence after the primary route went down, opposed to just switching immediatly to the equal route.
I see the MAC is not in the table when the flooding occurs. Could something else besides aging cause the MAC to be unlearned?
