I'm not in the environment right now but this one is a bit interesting. A 4507 has been running in the core of this network just fine for a while now. Something happened over night that caused it to stop routing... kinda. Basically, the only issue seems to be with routes that are redirected to a different address within the same network.
Example. The switch has an SVI in the VLAN and is acting as a default gateway for the VLAN. The switch is peering via EIGRP with a few other devices in the same VLAN. A packet destined for a different network might come into the switch at 10.0.0.1 but be redirected to 10.0.0.254 to reach a different network. This isn't working. If I were to change a host to use the .254 address as the default gateway, traffic to and from that network works. But when the switch is the default gateway, it doesn't.
This isn't a configuration issue since it has been running for well over a year without issue.
I'm not sure what this type of routing would be called... maybe hairpin "ish"? I've considered clearing the routing table but that would be service disruptive. I've thought about clearing the cef table to see if that would fix things but I'm not sure how disruptive that would be.
Any ideas on way to fix this without reloading the entire switch, and what could be causing this? Hasn't happened before.
(Also, this switch is slated for refresh soon since it is way EOL)
