ECS4100 Series CLI Reference Guide-R07

– 491 –
17 MAC-Thrashing Commands
A MAC-thrashing feature can be configured on the switch to detect traffic with
identical MAC addresses being received on different physical ports. Once detected,
the system can then take a configured action on a port. In this way, MAC-thrashing
can detect switching loops and take an appropriate action in order to eliminate the
loop condition and prevent CPU overload caused by the constant switching of the
learned MAC address in the Forwarding Database (FDB).
A loop may result in an previously learned MAC address being received from a
second station. Consider the following situation, switch 1 has learned, and
recorded in its FDB, the MAC address of a particular device it receives from switch 2.
However, because of a looping condition, station 2 also forwards the same packet
to station 3 which subsequently forwards the packet to switch 1. Switch 1 now
receives the same MAC address on a a second physical port, causing MAC-thrashing
to be detected. If the algorithm is enabled, MAC-thrashing will take action on the
last inbound port (in this case the port receiving switch 3’s traffic) effectively ending
the loop and discarding the packets that initially triggered the action.
When the MAC-thrashing function is enabled and MAC-thrashing is detected, one
of the following actions will take on the inbound port:
◆ stop-learning: the port stops learning new MAC addresses.
◆ port-disable: no packets can be transmited and received (the physical port can
be UP or DOWN).
◆ link-down: the port’s physical link is disabled.
While the MAC-thrashing action is in effect, the port’s LED will be toggled (i.e. green
to amber or amber to green). The MAC-thrashing resulting action will be invalid
after a configurable period. The action can be immediately cancelled by disabling
the MAC-thrashing function.
Note: If the port-disable action is enabled then it can also be cancelled by
removing and re-inserting the Ethernet cable connected to the port.