Common Link Failures; Flushing The Fdbs - Extreme Networks ExtremeWare XOS Guide Manual

Ethernet Automatic Protection Switching

Common Link Failures

If a single common link fails, the configured controller (S1) and partner (S2) take steps to prevent a
superloop.
Assuming there is a single data VLAN configured on all three EAPS domains, the controller (S1) keeps
one port open (called "Active-Open"). The remaining segment ports are "blocked" to prevent a
superloop.
In Figure 20, P2 is the "Active-Open" port on S1. Ports P3 and P4 are "blocked." The master nodes (S5,
S8, and S11) open their secondary ports.
Figure 20: EAPS domain common link failure
S3
S4
EAPS1
S5
Master
When the common link is restored, the controller goes into Preforwarding state. After the controller
receives notification from the master nodes that they have converged and blocked their secondary ports,
the controller opens all ports.
If you have an EAPS configuration with multiple common links and a second common link fails, the
controllers continue to take steps to prevent a superloop. In addition to having one controller with an
"Active-Open" port, the controller with the smallest link ID becomes the "root blocker." There can be
only one "root blocker" in the network.

Flushing the FDBs

When a controller goes into or out of the "blocking" state, the controller sends a "flush fdb" message to
flush all of the FDBs of the switches in its segments. Each switch in the path of the "flush fdb" message
flushes its FDB.
In a network with multiple EAPS ports in the blocking state, the "flush fdb" message gets propagated
across the boundaries of the EAPS domains.
ExtremeWare XOS 11.1 Concepts Guide
P3
P2
Active-Open
S1
P4
Controller
P1
x
EAPS2
S2
Partner
S9
EAPS3
S6
S10
S7
S8
Master
S11
Master
EW_102b
284