Extreme Networks ExtremeWare XOS 11.3, Manual

Page: 399 / 698

Fault Detection and Recovery
ExtremeWare XOS 11.3 Concepts Guide
399
Polling
The master node transmits a health check packet on the control VLAN at a user-configurable interval
(see
Figure 18 ). If the ring is complete, the master node receives the health-check packet on its
secondary port (the control VLAN is not blocked on the secondary port). When the master node
receives the health-check packet, it resets its failtimer and continues normal operation.
If the master node does not receive the health check packet before the failtimer interval expires and the
failtime expiry action is set to
open-secondary-port
, it declares a “failed” state and performs the same
steps described above:
●
Unblocks its secondary port for access by the protected VLANs.
●
Flushes its forwarding database (FDB).
●
Sends a “flush FDB” message to its associated transit nodes.
Restoration Operations
The master node continues sending health check packets out its primary port even when the master
node is operating in the failed state. As long as there is a break in the ring, the fail period timer of the
master node continues to expire, and the master node remains in the failed state.
When the broken link is restored, the master receives its health check packet back on its secondary port
and once again declares the ring to be complete. Again, the master node logically:
●
Blocks the protected VLANs on its secondary port.
●
Flushes its FDB.
●
Sends a “flush FDB” message to its associated transit nodes.
During the time between when the transit node detects that the link is operable again and when the
master node detects that the ring is complete, the secondary port on the master node is still open and
data could start traversing the transit node port that just came up.
To prevent the possibility of a such a temporary loop, when the transit node detects that its failed link is
up, it performs these steps:
1 For the port that just came up, put all the protected VLANs traversing that port into a temporary
blocked state.
2 Remember which port has been temporarily blocked.
3 Set the state to Preforwarding.
When the master node receives its health check packet back on its secondary port and detects that the
ring is once again complete, it sends a message to all its associated transit nodes to flush their
forwarding databases.
When the transit nodes receive the message to flush their forwarding databases, they perform these
steps:
1 Flush their forwarding databases on the protected VLANs.
2 If the port state is set to Preforwarding, unblock all the previously blocked protected VLANs for the
port.