Westermo RedFox Series User Manual
Page 332
Westermo OS Management Guide
Version 4.17.0-0
To mitigate issues with flapping uplinks, e.g., caused by bad cables, dual-homing
nodes and ring coupling nodes can be configured to use a sticky uplink, as op-
posed to the deterministic uplink election described above. With sticky uplink en-
abled, the priority component of an uplink’s cost vector is reduced with a given
value (the adjustment value) once that link is elected as active. That is, with
sticky uplink configured, the effective priority of an uplink can differ from the
configured priority.
Example
Consider three uplinks with same speed and duplex. Link A has ”priority
100”, link B has ”priority 110” with ”adjustment 20”, and link C has
”priority 120” with ”adjustment 40”. All nodes keep information about
each-others announced link cost (100, 110 and 120).
If Link A goes
down, link B will take over as it has lower (i.e., better) priority than link
C (110<120), and link B will decrease its effective priority to 90 in its an-
nouncements.
If link A comes up again, link B will continue to be active as ”90<100”. The
mechanism works in the same way for dual-homing, even though priority is
never ”announced” to any other node.
15.1.4
Handling Multicast
To provide fast fail-over of multicast traffic, FRNT Ring Coupling and Multi-Link
Dual-Homing uplinks are added to the list of multicast router ports, see
. This is both done at the Ring Coupling nodes and Dual-Homing
nodes, as well as on switches on the remote side of the uplink
. This means that
all layer-2 multicast traffic is always sent over the uplinks, even if IGMP snooping
is enabled.
2
An exception is when connecting a Dual-Homing uplink to a non-FRNT switch, the fail-over of
multicast traffic will instead occur on the next reception of an IGMP Report (if IGMP snooping is
enabled). See also
332
➞ 2015 Westermo Teleindustri AB