A trunk's allowed-VLAN list blinds both ends of the cable #1457
Labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
jaylouisw/packetforge#1457
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Two switches joined by a trunk. Somebody trims it to the VLAN it is meant to carry:
The cable is up,
show interfacesstill names its peer, STP still elects a root — and both ends have gone blind to each other. The only trace is a drop counter on a perfectly healthy port, blaming a VLAN nobody mentioned.Cause
The announcement goes through the egress VLAN filter in
_emitFrame: the LLDPDU is classified into the port's native VLAN (1), the allowed list does not contain it, and the frame is refused.LLDP is a link protocol, not customer traffic. It is sent untagged to the nearest-bridge multicast group (
01:80:c2:00:00:0e) and terminated by the nearest bridge; the allowed-VLAN list decides which switched frames the link carries, not whether its two ends may describe themselves to each other. Trimming a trunk's allowed list is routine, and it should not cost the neighbour table.What should be true
VLAN N not allowed on egress, and never reaches the wire;1 packets output, no drop line;down: peer shut downrather than forgotten, access ports in any VLAN still advertise, and a default trunk is unchanged.test/lldp-across-trunk.test.js, 32 checks. Revertingengine/simulator.jsfails 16. Neighbours green:lldp27,show-lldp23,vlan59,switchport-mismatch26,counters-agreement32,flood-refusal-reason24,flood-port-state19,link-state69,stp-port-state31.