17 Jul 2006 14:54
New trap
ravikumarb <ravikumarb <at> infosys.com>
2006-07-17 12:54:37 GMT
2006-07-17 12:54:37 GMT
Hi,
We would like to porpose a new trap for bridges so that they can detect malfunctioning neighbour bridges.
Lets call it as "neighbourUnreachable". This trap will be generated by the bridge if it doesn't receive 5
consecutive Hello BPDUs from a port, which was receiving these messages earlier. This trap will be useful
to detect cases where the bridge and its ports were functioning properly for some time and suddenly
started malfunctioning due to a hardware glitch or a software bug. Note that link failure (down) trap
won't be raised as the link is UP from the PHY point of view. This is analogous to Hello messages used by
Routing Protocols to detect routing peers. This trap will be very useful for the NMS to take appropriate
actions. Note that this is very different from "topologyChange" trap, which just tells that the port has
transitioned to Blocking State from Forwarding State. This trap tells that the bridge is not functioning
properly at all from the port's point of view. It has a specific meaning and is detected fast. Following
would be its structure:
neighbourUnreachable NOTIFICATION-TYPE
OBJECTS { dot1dBaseBridgeAddress dot1dBasePort dot1dBasePortIfIndex }
STATUS current
DESCRIPTION
"A neighbourUnreachable trap is sent by a bridge when it
doesn't receive 5 consecutive Hello BPDUs. This will
allow a bridge to detect malfunctions of the neighbour
bridge even when the link state to the neighbour bridge
is UP. It is analogous to Hello messages used by Routing
Protocols to detect whether the neighbour router is
functional."
::= { dot1dNotifications 3 }
Please let us know your thoughts regarding the proposed trap.
(Continue reading)
RSS Feed