CSPF uses a locally-maintained traffic-engineering database (TEDB) to calculate the shortest path to the destination address. Then, the signaling protocol applies for and reserves resources for the path. In the case of a link on a network is faulty, if the routing protocol fails to notify CSPF of updating the TEDB in time, this may cause the path calculated by CSPF to contain the faulty link.
As a result, the control packets, such as RSVP Path messages, of a signaling protocol are discarded on the faulty link. Then, the signaling protocol returns an error message to the upstream node. Receiving the link error message on the upstream node triggers CSPF to recalculate a path. The path recalculated by CSPF and returned to the signaling protocol still contains the faulty link because the TEDB is not updated. The control packets of the signaling protocol are still discarded and the signaling protocol returns an error message to trigger CSPF to recalculate a path. The procedure repeats until the TEDB is updated.
To avoid the preceding situation, when the signaling protocol returns an error message to notify CSPF of a link failure, CSPF sets the status of the faulty link to INACTIVE and enables a failed link timer. Then, CSPF does not use the faulty link in path calculation until CSPF receives a TEDB update event or the failed link timer expires.
Before the failed link timer expires, if a TEDB update event is received, CSPF deletes the failed link timer.
Perform the following configurations on the ingress node of an MPLS TE tunnel.
The system view is displayed.
The MPLS view is displayed.
The failed link timer is configured.
By default, the failed link timer is set to 10 seconds.
The failed link timer is a local configuration. If the failed link timers of nodes are set to different values, a failed link that is in ACTIVE state on one node may be in INACTIVE state on other nodes.