Failure-action failover

The failure-action failover command is supported for point-to-point RSVP LSPs (except mesh point-to-point and one-hop point-to-point auto-LSPs because these do not have a secondary path). When failure action failover is configured, the system triggers a failover from the currently active path to the secondary path, the next-best preference secondary path, or the secondary-standby path of an LSP when an LSP BFD session configured at the LSP level transitions from an up state to a down state. Unlike failure-action failover-or-down, this failure action does not affect how LSP paths are programmed in the data path and only runs LSP BFD on the active path.

The LSP is always marked as usable in the TTM, regardless of the BFD session state and BFD traps that are generated when the BFD state machine transitions. If BFD is enabled and failure-action failover is configured, the following conditions apply.

• It is possible to bring the LSP up regardless the current BFD session state.

• If the BFD session transitions from up to down, the current path immediately switches to the next-best preference standby path.

• If MBB is triggered, then this occurs immediately on the primary path, regardless the BFD session state.

• If the operator is concerned about detecting data path failures that may not be detected by the control plane, Nokia recommends that the revert timer be set to its maximum value.

• LSP BFD only runs on the currently active path. It cannot determine if any non-active paths (for example, a secondary path or primary path during reversion) that the system switches to is up and forwarding. The system relies on the normal control plane mechanisms.

Table 1 describes how the system behaves if a user changes the failure-action while BFD is down. The LSP remains on the current path unless (or until) the control plane takes action or the revert timer expires.