SR OS can determine the forwarding state of an LSP from the LSP BFD session, allowing users of the LSP to determine whether their transport is operational. If BFD is down on an LSP path, then the path is considered to be BFD degraded by the system.
Using the failure-action command, a user can configure the action taken by the system if BFD fails for an RSVP LSP or LDP prefix list. There are three possible failure actions:
failure-action down — the LSP is marked as unusable in TTM when BFD on the LSP goes down. This is applicable to RSVP and LDP LSPs.
failure-action failover — when LSP BFD goes down on the currently active path, then the LSP switches from the primary path to the secondary path, or from the currently active secondary path to the next-best preference secondary path. This is applicable to RSVP LSPs.
failure-action failover-or-down — similar to failure-action failover , when LSP BFD goes down on the currently active path, then the LSP switches from the primary path to the secondary path, or from the currently active secondary path to the next best preference secondary path. However, failure-action failover-or-down also supports the ability to run BFD sessions simultaneously on the primary and up to two other secondary or standby paths. The system does not switch to a standby path for which the BFD session is down. If all BFD sessions for the LSP are down, then the LSP is marked as unusable in TTM. This is applicable to RSVP LSPs and SR-TE LSPs. See the 7750 SR and 7950 XRS Segment Routing and PCE User Guide for further details of its use for SR-TE LSPs.
In all cases, an SNMP trap is raised indicating that BFD has gone down on the LSP.
It is recommended that BFD control packet timers are configured to a value that is large enough to allow for transient data path disruptions that may occur when the underlying transport network recovers following a failure.