PBR and next-hop router reachability

When a policy with PBR actions is applied to an interface, all next-hop and default-nexthop entries in the associated action list are continuously monitored for reachability. Probing occurs every 5 seconds for each such entry and guarantees that the reaction to the loss of reachability, or detection of the new reachability of any such entry will be approximately 5 seconds.

For example, if there are two reachable next-hops in an applied action list and the active next-hop (highest priority by lowest sequence number) entry becomes unreachable (loss of link or next-hop power event for example), the switch over to the lower priority next-hop will occur within approximately 5 seconds. If there is only one nexthop/default-nexthop entry and no entries of any other type in the action list, routing decisions will return to the system routing table.

Conversely if an applied action list with a single unreachable next-hop becomes reachable, the switch back to routing to that next-hop will occur when reachability of the next-hop on the network is achieved. On an action list with multiple next-hops (where the active next-hop is not the highest priority entry); once the next-hop with the higher priority becomes available, the now reachable next-hop will promote to active once new reachability on the network is achieved.