I came across this excerpt from the Cisco documentation:
"Not all routes can have repair paths. Multipath primary routes might have repair paths for all, some, or no primary paths, depending on network topology, the connectivity of the computing router, and the attributes required of repair paths."
My question is, what actually determines if a path is eligible to become a repair path? I'm unable to find any clear explanation of how OSPF chooses these.
Background: I have OSPF setup on some L3 switches that are in a physical ring topology terminated on the ends by routers (more of a crescent than a fully closed ring). Each switch is receiving routes originating from the routers via both its east and west neighbor. Some of the switches have repair paths for these routes going the other direction, and some do not. The ones that do, however, are in the "middle" of the ring. Based on that, I assumed it has to do with metric, but after checking the OSPF external database the metric of the routes are the same in all the switches (I redistributed these routes into OSPF from a separate process using the max metric of 65535). Can someone explain?
The setting that I'm using is 'fast-reroute per-prefix enable prefix-priority low'