Symptom
When using per-prefix fast-reroute feature with ISIS routing protocol, under specific timing condition in large scale deployments, some or all of the FRR protected routes might temporarily lose the backup path for a very short period of time.
The user visible symptom is that affected routes have the timestamp reset in the RIB when the problem happens.
Conditions
Problem might happen under specific conditions in large scale deployment (hundreds of IS-IS neighbors), when multiple SPF calculations in both L2 and L1 levels happens very close to each other.
Workaround
There is no workaround.
No workaround is usually needed, the problem doesn't cause any traffic loss.
Howerer, if this defect is causing any secondary problems (massive updates ISIS->RIB->FIB->HW, or frequent resets of route uptime in RIB), then the only solution is to install image or SMU with this fix.
Further Problem Description
See also CSCux78200, which fixes another problem with the same symptoms.
