Symptom
OSPF process exception and restart
Conditions
This is only applicable when OSPF is enabled for Segment Routing.
The condition is triggered by a rare timing condition when one of the router's stale (i.e. previously purged) Extended Prefix or Link LSA is received by the router. Such a scenario may happen with high scale and route/topology churn.
E.g. this was seen during test on an OSPF ABR, during OSPF session flap and route churn
There were ~20K (~20K prefix SIDs) nodes and ~40K OSPF routes in the topology.
Workaround
No Workaround. The OSPF process restarts and recovers from the error.
Further Problem Description
This issue is applicable to all IOS-XR platforms from Release 6.0.0 onwards.
This is a process exception and can be traffic impacting.
There is no workaround for this problem.
This is a timing condition issue, when a router receives its pre-restart Extended Prefix LSAs and the router does not have a copy of it.
The system will recover itself after restart, but the restart may lead to the same timing condition again, and lead to a repeated process exception.
