Symptom
LC: VSM-500
AIB gets to OOR state, following trace is observed,
aib/server/fast/aib_server 0/3/CPU0 1884818# t1 AIB server runs into RED OOR state with 512000 adj entries,blocking update until back to GREEN
And AIB is not coming out of OOR. Symptoms of that are -
quarantine list doesn't emptied after the quarantine timer expires.
i.e. either
show aib quarantine-list location
or
show aib quarantine-list remote location
Have entries, that don't get deleted.
Another thing to note is that in the output of the quarantine-list the version listed is greater than UINT32_MAX (4294967295).
AIB will not send updates or create new adjancencies when it is in this state.
Conditions
Very fast churn on the router such that the adjacency version number can increase to be greater than the UINT32_MAX number.
Workaround
No workaround. Router/LC needs to be reloaded to get out of this state. Process restart of AIB or FIB will not solve the problem.
Further Problem Description
This problem can happen anytime from 5.2.2 release onwards, since the adjacency version was increased to a 64bit number, but other places were not modified.
This is a PI problem, so it can impact all platforms.
However, it is being seen more on VSM LCs due to the EINT interfaces not resolving, and continually retrying. As a result, the adjacency version grows to be > UINT32_MAX in a few months.