[babel] Restarting nodes and seqno requests

[Toke Høiland-Jørgensen <[e-mail]>]

I was debugging an issue with routes staying unreachable for a long time
after a node reboots; and decided I wasn't actually sure what the right
behaviour is, so I thought I'd ask.

So, the simple scenario is this:

- Babel nodes A and B have established a neighbour relationship, and node
  B announces prefix P with seqno S.

- Node B restarts (i.e. shuts down, loses its transient state such as
  seqnos and comes back up either immediately or after a relatively
  short time). It will then start announcing prefix P again, but now
  with seqno S' < S.

- Node A sees the new announcements from node B with a lower seqno.
  Since the hold time for P has not expired, these updates are
  unfeasible, so node A sends a seqno request.

- Node B updates its seqno to S'' = S' + 1 and resends the update.

- In many cases, S'' < S, so the update is still unfeasible from A's
  point of view.

The question is, how is this supposed to be resolved? Should A keep
resending the seqno requests each time it gets a new unfeasible update -
and if so, is there any limit to the frequency? Or should A immediately
consider the update with seqno S' as feasible because its selected route
has metric infinity?

I *think* the answer is the former (and that no rate limit is needed),
but I may be missing a reason why it is safe to accept the unfeasible
update during hold time... So what does everyone else think? :)

-Toke