when guessing node is restarted, it does not begin guessing again until next block.

Actually, I'm not certain. It might start guessing if a new block proposal is received for the current block.

But really, it should start guessing soon after the node is restarted and connected. That could mean:

1. asking peers for block-proposal(s)
2. storing best block-proposal locally, in case of restart.

For now, this is mostly a nuisance and seems low priority. But it should be fixed at some point.

The basic behavior I am seeing is:

1. node is guessing on block 400
2. node is shutdown, and restarted.
3. block 401 has not been found yet.
4. node remains inactive for a lengthy period, when it should start mining instead.