fix: prevent race condition between block propagation and sync

[sdbondi]

Description

• prevents base node from switching to sync mode before the block has a chance to arrive via propagation

Motivation and Context

Base node may switch to sync mode many times as the block arrives at peers, syncing a single block and returning to
listening state. It is more efficient to stay in listening state for a while to give the block chance to arrive via propagation.

How Has This Been Tested?

Manually tested:

• more than 1 behind tip - sync as normal
• startup 1 behind missed propagation, enters sync mode
• 1 behind, prevented from entering sync mode, block arrives shortly after

Prevented 3 times in a few minutes on my base node.

[sdbondi]

I've just noticed the blocks_behind_before_considered_lagging setting - think we should default this to 1 because this causes unnecessary network overhead

[stringhandler]

I've just noticed the blocks_behind_before_considered_lagging setting - think we should default this to 1 because this causes unnecessary network overhead

This setting is only intended to be used in debugging and tests. It should be set to 0 by default

[sdbondi]

There is a race condition as described in the PR which puts the base node in header sync before the block arrives via propagation.

[stringhandler]

cool

[sdbondi]

Yeah so maybe that setting has a slightly different use case. than this which would never need to wait > 1 - ok cool makes sense