[1.0 -> main] Normally process blocks from the forkdb on startup

[heifner]

In Spring 1.0.0, unlike Leap, we process blocks into the fork database immediately. This can cause the fork database to grow very large when syncing and shutdown due to hitting the new max-reversible-blocks limit (before #545). When a node is shutdown, in Leap, it was assumed that when restarting a node if that node did not receive any blocks from the network it would be at the same height as when it was shutdown. The existing tests/nodeos_read_terminate_at_block_test.py integration test verifies this behavior.

However, if the node shutdown because of max-reversible-blocks you would like on restart for the node to process blocks out of the fork database potentially allowing LIB to progress and reversible blocks to be consumed and shrink in size. This is at odds with the expected behavior of a node starting up after a terminate-at-block. A user might find it very odd to terminate a node at a block, but find on restart that the node is actually beyond that block.

To fix the issue reported in #565, we would like to process blocks out of the fork database on restart. However, we want to maintain the existing expectations of starting a node after a shutdown via terminate-at-block being at the same block height on restart.

Therefore, this PR modifies nodeos to normally process blocks out of the fork database on startup unless the node has no configured peers. The idea being that if a user has terminated a node with terminate-at-block and wants the node to remain at that block, they will restart the node without any p2p-peer-address configured. This is a bit of a hack until #570 can be added. Since #570 is a new feature, it will not come until a future release and will not be part of Spring 1.0.0.

Merges release/1.0 into main including #572

Resolves #565

[ericpassmore]

Note:start
group: STABILITY
category: INTERNALS
summary: Until pause at block height is available this fix enables use of forkdb to advance LIB on startup when connected to peers, but does not read from forkdb on startup when no peers are connected.
Note:end