nodeos should not sync blocks while running a snapshot and in catch-up mode

[matthewdarwin]

trafficstars

Nodeos has 2 methods of doing p2p transactions:

• normal mode (within the last 5 minutes of head block)
• catch-up mode (older)

If nodeos is in "catch-up mode" and it is asked to create a snapshot via the producer plugin, nodeos actually keeps syncing blocks. When blocks are big and snapshots take a long time (eg on the wax network), this blows up the amount of RAM needed by nodeos significantly.

When in catch-up mode, nodeos should stop syncing blocks from the network and just run the snapshot. Block sync should resume after snapshot is complete.

Memory usage is flat while snapshot is not running. Memory usage increases while snapshot is running.