Blockstore contains unrooted slots past the latest optimistic slot

[alpeng-jump]

In a cluster halt, Firedancer always has a blockstore with unrooted slots past the latest optimistically confirmed slot, likely due to buffering. After the restart, a node with a locally created snapshot needs to purge these slots before booting off it, which is not ideal, or pass all 3 specified arguments --wait-for-supermajority, --expected-bank-hash, --expected-shred-version, which is not intuitive.