#OperationSaveHarddrives

At some point we'd like full nodes to be able to prune old blocks to free up hard drive space. Of course doing so would mean new nodes can't sync the chain from before the last pruning point, and users won't be able to query for txs from before this time. The former is solved by state sync (ie. #828), and the later by the good people who operate archive nodes.

Aside from actually implementing state sync, we need to expose some way to responsibly prune the history. While this should work together with state sync, one could imagine other ways to sync state out-of-band, so that this feature could be used even independent of an in-protocol state-sync reactor.

Related #2543