Handle stale transactions on the backlog table

[r-marques]

We need to handle the case in which a transactions in the backlog is assigned to nodeA and nodeA becomes unavailable (network connection, the node went down, couldn't handle the awesomeness of BigchainDB and caught on fire)

The solution would be a process that would check the backlog periodically for stale transactions and reassign them.

[ttmc]

For reference, this is the routine reassignOldTransactions() in Listing 3 in the whitepaper (around page 23, although the page number may change as we edit the whitepaper).