deadlocks and lock wait timeouts

This patch adds the option to process the update query in a batch format where a limited number of rows are processed per update hook. This is important because the query requires a full table scan which will ultimately lock every record in the table and the update query can take an appreciable amount of time on a large table. Processing in batches, combined with the new index, requires a lock on only a few records at a time. On busy sites, the time taken while locking the table can result in deadlocks or lock wait timeouts. By default, this query will behave as it did before, unless radioactivity_decay_batch_size is set. Additionally, I have wrapped the entire function in a lock_acquire to ensure this function is not called multiple times concurrently as that is a sure-fire way to cause deadlocks.