Optimize very common index state transitions (rebalance)

Optimize 2 types of state transitions:
    
    1) If a state transition consists only of moving partitions
    between the passive and active states (or vice-versa), that
    is, no new partitions are added nor existing partitions are
    marked for cleanup, don't restart the updater (at the expense
    of more complex logic to correct snapshots received from a
    currently running updater).
    
    2) If a state transition only adds new partitions to the
    passive state (as ns_server's rebalance frequently calls),
    don't restart the updater, and send the updater the new
    list of passive partitions, so that it has a chance to
    iterate over those partitions' content and index them in
    the same run.
    
    As long as the rebalancer in ns_server doesn't keep changing
    its logic regarding state transitions all the time, this
    provides a significant reduction of the indexing time. However
    for at least some scenarios, there's still significant portion
    of rebalance time not spent on indexing nor index compaction.