scrub should pause when there is some dirty data
Scrub/resilver should pause if there is enough dirty data. Otherwise, we could
scrub for the maximum of zfs_txg_timeout seconds, despite a high % dirty data
and subsequent throttling of incoming write requests.
We should pause when dirty data is >
zfs_vdev_async_write_active_min_dirty_percent. i.e. once we start to increase
the number of concurrent async writes.