When splitting, the last bucket can end up much larger than the rest of the buckets

Reported to me this morning--Splitting <40TB of data into 5 buckets, buckets 1-4 have 5TB in them, bucket 5 has 18TB.

I've seen this before long ago and worked around it by increasing the bucket count. We will try the same workaround.