I’ve turned off the balancer and deleted my one index and it seems to have sped things up ( down from 111 minutes per batch to ~ 7 minutes ).
So far I’m pretty unimpressed by the performance of MongoDB vis-a-vis SQL Server. SQL Server loaded this dataset, WITH indexes WITH concurrent data processing steps thrown in, in less than a week.
With all the hooplah around MongoDB, it makes me think I’m missing something, some kind of misconfiguration or hidden bottlenexk / overhead somewhere, that is keeping me from the promised land.