I am trying to make a map reduce with sharded option true.
The output collection does not exist before the mapreduce.

I have 8 shards, however when I run the map reduce:
with the output options:
out:

{ "reduce" : "testCollection" , "sharded" : true , "nonAtomic" : false}

,

It is writing just to one shard, instead of balancing through the 8 shards.

Totals
data : 1.03GiB docs : 2768955 chunks : 16
Shard shard0000 contains 0% data, 0% docs in cluster, avg obj size on shard : NaNGiB
Shard shard0001 contains 0% data, 0% docs in cluster, avg obj size on shard : NaNGiB
Shard shard0002 contains 0% data, 0% docs in cluster, avg obj size on shard : NaNGiB
Shard shard0003 contains 0% data, 0% docs in cluster, avg obj size on shard : NaNGiB
Shard shard0004 contains 0% data, 0% docs in cluster, avg obj size on shard : NaNGiB
Shard shard0005 contains 0% data, 0% docs in cluster, avg obj size on shard : NaNGiB
Shard shard0006 contains 100% data, 100% docs in cluster, avg obj size on shard : 400B
Shard shard0007 contains 0% data, 0% docs in cluster, avg obj size on shard : NaNGiB