Details
-
Improvement
-
Resolution: Cannot Reproduce
-
Major - P3
-
None
-
None
-
None
Description
When inserting the data attached on a three shard cluster (three separate machines, all of them running v1.3.4) 99% of the documents go to the same shard and chunk ranges don't make any sense as seen on printShardingStatus( db.getSisterDB( "config" ) );
The key used for sharding was "mts".