-
Type: Improvement
-
Resolution: Duplicate
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: Sharding
-
None
-
Sharding
Currently this value is taken from the server mapped memory number and results in no chunks being moved to the shard even if there is room in existing files (from deleted extents or docs).
Taking the sum of each databases datasize would be a better I believe.
- duplicates
-
SERVER-35431 rollback does not correct sizeStorer data sizes
- Backlog