Now that mongos has a better implementation of cursors, it should be possible to implement this for both sharded and unsharded collections. For sharded collections an initial implementation could just use a normal find() query to each shard if we don't want to add a ShardFilterStage in parallelCollectionScan.
- is depended on by
-
TOOLS-1468 Allow mongodump to leverage parallelCollectionScan to speed up dumping of large collections
- Closed