Just having a discussion with Bela about this.
I guess having "grouping" on GridFS' content would make sense.
e.g. put all chunks on the same node
Is this doable?
Afaiu, we would need to have some sort of "similarity" function for
content's metadata?
-Ales