> Could you provide more details on this? Are you saying accessing the metadata will be a bottleneck (via E.g., a relational database)? It's the problem of hav

> Could you provide more details on this? Are you saying accessing the metadata will be a bottleneck (via E.g., a relational database)? It's the problem of having too many small files. Explained here https://github.com/chrislusf/seaweedfs#introduction https://github.com/chrislusf/seaweedfs#introduction. > Also, what do you mean by fill/rebalance in this context? Assuming you are replicating the data. You have to be careful of cascading failures if repairing replication is too slow because of slow network/disk-io/cpu.