Could you elaborate on the architectural nuances of sharding and partitioning methodologies within distributed database systems, exploring their implications for data distribution, query routing, and fault tolerance, and how do these techniques intersect with data replication and consistency mechanisms to ensure robustness and resilience?