No to what exactly? That's replication, not partitioning, and involves scaling up every instance to be big enough to hold the entire dataset. You must scale up

No to what exactly? That's replication, not partitioning, and involves scaling up every instance to be big enough to hold the entire dataset. You must scale up first before you can scale out.