Looking at one of the CouchDB 2.0 lead-up blog posts (here), there is mention of shards (q), copies (n) and nodes.
Reading up on Couch 2.0 suggests (to me at least) that a node is basically an instance of Couch, which can either be a shard or a shard-copy. Is that correct? i.e. if you have 8 shards (q=8) and 3 copies of each shard (n=3), then you have 24 nodes?
How should these nodes be split across servers and will it do this automatically? If you have 8 shards should you have a cluster of 8 servers?
