In the node.js SDK, if you specify durability requirements which exceed available limits, the requirements are automatically rounded to the “maximum available”: https://github.com/couchbase/couchnode/blob/master/src/operations.cc#L373
Feel free to file an issue if you think this is a bug. It would seem that in most cases, the intent behind durability operations is to provide increased durability (to the extent of “as much as possible”) rather than a fixed form of durability to a specific number of nodes. Additionally, if a server is removed from the cluster and the durability requirements remain fixed in code, operations would start needlessly failing (unless of course, the application has some certain strict requirements to only function if an item is sufficiently durable to n number of nodes, rather than “as many as possible”).