Comment From: itamarhaber

Hello @huangxiaoxie

Keep in mind that this issue tracker should be used for reporting bugs or proposing improvements to the Redis server. Questions should be directed to the community:

Comment From: itamarhaber

As far as I can tell this was done because of two main reasons:

  1. The concept of shared/numbered/logical Redis databases was deemed long ago as an anti-pattern in Redis. The general thinking was to deprecate it, so adding it to the cluster (v3) wasn't one of the goals.
  2. That said, supporting databases in the cluster would've (probably) made the implementation more complex with all the associated downsides of that.

Closing as semi-answered.