Add rack concept like Hadoop

[FrommyMind]

Feature request

Is your feature request related to a problem? Please describe.

No Describe the solution you'd like

Each Fe and BE has it's own rack properties.

Describe alternatives you've considered

As we all know, distribution system, like hadoop, has a rack concept. Replicas of one data will spill into different server in different rack to prevent data lose when one entire rack lose from cluster. Additional context

[jzhou77]

Listing the ID is easy, IP:port is a bit tricky and may be stale. That said, even the storage server ID is very useful for debugging purposes.

@xumengpanda assign to you to prioritize.