Use local shuffle reader in containerized environments

[andygrove]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.
#347 adds a local shuffle reader. It uses the shuffle read URL to determine if a fetch is local or not. This works well when running as standalone processes but does not work so well in a containerized environment such as k8s where multiple executors could be running on the same host but in different containers.

Describe the solution you'd like
Why don't we just look on disk to see if the shuffle files exist locally

Describe alternatives you've considered
None

Additional context
See discussion at #347 (comment)

[andygrove]

@Ted-Jiang What do you think of this idea?

[Ted-Jiang]

Sounds like a better approach! 👍 I will make another PR.