User Story

As a data scientist, data engineer, or machine learning analyst, I want to be able to access an HDFS cluster that uses Kerberos from a Kubernetes cluster so that I can easily access and analyse large datasets stored in HDFS without having to manually configure authentication and authorisation.

Tasks

• Sprint 9 https://phabricator.wikimedia.org/T327257

Acceptance Criteria

• The person should be able to access and load data stored in HDFS from a K8 Pod.
• The user should be able to do this securely and with minimal configuration.

Outstanding Questions:

• Can a user use their existing credentials or will it be a shared key?
• Is there a sufficiently robust audit trail?