This task encompasses analysis and design discussions needed to define requirements (i.e. user stories) for the multi-node cluster configuration. Output from this task will be a Confluence page defining requirements and a set of JIRA tickets/user stories reflecting the development tasks required to implement the multi-node Kubernetes cluster.

A Confluence page has been started, documenting a simple prototype cluster based on an existing ansible playbook:
https://opensource.ncsa.illinois.edu/confluence/display/NDS/Multinode+cluster

Questions we might consider:

• Can we use the kubernetes/contrib/ansible playbook as a basis for our own deployment?
• Do we need all of the features/any other features not enabled by this playbook/
• Features included in the contrib/ansible include: flannel networking, logging, monitoring, Kubernetes dashboard, DNS, etc.
• Do we expect to support others bringing up NDSL clusters using this deploy strategy?