Support OCP workloads on OSP DCN for Enterprise use-cases - tech preview

Feature Overview

As an Enterprise customer, my OpenStack deployment is distributed across multiple sites (called Distributed Compute Nodes, DCN), called Edge Sites and I run OpenShift clusters on these sites.
I run OCP cluster at the remotes OSP sites (using DCN, Distributed Compute Nodes) and can consume OpenStack resources on the same site.

Goal

As of today, OCP workloads aren't supported on DCN and have never been tested by QE, so we want to design a Reference Architecture Role Profile of what can be tested now, in the OSP16.2 timeframe, (therefore supported at some point) and potentially delivered to customers.

And then a next step will be to work on the evolution of this architecture based on the OSP and OCP roadmaps, and find a conjuncture of what customers need and what we'll be able to deliver in the OSP17 timeframe.

Why is this important

With a growing demand from customers to deploy OCP at the Edge of OpenStack clouds (using DCN, Distributed Compute Nodes), we started this first Epic so we can provide be ready when it'll happen. This EPIC is important because today OCP isn't considered as a supported workload in DCN, and we need to figure out our needs in term of OSP features, documenting potential gap, drafting a testing procedure and provide a Reference Architecture Role Profile as a response to our potential customers.

Requirements

• Plan would be Tech Preview in OCP 4.10 and GA in 4.11
• Deliverable:
  • Role Profile of OCP at the Edge for Enterprise, that we can test and support
  • Documentation
  • QE testing
  • Upstream CI testing
    • AZ testing strategy
• OCP workload is supported at the Edge sites and workloads would still be operational if the network link between the Central and Edge sites would break for a moment.
  • This requirement is basically to track all the work around Topology Awareness (Cinder/Nova/Neutron/OVN AZ support), disconnected environments, etc
• Networking
  • Support for IPv6 provisioning
  • IPv4 Provider networks is the target for now.
  • Dual stack (IPv4/IPv6) for VM and pods

Resources
https://docs.google.com/document/d/1aqun-QLkHkik-6uNLGnMDR5C0sNv7pXTIVcxyUVEayA/