Scenario

As a customer experiencing an emergency,

I want to be able to optionally apply an overlay on top of a cluster as a hotfix,

so that I can have granular control of custom configurations to remediate.

Summary

Based on the conversation Aug. 26, the MCO's role in MCBS will be an "unopened envelop". We should move away from thinking in terms of files and towards a layered container.

RHCOS Phase 1:

Creating some tooling that allows us to package some rpms into a container layer.

Therefore, MCO's Phase 1 will try to locally:

1. build a container by any means possible
2. put that container somewhere
3. pull the container
4. apply on top of rpm os-tree

Outcome:

• Updated use cases doc that everyone contributes to as we find new use cases.
• Share the MCO team's learnings in a doc throughout a week long hackweek
• Spikes and stories for next steps.

Things to figure out:

• How to apply config?
• What are our unknowns?
• How to have separation of concerns?
• What are the downsides? (performance? security? usability? complexity of use cases? etc.)

Out of scope:

• Shippable code
• Determine a build strategy.