Attendees:
No need to add - will be done after the call.
Agenda and Minutes:
- Antitrust notices
- Walk-in Items
- WSL/co-leads: Tom Kivlinto step down at next nomination period (February 2021 would be 12 months after RI2 formally started)
- Nomination / selection process for new WSL will follow documented governance process: https://github.com/cntt-n/CNTT/blob/master/doc/gov/chapters/chapter04.md
- Project board review: https://github.com/cntt-n/CNTT/projects/27
- RI2
- How do we progress with the document?
- Can we start adding content based on RA2 directly
- Do we need to implement in Kuberef first, and then use that as "reference material" for RI2?
- Currenly only covers HW provisioning and Kubernetes deployment
- How do we progress with the document?
Kuberef
Roadmap/Requirements Planning Discussion
- What features do we want in the next Kuberef release?
- Upgrade to CentOS 8
- Not high priority as BMRA supports CentOS 7. CentOS 8 might be EOL 2021
- Should keep track of CentOS distribution status
- Run RC2 test suite against BMRA v2.0
- Pankaj/Trevor - Current RC2 testcases doesn't take into account all RA2 specs, only k8s ones
- More alignment needs to be done in this area
- Feedback loop is missing
- What should a test case cover? Should it take into account specific implementations or should it be generic?
- Ability to run on virtualized infrastructure (VMs)
- Could make it easier for contributors to help with project
- Might have some feature limitations (CPU, Network, Memory, PCIe)
- Tinkerbell as a hardware provisioning tool
- Upgrade to CentOS 8
- What are the current gaps?
- Some ideas listed below (Michael Pedersen)
- What features do we want in the next Kuberef release?
- Alignment with RA2 and RC2 - Should be high priority
- It is expected that running RC2 towards Kuberef/RI2 will result in feedback for RC2
- Coverage of some tests might not be exhaustive (e.g. security)
- Given the dynamic nature of Kubernetes, there will be several ways to handle many of the requirements
- Tests will have to go beyond Kubernetes functionality and APIs
- E.g. networking. There are several ways to add interfaces to a pod/workload. Is it sufficient that the interface is available and has been assigned an IP?
- If not, then each different method for implementing network functionality likely needs it's own set of tests (created/maintained by the developers?)
- It is expected that running RC2 towards Kuberef/RI2 will result in feedback for RC2
- Presentation for Kubecon EU 2021 - DONE
- May 4-7 (Virtual)
- CFP closes Dec 13, 2020 (end of week)
- Jira review - https://jira.opnfv.org/secure/RapidBoard.jspa?rapidView=240&projectKey=KUB&selectedIssue=KUB-10
- Michael created new issues in Jira
- Gerrit review - https://gerrit.opnfv.org/gerrit/q/project:kuberef+
- AOB
Ideas for kuberef short-term plan:
- Integrate BMRA (v2.0) in the Kuberef framework
- Expected done end of 2020
- Configure CI to run most recent RC2 test suite
- Decide what should be run when
- Estimate TBD
- Start enabling and testing additional cluster/k8s features
- Enabling started as part of BMRA integration
- Expected done in Jan 2021
- Add example CNFs/PODs/Workloads
- Is this something we want to include in Kuberef or keep separate?
- Might be (partially) covered by deployments/installs done through RC2
- Simple examples could be used to verify K8s features (maybe even through CI)
- Estimate TBD