2018-08-15 Meeting notes

Multi-CPU images

• Build pipeline

  • LF currently setting up servers and jobs

  • JDK issue being addresses

• OpenShift has an integrated CI/CD pipeline that could use arm hardware

  • Lab will initially be internal

  • Developing Ansible scripts

• Lab access

  • UNH

  • https://github.com/worksonarm/cluster

  • Auto Arm Pod at IOL-UNH

Best practices

• Addressing immediate needs and pain points

• Initial version of best practices on the wiki

• Mid-term and long term approaches

• Consider expanding scope to provide input on startup for different projects

  • OOM and CIA share a similar concern on project and startup and configuration

  • Build guides will be helpful

• Should we define strict rules?

  • We should be changing the nature of how everybody builds images

  • Perhaps we should ask for TSC approval

• We should also look at the implementers' pain points as well

  • Tal is working on a PoC

Collaboration with SO, SDC, VNFSDK, OOF, Logging

• Collaboration with OOM

  • Beijing is 5 1/2 times bigger

  • It takes a long time to build

  • It is error prone

  • Sizing resource needs (S,M,L)

Focus on a single project: SO

• Turn it into a cloud native app

• Measure/assess current state

  • Container is huge b/c of base image

• Introduce changes

• Observe impact of contributions