Sprint 39 retrospective
- What went well
- Focus
- Stability – more confidence in the stability of the release because of testing
- Velocity – more accurate estimation, we care, and we got things done
- Working together on a branch
- Testing mentality
- Declaring "this is good enough" early enough to review
- What didn't go well
- We didn't release
- Lack of a CI environment (disconnect between environments)
- Things are still in review
- Noise: ESIP, HackIllinois, DataDrivenAg, NASA
- What to continue
- Point estimation
- Staying focused (keep integrity of sprint)
- Testing mentality
Current conditions:
- ESIP
- Christine has approved ESIP allocation at SDSC
- Kevin has cycles (limited cycles)
- DataDrivenAg
- Commercial AWS/GCE support (roadmap item)
- Tasks:
- Cloud9all - terra
- ML-environment (tensorflow)
- JupyterLab
- Rodeo?
- Provision at SDSC
- Monitoring
- DNS/TLS problem
- Account import
- Data subset definition and transfer
- HackIllinois
- Nothing in sprint
- GlobusWorld
- Abstract in sprint
- Personal endpoint support – after Globus auth
- PEARC panel proposal due 2/19
- Globus auth
- EUDAT/FRDR
- 2017-10-19 Roadmap
- Kubernetes 1.8, Travis, CI, Unit testing = stability and maintainability of the thing we're creating
- Documentation
- Si2 proposal
GCE Support
- 1.6.9?
- Can use gcePersitentDisk
- Don't know if the client nodes work
- Risk: still too high, but could move the ball forward