FYI: TWS call next week will include blueprint review
Moved from Juniper Jira to TF Jira (woo!)
Infra group is changing their meeting frequency
No need to meet weekly at the moment
Move to biweekly
Monthly releases: Contrail vs TF
Only introducing the subject now; longer conversation needed
Juniper is releasing Contrail every month now; TF…is not
Can the TF CI even match what Juniper has right now?
Hardware requirements
Manual QA required
If TF keeps consuming/repackaging Juniper, setting expectation of monthly releases but those releases are tied to Juniper release capabilities that TF can't duplicate
We need to figure out the best path forward
VMB: Make this our single goal for 2020: reach parity w/Juniper so TF is consumable
Draft plan
Start figuring out the gaps between where we are & where we want to be
Where are we now?
Where to we want to be?
What do we need to work on to get there?
Use DDF to work on filling in those gaps and/or plan for filling them in 2020
swill: Would it make sense to use the CNCF cross-cloud tests?
VMB: maybe? Dunno what we need, dunno what they do. But if there's overlap then sure.
Abhijeet: ATT has Tempest tests
RLB: Want to use these, but current test framework is Heat/python & totally different; wanna get to that
AS: Useful to open source these tests anyway?
RLB: Yes, definitely
AS: Point of contact for testing?
ATT has build Torpedo, a new project and container testing
Could have the testing SME & have them fill folks in?
RLB: reach out to RLB, VMB, Alex
Next step: RLB talk to alevine & make the gap analysis a priority
VMB: Documented in the wiki, please
swill: Published list of passing/failing tests?
RLB: alevine has it; Zuul
swill: k8s architectures
(note taker got pulled onto something else & missed a bunch of this)
tl;dr: Will's gonna make this into a blueprint.
Action items
Casey Cain Start a wiki page for planning the KubeCon DDF (topic proposals, etc)
Randy Bias Ask Progmatic to link to the Zuul on the wiki
Casey Cain Will get the blueprint review on the agenda for next week's TWS call
Randy Bias To talk to Alex about a documented GAP analysis of what is missing to achieve parity with Juniper testing by Kubecon NA and the TF DDF there