[OpenStack-DefCore] Meeting 5 Notes from 3/3
Rob_Hirschfeld at Dell.com
Rob_Hirschfeld at Dell.com
Tue Mar 4 17:47:12 UTC 2014
Recording link included.
# Summary
* We are making steady progress, some places are slightly behind but we are on schedule with a lot of key items
* Critieria Process for test selection bearing fruit - ready to expose to larger audiencec
* Demo for RefStack / TCUP looks ok. Will set date for rollout to the board & TC for alpha testing
* Oh, we're in Stackforge, check out that shizzle
* Planning some interactive prep meetings with TC around designate sections. We asked them to resconsider their 100% of Nova & Swift only stance.
* Driver Testing - this may be more alarming than it first appears. OpenStack board needs to be involved if "certification" used in conjunction with the Brand
MEETING AGENDA (3/3 @ 3 PDT, 120 minutes) Elephant.5
#######################################################
NB: No time is perfect, but if you want to encourage presence of PTLs / TC members, please avoid 12:00pm-2pm PDT on Tuesdays, as that is when we have the TC and PTL weekly meetings.
Recording: http://youtu.be/eGiRrpDGzGU
* was FACE TO FACE MEETING at Piston HQ
*
Previous Meeting: https://etherpad.openstack.org/p/DefCoreElephant.4
Next Meeting: https://etherpad.openstack.org/p/DefCoreElephant.6
# Agenda:
1. Status Update & Calendar Review
2. Update on Criteria Scoring Process (not a "working" meeting)
3. Board Demo of TCUP
1. Answer questions from TC's "Designated Code" proposal
1. https://review.openstack.org/#/c/74483/2/resolutions/20140211-tc_defcore_response (now approved by the TC)
5. Mirantis Driver Testing & DefCore Alignment
* Receommendation for position
Not discussed: program vs project
Roll Call:
* Rob Hirschfeld
* Joshua McKenty
* Troy Toman
* Michael Still
* Rochelle Grober
* Mark Atwood
* Catherine Diep
* Raymond Wong
* Alan Clark
* Mark Radcliffe
* Van Lindberg
# Minutes
## Status Update & Calandar 3:07pm
We are running behind our aggressive schedule, but hope is not lost
DECISION POINT: We will score based on capability groups not tests. Tests will inherit must-pass status from their capabilty groups.
1. Requirements and ongoing development of RefStack & TCUP (slow, but we have a thinner set that is workable)
1. Development of capabilities list (on track)
1. Development of scoring criteria (on track)
1. Application of criteria to capabilities list (ready for broader community input)
1. Review of criteria effectiveness (on track)
1. Review of weighting (as planned, not started yet)
1. Review of threshold for "coreness" (as planned, not started yet)
1. Additional findings and committee outputs
* - Recommendation of process on deprecation and documentation
* - Additional recommendations on nomenclature
DECISION POINT: At 3/4 board meeting, we are asking to take the capabilities & criteria to a larger audience.
## Update on Criteria Process (fuzzed into, 3:30pm)
Link to current scoring sheet: https://docs.google.com/spreadsheet/ccc?key=0Av62KoL8f9kAdFo4V1ZLUFM0OHlrRnFpQUkxSHJ5QWc&usp=drive_web#gid=6
HELP NEEDED (added to E.6 agenda): From TC to help figure out deprecation
## Board Demo of RefStack & TCUP (3:46pm)
We can show TCUP running against multiple end-points. Upload is not working yet
Refstack is now on stackforge, but needs to be part of an offical program to award ATC status (https://wiki.openstack.org/wiki/Governance/Foundation/TechnicalCommittee)
BONUS: RefStack is managed in Stackforge
ACTION: Board to be asked to run this on their own clouds
## Designated Code Section
The TC have been actively discussing this and formulated a response: https://review.openstack.org/#/c/74483/2/resolutions/20140211-tc_defcore_response
TC is leaning towards saying 100% of the code is designated.
* confusion on reading - is this 100% of everything or 100% of Nova & Swift only? Reading is 100% of Nova & Swift.
ACTION: Draft a response based on TC proposal accepting 100% of Nova and Swift as Designated. All other code is NOT designated. Implication is that implementations must ship those projects AS IS and all other additions or omissions are acceptable AS LONG AS MUST-PASS TESTS pass. Rob and Josh to draft a response with implications of this definition.
GUIDANCE: We are NOT creating a requirement exact specification, we are asking for guideance that can drive fair-use
Michael suggested a join DefCore / TC meeting w/ a planned agenda.
* DefCore is receptive to an interactive meeting (not physical, but voice)
* Rob offered to help coordinate the agenda.
* What about using the pre-summit joint session for this?
DRAFT OF DESIGNATED SECTIONS: https://etherpad.openstack.org/p/openstack-designated-sections
NOTE: We do NOT need all projects to designate, only the ones that are likely to have capabilities flagged by must-pass tests
## Driver Testing 4:51pm
There's a community need to have validation of drivers, hypervisors, etc
Issues
* seems like a lot of duplication for people writing ways to show Tempest results (this was not a primary focus, but it seems like we could collaborate more)
If the efforts are upstreaming their tests and helping add test coverage, then it's good
Concern > are these efforts aligned & support or actually parallel without collaborating.
Concern > do we need to have DefCore coordinate this?
QUESTION TO RESOLVE: Does lack of control of these efforts by DefCore weaken the Brand?
ACTION: Using the OpenStack mark for "certification" requires DefCore & Board approval. This should be brought before the board as an action item. We will ask the Foundation to police this.
## We re-opened designed sections 5:15pm
# Elephant.6 Agenda
* How to crowdsource data, we need a plan
* Need help from TC on deprecation & documentation in capabilities. What about very long deprecation cycles, like major API version bumps?
* Subcomittee update from Bylaws subcommittee
* Becoming an official OpenStack program
* Discuss Designated Sections: https://etherpad.openstack.org/p/openstack-designated-sections
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/defcore-committee/attachments/20140304/a5b2852c/attachment-0001.html>
More information about the Defcore-committee
mailing list