Meeting agenda and notes for the Dec 18th meeting

· Welcome!

· Goals

• Complete the work described in Brent's Patch Resolution proposal for the non-Neutron OpenStack content in StarlingX

• Rebase StarlingX to OpenStack master and release in May with Stein.

• Keep current with OpenStack master on an on-going basis

· Branch handling

• Need feature branches created immediately to start pulling from OSF master .  For projects: stx-integ, others?

§ Hold off on branching for now, likely target mid-January.

• Ensure all dependent services are updated if needed (e.g. rabbit, memcached, others?)

• Build / test of the branch

§ Who will do the builds?  Who will handle the inevitable build issues?  

· We need to merge with containerized services - turn on config flag - no support for non-containers

· Container team to do initial work to get the builds working again

§ How do we handle carried patches?  Ignore?  Rebase?  Selectively apply?  No carried patches to be applied initially.  Run against master and address issues if/when they arise.

• The following critical refactoring needs to be accelerated as it is the bare minimum to move from track 2

§ Neutron host state management moved to the STX-NFV  https://storyboard.openstack.org/#!/story/2003857

§ Nova service create moved to the STX-NFV  https://storyboard.openstack.org/#!/story/2004583

§ Modeling the provider networks in sysinv      Story: https://storyboard.openstack.org/#!/story/2004455

§ Need Neutron Spec implemented to cutover to track 2 (Which spec????)  Bruce to find in chart deck.

· Work item status and discussion

• Intel response

§ Can cover Yongli's 3 items for Stein

§ Need the spec approved for vCPU model - how can we help?  Risk very high for missing Stein.  Jack to update to address latest comments, Yongli to review.  AR Bruce to ask the Intel core to review the spec.

§ Need the patches unsquashed for the 3 squashed bugs

§ Intel looking into internal/external funding for the Nova bug list.  We will need help reproducing issues

§ Intel willing to discuss funding Horizon outsourcing to be PM'd by WR.  Bruce to ask Denise - do the vendors we are considering have Horizon experience?  UI experience?

• Review the tracking spreadsheet: https://docs.google.com/spreadsheets/d/s/edit?usp=sharing

§ Column U is blank.  Do we care?  Maybe - Brent thinks it's not needed, Dean thinks it would be good to have the back-pointers while doing the work.  Handle on-demand as needed.

§ Question - Move to a new sheet?  Permissions issues with this version.

· Let's move to a clean spreadsheet - Bruce to create.  Should be R/O for anyone with the link and we should have control over who can edit it.

• How do we want to track the Strategy part of this.  In the past this has been conversations, emails or other informal (non-tracked) methods.  Do we want to create short specs to submit for review?

• Testing - for features we inherit from Upstream can the Test team provide regression test cases?  Bruce to check with Ada & Numan.

• Raw cache (lixiaoy1) Ovidiu and Lisa  agreed on that we can replace StarlingX raw cache with Cinder image cache, and lists

           what we need to do in starlingx config and what we can enhance in Cinder image cache (add some tags in image properties to

           impact when the image is evicted from cache.)  We need Brent's confirmation. More detailed info please see the email threads 

           thread with subject "[Starlingx-discuss] [StarlingX] Use Cinder generic image cache to replace raw caching".

· Long term strategy

• Assuming STX r/2019.05 will rebase to Stein branches and STX master will continue on master