Here’s my attempt to recap the goal selection discussion we had last week at the Forum. Feel free to correct any misstatements and continue the discussion.
For reference, here’s the etherpad from the discussion:
Overall Goal Discussion
We started off by discussing the reason for having the release cycle goals and what we should actually be trying to achieve with them. There were some who expressed concerns about the current Rocky goal selections not being the right things we should be focusing on. The hope with the first part was to come to some sort of agreement, or at least common understanding, that would inform our selections for Stein.
Some thought the goals should be entirely operator facing. So things that have an obvious and direct improvement for operators. At least so far, our goal selection has mostly been to try to get one goal that is a visible thing like that while the other is more of a technical debt cleanup. The idea being that the tech debt ones will keep us in a good and healthy position to be able to continue to address operator and user needs more easily.
There was also a desire to make the goals more “fully baked” before making them a goal. This would mean having the necessary changes well documented with example patches for teams to refer to to help guide them in figuring out what needs to be done in their own repos.
There was also the desire to make these goals something that can generate some excitement and be things that can be more of a marketing message. Things like “OpenStack services now support live configuration changes” vs. “OpenStack got rid of a testing library that no one has heard of”.
And I almost missed it, but there was a great suggestion to have a #goals channel for folks to go to for help and to discuss goal implementation. I really like this idea and will bring it up in the next TC office hour to see if we can get something official set up.
We ended up with only 10-15 minutes to actually discuss some ideas for goal selection for Stein. This was expected and planned. It will take some further discussion and thought before I think we are ready to actually pick some goals.
Some of the more popular ones brought up in the session were:
- Cold upgrade support
- Python 3 first
- Addition of “upgrade check” CLIs
We were also able to already identify some possible goals for the T cycle:
- Move all CLIs to python-openstackclient
- Adopt a larger set of default roles
We’ve been collecting a “goal backlog” with these and other ideas here:
All ideas (and support implementing them!) are welcome. Please join the discussion on the openstack-dev mailing list if you would like to get involved.