Placement update 19-11! Our little friend has come so far: a
stable/stein branch was cut this week.
Soon after RC1 was created, we discovered an issue in the PlacementFixture, used by nova. This was fixed and backported so there will be an RC2, toward the end of next week. In the meantime we should be trying to find release critical bugs and looking for feedback from others.
There are (thus far) two PTG related etherpads where you may want to leave some placement-planning-related thoughts:
- There's been an initial stein release.
- That release included install from PyPI and Upgrading from Nova instructions. Both of which probably need to be validated by people who are not close to the details. Note that the state of the distro-related docs remains undefined.
The Monday nova scheduler meeting has been renamed to placement. I'll be taking over from Eric as the usual chair as he's busy with other things. While we are still establishing ourselves as a new project, the meeting seems like a good idea, but we should decide if/how we want to phase it out (or into office hours). Feel free to respond with your thoughts (on this or anything else in the post).
I sent out a message earlier in the week asking Nova cores to confirm if they'd like to stay on as Placement core. I'll give it a few more days and then remove nova-cores as an included member.
In-tree specs are going to happen (see next section).
A message was sent asking for volunteers for cross project liaisons.
We've got a StoryBoard project
now. I've started using it. Tagging bugs with
bug and also making
rfe tags to indicate things that needs to be
cleaned up or feature requests.
Please be prepared for these structures to evolve as we gain some understanding of how StoryBoard works.
There are still bugs in launchpad and we need to continue to watch there:
osc-placement is currently behind by 13 microversions.
Prepping the RC
Everything listed last week has been done except for:
- Ensuring the install docs are sane and complete. I have asked packaging-related people for their input, as they're the ones who know how their packages are (or will be) set up.
Be thinking about what you'd like the main themes to be. Put them on the PTG etherpad.
With regard to the PTG, because we will have limited time, we should do as much of the discussion in email prior to the PTG so that when we get to the PTG we are resolving the difficult problems, not discovering what they are.
https://review.openstack.org/#/q/topic:cd/gabbi-tempest-job Gabbi-based integration tests of placement. These recently found a bug that none of the functional, grenade, nor tempest tests did. Not release related, but useful testing.
https://review.openstack.org/#/q/topic:bp/negative-aggregate-membership Negative member of aggregate filtering resource providers and allocation candidates. Work on this can go ahead now that
stable/steinhas been cut.
https://review.openstack.org/#/c/645255/ This is a start at unit tests for the PlacementFixture. It is proving a bit "fun" to get right, as there are many layers involved. Making sure seemingly unrelated changes in placement don't break the nova gate is important. Besides these unit tests, there's discussion on the PTG etherpad of running the nova functional tests, or a subset thereof, in placement's check run.
On the one hand this is a pain and messy, but on the other consider what we're enabling: Functional tests that use the real functionality of an external service (real data, real web requests), not stubs or fakes.
Other Service Users
We'll hold off here until the final RC is cut. In the future if you stick "placement" somewhere in your commit message I'll probably eventually find your in-progress placement-related changes. A quick scan indicates there's quite a lot of interesting work in progress.