71 lines
3.3 KiB
Plaintext
71 lines
3.3 KiB
Plaintext
Howdy!
|
|
|
|
I am submitting my name to continue as PTL for Chef OpenStack. If you
|
|
don't know me, I am scas on Freenode. I work for Workday, where I am an
|
|
active operator and upstream developer. I have contributed to OpenStack
|
|
since 2014, and joined the core team in early 2015. Since then, I have
|
|
served as PTL for four cycles. I am also an active member of the
|
|
Sous-Chefs organization, which fosters maintainership of community Chef
|
|
cookbooks that could no longer be maintained by their author(s). My life
|
|
as a triple threat, as well as being largely in the deploy automation
|
|
space, gives me a unique perspective on the use cases for Chef
|
|
OpenStack.
|
|
|
|
Development continues to run about a release behind the coordinated
|
|
release to stabilize due to contributor availability. In that time,
|
|
overall testing has improved to raise the overall testing confidence in
|
|
landing more aggressive changes. Local testing infrastructure tends to
|
|
run closer to trunk to keep a pulse on how upstream changes will affect
|
|
the cookbooks closer to review time. This, in turn, influences the
|
|
changes that do pass the sniff test.
|
|
|
|
For Stein, I would like to focus on some of the efforts started during
|
|
Rocky.
|
|
|
|
* Awareness and Community
|
|
|
|
Chef OpenStack is extremely powerful and flexible, but it is not easy
|
|
for new contributors to get involved. That is, if they can find it,
|
|
down the dark alley, through the barber shop, and behind the door with
|
|
a secret knock. Documentation has been a handful of terse Markdown
|
|
docs and READMEs that do not evolve as fast as the code, which I think
|
|
impacts visibility and artificially creates a barrier to entry. I
|
|
would like to place more emphasis on providing this more well-lit
|
|
entry point for new and existing users alike.
|
|
|
|
* Consistency and HA
|
|
|
|
Stability is never a given, but it is pretty close with Chef
|
|
OpenStack. Each change runs through multiple, iterative tests before
|
|
it hits Gerrit. However, not every change runs through those same
|
|
tests in the gate due to the gap between local and integration. This
|
|
natural gap has resulted in multiple chef-client versions and
|
|
OpenStack configurations testing each change. There have existed HA
|
|
primitives in the cookbooks for years, but there are no published
|
|
working examples. I am aiming to continue this effort to further
|
|
reducing the human element in executing the tests.
|
|
|
|
* Continued work on containerization
|
|
|
|
With efforts to deploy OpenStack in the context of containers, Chef
|
|
OpenStack has not shared in the fanfare. I shipped a very shaky dokken
|
|
support out of a hack day at the 2017 Chef Community Summit in
|
|
Seattle, and have refined it over time to where it's consistently
|
|
Doing A Thing. I have found regressions upstream (e.g. packaging), and
|
|
have conservatively implemented workarounds to coax things into
|
|
submission when the actual fix would take more months to land. I wish
|
|
to continue that effort, and expand to other Ansible-based and
|
|
Kitchen-based integration scenarios to provide examples of how to get
|
|
to OpenStack using Chef.
|
|
|
|
These are but some of my personal goals and aspirations. I hope to be
|
|
able to make progress on them all, but reality may temper those
|
|
aspirations.
|
|
|
|
I would love to connect with more new users and contributors. You can
|
|
reach out to me directly, or find me in #openstack-chef.
|
|
|
|
Thanks!
|
|
|
|
-scas
|