d7c037b7c0
Change-Id: I9769a314e6d2e1af2a6cc96abe906556edd72684
69 lines
3.4 KiB
Plaintext
69 lines
3.4 KiB
Plaintext
Hi all,
|
||
|
||
I'd like to formally communicate my desire to continue serving as the keystone
|
||
PTL for the upcoming Queen’s release. Despite some turbulence throughout the
|
||
Pike development cycle, keystone has managed to make progress on some long
|
||
standing issues. Even though the pace of development has decreased, I think we
|
||
can build momentum from the small victories in Pike and have a productive
|
||
Queens release. I'd like to direct our focus on the following areas for
|
||
continued improvement and stability of the keystone project throughout the
|
||
Queen’s cycle.
|
||
|
||
*Policy Improvements*
|
||
|
||
We dedicated a significant portion of our time this release getting policy into
|
||
code and documented. We also had meaningful discussions with operators and
|
||
developers, resulting in a plan that improves long-standing issues with
|
||
OpenStack policy enforcement. I'd like to carry this momentum into Queens and
|
||
ensure we’re implementing global role assignments. Additionally, I'd like to
|
||
work closely with the oslo team to find ways we can signal deprecations to
|
||
operators though the oslo.policy library. This will help us define a better set
|
||
of default roles in Rocky, and clean-up policy enforcement at each service.
|
||
Lastly, I look forward to championing the community goal to move policy and
|
||
documentation into code for all applicable projects.
|
||
|
||
*Application Credentials*
|
||
|
||
At the forum and while reviewing the specification, we realized just how
|
||
important this work is to our users. While it's unfortunate we didn't make as
|
||
much progress here as we hoped during Pike, the discussions this cycle
|
||
highlighted a lot of concerns with design as well as usability. I think we're
|
||
all better off and more prepared to address the last few tough design bits
|
||
during the PTG. One of my goals during the PTG is to facilitate those
|
||
conversations and verbosely communicate our approach. I think that will help us
|
||
keep the goal in mind as we drive towards delivering this in Queens.
|
||
|
||
*Unified Limits*
|
||
|
||
Addressing unified limits was another long-standing issue that we did a good
|
||
job of capturing and documenting throughout the Pike release [0]. Pending
|
||
available resources, it would be great to push this forward, starting with
|
||
unified limits in keystone. This will have a positive impact on any project
|
||
currently experiencing issues with quota and will make quota usability more
|
||
consistent overall.
|
||
|
||
*Testing*
|
||
|
||
Throughout Pike, our team spent a significant amount of time paying down
|
||
testing, technical debt. We cleaned-up and extended support for federated
|
||
testing, and we've integrated testing with our tempest plugin. We added
|
||
experimental support to test rolling upgrades and look forward to gating on
|
||
rolling upgrades in Queens. In the upcoming months, we need to focus our
|
||
efforts on better LDAP integration testing, which has been on our TODO list for
|
||
too long.
|
||
|
||
*Team Building*
|
||
|
||
Last, but certainly not least, we need to recruit new keystone contributors -
|
||
part or full-time. Some of our most experienced and well-versed developers
|
||
moved onto other projects outside of OpenStack. Luckily, the remaining team
|
||
members have ambitiously stepped up to fill the gaps. I want to ensure this
|
||
project is working optimally on all cylinders. In order for us to do this and
|
||
achieve our Queen’s goals, we need more contributors.
|
||
|
||
Thanks for reading and I look forward to seeing everyone in Denver,
|
||
|
||
Lance
|
||
|
||
[0] http://specs.openstack.org/openstack/keystone-specs/specs/keystone/ongoing/unified-limits.html
|