70 lines
3.7 KiB
Plaintext
70 lines
3.7 KiB
Plaintext
Hello everyone,
|
|
|
|
I'm asking for the opportunity to continue to serve as the PTL of Glance
|
|
into the Queens cycle.
|
|
|
|
Things looked good for Glance at the start of Pike. We had a strong set
|
|
of active cores, and discussed plans at the PTG for recruiting more
|
|
developers to the team through holding bugsmash events and developing a
|
|
more organized approach to triaging bugs to a point where they could be
|
|
worked on by developers not yet experienced in working with Glance. We
|
|
were interested in attracting new talent because we had recently lost
|
|
our team:diverse-affiliation tag [0] which was a warning that Glance had
|
|
become too reliant on two companies. And in fact when those companies
|
|
decided to change direction early in the cycle, the impact on Glance was
|
|
devastating. However, due to the hard work of those who remained and
|
|
developers from the wider community who pitched in [1], we were able to
|
|
complete a revised list of priorities for Pike and did not require any
|
|
kind of drastic intervention (such as going into maintenance mode, which
|
|
looked like a real possibility in April). And, on the plus side, we'll
|
|
definitely be receiving the team:diverse-affiliation tag the next time
|
|
it's assessed.
|
|
|
|
The point of mentioning the above history is that I'd like to continue
|
|
the reorganization and recruitment work we'd discussed at the beginning
|
|
of Pike, which should improve the project's health for the Rocky PTL,
|
|
who will most likely not be me. Or at least there's no reason why it
|
|
should be me, as some of my goals for Queens are to help grow the Glance
|
|
community and help any Glancer who's interested (hopefully, more than
|
|
one) get themselves into a position to become an effective PTL.
|
|
|
|
So that's the general nature of what I think we should focus on in
|
|
Queens. There are a few specifics that I should probably mention:
|
|
|
|
* Image import continued -- thanks to Erno's hard work, a minimal viable
|
|
product of image import has landed in Pike. There are a few obvious
|
|
Glance-side improvements we can make, but at Pike RC-1 time, I plan to
|
|
notify operators that the MVP of image import is available, they
|
|
should look it over, and give us feedback on what enhancements are
|
|
necessary, and think about how they can direct development resources
|
|
to Glance to make those changes happen.
|
|
|
|
* Rolling upgrades -- we released zero-downtime upgrades as an
|
|
experimental feature in Ocata, but further work planned for Pike
|
|
(primarily, in-gate testing of upgrades) did not happen. It's
|
|
important to get this completed in Queens.
|
|
|
|
* Operator communication -- the popular operator surveys that we were
|
|
using to gather usage information about Glance were another casuality
|
|
of the personnel situation in Pike; I'd like to get those going again.
|
|
|
|
Additionally, there are some small worthwhile features among the
|
|
untargeted approved specs that would be nice to land, and new specs are
|
|
already showing up in the Queens directory. And of course Glance will
|
|
participate in the Queens community goals. We'll sort out the
|
|
priorities at the PTG [2].
|
|
|
|
Finally, there's been some discussion on the ML about whether Glance has
|
|
outlived its usefulness and should be replaced [3,4]. I'm willing to
|
|
discuss that at the PTG, but as you can tell from the fact that I listed
|
|
it last, I think there's still a role for Glance in the Queens cycle.
|
|
|
|
Thanks for reading, and thanks for your consideration,
|
|
brian
|
|
|
|
[0] https://git.openstack.org/cgit/openstack/governance/commit/?id=c0359e95153c3d3688d4ef6902a058204913a87a
|
|
[1] http://lists.openstack.org/pipermail/openstack-tc/2017-July/001444.html
|
|
[2] https://etherpad.openstack.org/p/glance-queens-ptg-planning
|
|
[3] http://lists.openstack.org/pipermail/openstack-dev/2017-June/118228.html
|
|
[4] http://lists.openstack.org/pipermail/openstack-dev/2017-July/119442.html
|