Open Stack

On 22 November 2017 at 09:32, Thierry Carrez <thierry at openstack.org> wrote:
> Thierry Carrez wrote:
>> Now one could argue that the development of OpenStack has slowed down
>> enough that a release cycle of 9 months or one year (including only one
>> PTG per cycle) would be a better fit.
>
> Let me expand on that.
>
> As OpenStack matures, it takes longer for features to land as we are
> much more careful about not breaking our existing users. There is less
> pressure to get a feature landed in a given release. Time just generally
> goes slower, and personally I feel like cycles are getting shorter, with
> the next release always just around the corner. Who didn't get the
> feeling that the Denver PTG was pretty close to the Atlanta one ?

I agree.

>
> As we are working to make OpenStack development more palatable to
> part-time contributors, removing a bit of the time pressure might
> actually be a good idea too. Development goes slower when you can only
> spend 20% of your work time on it.

I agree there too.

>
> Yes, we need to be able to ship early and often, but projects can use
> the intermediary releases model to do that. The difference with the
> "final" release of a cycle is that we cut a stable branch out of it. We
> might not need to maintain a stable branch every 6 months.

There are pros and cons of making these cycles longer.
What I see for the project I contribute the most:
Pros:
- Less pressure to contributors
- Probably less backports (more things will stay in a cycle)
- Less branches to maintain (except if LTS conversation opens, but
that's another _related_ topic)
Cons:
- Backports could probably be more painful and less straightforward
I'll not add the likelihood (or not) to introduce larger features, the
travel fatigue.

I generally think it's a good thing, if done well.

One important thing would be, at least for me, that the community
still aligns together, by keeping (minimum):
1) some kind of feature freeze
2) an official _all across the board_ release
3) regular aligned milestones

>
> Again, this won't solve the upgrade pressure/pain (skip-level upgrades
> is your friend there), nor is it going to solve your LTS needs (won't
> make branches live for 3+ years). It's just acknowledging that OpenStack
> is maturing, development pace is less hectic, and our 6-months cycles
> might (might) put unnecessary pressure and overhead on us for limited gains.

Agreed.

>
> --
> Thierry Carrez (ttx)
>
> _______________________________________________
> openstack-sigs mailing list
> openstack-sigs at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-sigs

Best regards,
JP