[Gluster-devel] [Gluster-users] Idea: Alternate Release process

Kaushal M kshlmster at gmail.com
Thu May 19 05:43:08 UTC 2016 

I'm in favour of a stable release every 2 months and an LTS once a
year (option 2).

As Oleksander already suggested, I'm in favour of having well defined
merge windows, freeze dates and testing period.
(A slightly modified timeline from Oleksander's proposal follows)
For every 2 month window,
- 1 month of development (merge window). New features will not be
accepted post this period.
- At the end of the development period a release-candidate 1 is tagged.
- A 1 month testing/bug-fixing window follows. This time is for
testing and fixing bugs found.
 Feature development and reviews can happen during this period on
gerrit, but nothing gets merged. Merges will be held till the next
merge window.
- At least 2 more release-candidates will be release once every fortnight
- The final release-candidate becomes the release if it passes all our tests.

One of the 6 releases of the year will become a LTS release.
The 2 month window for this release will mainly be targeted at making
the release stable.
New features should be minimal for this release.

During every 2 month window, the LTS release will get any required bug
fixes and stability improvements backported.
For fix to be backported it needs to be present in a stable release.

~kaushal

On Wed, May 18, 2016 at 11:46 PM, Atin Mukherjee <amukherj at redhat.com> wrote:
> A bit late but better than never. My vote is for option 2.
>
> ~Atin
>
> On 05/18/2016 07:19 PM, Vijay Bellur wrote:
>> [Adding gluster-users]
>>
>> I would like to wrap this poll by the next community meeting on 25th
>> May. Can you please weigh in with your opinions on the options
>> provided by Aravinda?
>>
>> Thanks!
>> Vijay
>>
>>
>> On Fri, May 13, 2016 at 4:16 AM, Aravinda <avishwan at redhat.com> wrote:
>>> Hi,
>>>
>>> Based on the discussion in last community meeting and previous discussions,
>>>
>>> 1. Too frequent releases are difficult to manage.(without dedicated release
>>> manager)
>>> 2. Users wants to see features early for testing or POC.
>>> 3. Backporting patches to more than two release branches is pain
>>>
>>> Enclosed visualizations to understand existing release and support cycle and
>>> proposed alternatives.
>>>
>>> - Each grid interval is 6 months
>>> - Green rectangle shows supported release or LTS
>>> - Black dots are minor releases till it is supported(once a month)
>>> - Orange rectangle is non LTS release with minor releases(Support ends when
>>> next version released)
>>>
>>> Enclosed following images
>>> 1. Existing Release cycle and support plan(6 months release cycle, 3
>>> releases supported all the time)
>>> 2. Proposed alternative 1 - One LTS every year and non LTS stable release
>>> once in every 2 months
>>> 3. Proposed alternative 2 - One LTS every year and non LTS stable release
>>> once in every 3 months
>>> 4. Proposed alternative 3 - One LTS every year and non LTS stable release
>>> once in every 4 months
>>> 5. Proposed alternative 4 - One LTS every year and non LTS stable release
>>> once in every 6 months (Similar to existing but only alternate one will
>>> become LTS)
>>>
>>> Please do vote for the proposed alternatives about release intervals and LTS
>>> releases. You can also vote for the existing plan.
>>>
>>> Do let me know if I missed anything.
>>>
>>> regards
>>> Aravinda
>>>
>>> On 05/11/2016 12:01 AM, Aravinda wrote:
>>>
>>> I couldn't find any solution for the backward incompatible changes. As you
>>> mentioned this model will not work for LTS.
>>>
>>> How about adopting this only for non LTS releases? We will not have backward
>>> incompatibility problem since we need not release minor updates to non LTS
>>> releases.
>>>
>>> regards
>>> Aravinda
>>>
>>> On 05/05/2016 04:46 PM, Aravinda wrote:
>>>
>>>
>>> regards
>>> Aravinda
>>>
>>> On 05/05/2016 03:54 PM, Kaushal M wrote:
>>>
>>> On Thu, May 5, 2016 at 11:48 AM, Aravinda <avishwan at redhat.com> wrote:
>>>
>>> Hi,
>>>
>>> Sharing an idea to manage multiple releases without maintaining
>>> multiple release branches and backports.
>>>
>>> This idea is heavily inspired by the Rust release model(you may feel
>>> exactly same except the LTS part). I think Chrome/Firefox also follows
>>> the same model.
>>>
>>> http://blog.rust-lang.org/2014/10/30/Stability.html
>>>
>>> Feature Flag:
>>> --------------
>>> Compile time variable to prevent compiling featurerelated code when
>>> disabled. (For example, ./configure--disable-geo-replication
>>> or ./configure --disable-xml etc)
>>>
>>> Plan
>>> -----
>>> - Nightly build with all the features enabled(./build --nightly)
>>>
>>> - All new patches will land in Master, if the patch belongs to a
>>>    existing feature then it should be written behind that feature flag.
>>>
>>> - If a feature is still work in progress then it will be only enabled in
>>>    nightly build and not enabled in beta or stable builds.
>>>    Once the maintainer thinks the feature is ready for testing then that
>>>    feature will be enabled in beta build.
>>>
>>> - Every 6 weeks, beta branch will be created by enabling all the
>>>    features which maintainers thinks it is stable and previous beta
>>>    branch will be promoted as stable.
>>>    All the previous beta features will be enabled in stable unless it
>>>    is marked as unstable during beta testing.
>>>
>>> - LTS builds are same as stable builds but without enabling all the
>>>    features. If we decide last stable build will become LTS release,
>>>    then the feature list from last stable build will be saved as
>>>    `features-release-<NUM>.yaml`, For example:
>>>    features-release-3.9.yaml`
>>>    Same feature list will be used while building minor releases for the
>>>    LTS. For example, `./build --stable --features features-release-3.8.yaml`
>>>
>>> - Three branches, nightly/master, testing/beta, stable
>>>
>>> To summarize,
>>> - One stable release once in 6 weeks
>>> - One Beta release once in 6 weeks
>>> - Nightly builds every day
>>> - LTS release once in 6 months or 1 year, Minor releases once in 6 weeks.
>>>
>>> Advantageous:
>>> -------------
>>> 1. No more backports required to different release branches.(only
>>>     exceptional backports, discussed below)
>>> 2. Non feature Bugfix will never get missed in releases.
>>> 3. Release process can be automated.
>>> 4. Bugzilla process can be simplified.
>>>
>>> Challenges:
>>> ------------
>>> 1. Enforcing Feature flag for every patch
>>> 2. Tests also should be behind feature flag
>>> 3. New release process
>>>
>>> Backports, Bug Fixes and Features:
>>> ----------------------------------
>>> - Release bug fix - Patch only to Master, which will be available in
>>>    next beta/stable build.
>>> - Urgent bug fix - Patch to Master and Backport to beta and stable
>>>    branch, and early release stable and beta build.
>>> - Beta bug fix - Patch to Master and Backport to Beta branch if urgent.
>>> - Security fix - Patch to Master, Beta and last stable branch and build
>>>    all LTS releases.
>>> - Features - Patch only to Master, which will be available in
>>>    stable/beta builds once feature becomes stable.
>>>
>>> FAQs:
>>> -----
>>> - Can a feature development take more than one release cycle(6 weeks)?
>>> Yes, the feature will be enabled only in nightly build and not in
>>> beta/stable builds. Once the feature is complete mark it as
>>> stable so that it will be included in next beta build and stable
>>> build.
>>>
>>>
>>> ---
>>>
>>> Do you like the idea? Let me know what you guys think.
>>>
>>> This reduces the number of versions that we need to maintain, which I like.
>>> Having official test (beta) releases should help get features out to
>>> testers hand faster,
>>> and get quicker feedback.
>>>
>>> One thing that's still not quite clear to is the issue of backwards
>>> compatibility.
>>> I'm still thinking it thorough and don't have a proper answer to this yet.
>>> Would a new release be backwards compatible with the previous release?
>>> Should we be maintaining compatibility with LTS releases with the
>>> latest release?
>>>
>>> Each LTS release will have seperate list of features to be enabled. If we
>>> make any breaking changes(which are not backward compatible) then it will
>>> affect LTS releases as you mentioned. But we should not break compatibility
>>> unless it is major version change like 4.0. I have to workout how we can
>>> handle backward incompatible changes.
>>>
>>> With our current strategy, we at least have a long term release branch,
>>> so we get some guarantees of compatibility with releases on the same branch.
>>>
>>> As I understand the proposed approach, we'd be replacing a stable
>>> branch with the beta branch.
>>> So we don't have a long-term release branch (apart from LTS).
>>>
>>> Stable branch is common for LTS releases also. Builds will be different
>>> using different list of features.
>>>
>>> Below example shows stable release once in 6 weeks, and two LTS releases in
>>> 6 months gap(3.8 and 3.12)
>>>
>>> LTS 1 : 3.8    3.8.1  3.8.2  3.8.3   3.8.4   3.8.5...
>>> LTS 2 :                              3.12    3.12.1...
>>> Stable: 3.8    3.9    3.10   3.11    3.12    3.13...
>>>
>>> A user would be upgrading from one branch to another for every release.
>>> Can we sketch out how compatibility would work in this case?
>>>
>>> User will not upgrade from one branch to other branch, If user interested in
>>> stable channel then upgrade once in 6 weeks. (Same as minor update in
>>> current release style)
>>>
>>>
>>> This approach work well for projects like Chromium and Firefox, single
>>> system apps
>>>   which generally don't need to be compatible with the previous release.
>>> I don't understand how the Rust  project uses this (I am yet to read
>>> the linked blog post),
>>> as it requires some sort of backwards compatibility. But it too is a
>>> single system app,
>>> and doesn't have the compatibility problems we face.
>>>
>>> Gluster is a distributed system, that can involve multiple different
>>> versions interacting with each other.
>>> This is something we need to think about.
>>>
>>> I need to think about compatibility, What new problems about the
>>> compatibility with this approach compared to our existing release plan?
>>>
>>>
>>> We could work out some sort of a solution for this though.
>>> It might be something very obvious I'm missing right now.
>>>
>>> ~kaushal
>>>
>>> --
>>> regards
>>> Aravinda
>>>
>>> _______________________________________________
>>> Gluster-devel mailing list
>>> Gluster-devel at gluster.org
>>> http://www.gluster.org/mailman/listinfo/gluster-devel
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Gluster-devel mailing list
>>> Gluster-devel at gluster.org
>>> http://www.gluster.org/mailman/listinfo/gluster-devel
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://www.gluster.org/mailman/listinfo/gluster-users
>>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-devel

More information about the Gluster-devel mailing list