Re: Jenkins External Artifact Storage core changes in LTS

[Oliver Gondža]

On 2018-04-13 18:23, Carlos Sanchez wrote:
> Hi Oliver,
>
> There are some changes for core that would be really helpful to have in
> hands of people for testing sooner than later
>
> https://github.com/jenkinsci/jenkins/pull/3302
>
> All new methods are marked with the new @Beta annotation so it should be
> clear that it's not final.
>
> We expect this to be in the next weekly or so,but I would really like to
> ensure we can pull this into the next LTS if at all possible,what can I
> do to help prepare these changes for the LTS,even if we tried to
> backport it?

Hi, Let's make the discussion public.

My personal take is the LTS and Beta.class are quite orthogonal in
principle. When it comes to the actual change, it feels far from trivial
for such an exception IMO. What is the motivation here? Backporting this
will get it in LTS in 4 weeks, otherwise it will get in in 8 weeks.

Thoughts? I would really like to avoid backporting something before it
has its final shape (IOW, merged to master).
--
oliver

[R. Tyler Croy]

(replies inline)

Hey Oliver, I was chatting with Carlos about this earlier and encouraged him to
reach out to you to better understand your thoughts for considering code coming
into the LTS.

From my understanding of what Carlos is asking about, is not pulling anything
into an LTS which isn't already in a Weekly release. Rather, I think Carlos is
apprehensive about the timing between Weekly releases and LTS, and is worried
about missing the last Weekly before our typical 2-weeks-before LTS cutoff."

I think Carlos's team has worked quite hard to get the underlying code to
enable external artifacts ready, and considers one indication of "success" to
have that code delivered in our next LTS cycle, since LTSes are consumed by the
vast majority of Jenkins users who would benefit from this functionality.


I *hope* I've captured the situation clearly.


The questions Carlos has about LTS timing I think are all moot assuming that
the aforementioned pull request gets reviewed and merged in a timely manner for
next week's release (from my understanding).



Cheers

[Carlos Sanchez]

Tyler has summarized it correctly, I expect this to be in master soon but just want to make sure we don't miss the LTS.

Getting useful feedback and having final implementations ready in the future for this type of change relies on putting this in the hands of as many people as possible.

That's why the API changes use @beta annotations (and I believe we would be the first ones using it), to get new functionality in the hands of users asap, improve the velocity of new feature delivery while ensuring that stability is not affected.

Cheers

[Oliver Gondža]

Thanks for the details, guys. I have rechecked all new API is annotated
with Beta. I am wondering, what plugins are going to consume this? I am
trying to understand what userbase will be exposed to the change, what
use cases can be affected.

Having said that, I see this backportable into .3 provided it will be
merged by the end of the backporting window. Also, It would be great if
you can either provide ATH tests for the new plugin functionality for RC
testing window or retest and report during that window.

Thanks!
--
oliver

[Carlos Sanchez]

On Mon, Apr 16, 2018 at 9:32 AM, Oliver Gondža <ogo...@gmail.com> wrote:

Thanks for the details, guys. I have rechecked all new API is annotated with Beta. I am wondering, what plugins are going to consume this? I am trying to understand what userbase will be exposed to the change, what use cases can be affected.

To be clear, this would be for the next LTS picked in May, not for the current in progress one.

Existing artifact manager usage is unaffected, everything continues to work as before storing things on disk.

Only when the new plugin implementing the artifact manager for s3 is enabled artifacts would go to S3 

https://github.com/jenkinsci/jep/blob/master/jep/202/README.adoc#prototype-implementation

The new one is artifact-manager-s3-plugin

Related PRs

• JENKINS-26810 File attribute/symlink support in VirtualFile - fixed as a side effect

• JENKINS-49635 Permit VirtualFile to serve external file contents

• Jenkins core PR-3302

• WAR packaging implementation for the reference implementation

Downstream improvements to use the new APIs

• workflow-api-plugin PR-67

• workflow-basic-steps-plugin PR-60

• copyartifact-plugin PR-100

• compress-artifacts-plugin PR-1

 

Having said that, I see this backportable into .3 provided it will be merged by the end of the backporting window. Also, It would be great if you can either provide ATH tests for the new plugin functionality for RC testing window or retest and report during that window. 

Thanks!
--
oliver

--
You received this message because you are subscribed to the Google Groups "Jenkins Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-dev+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-dev/9760335c-57cd-0d30-cc90-0053107cfbcb%40gmail.com.
For more options, visit https://groups.google.com/d/optout.