[ 
https://issues.apache.org/jira/browse/CASSANDRA-21629?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18109870#comment-18109870
 ] 

Brandon Williams edited comment on CASSANDRA-21629 at 8/31/26 12:43 PM:
------------------------------------------------------------------------

The 4.x PRs contain the backports, and the rest contain a utility to run CI 
locally and a couple small bugfixes encountered during the backports.

The way I approached this was to first create the `run-ci-local` script in 5.0 
that runs a test via the docker images just like CI, and then used that to spot 
check a cell of each type and iterate on those on 4.x, where I first copied the 
5.0 CI infra wholesale.  This results in a commit of just the needed changes 
for 4.0 or 4.1 compared to the existing infra, which I hope facilitates review.

Currently `run-ci-local` duplicates what is in the `pipelineProfiles` in the 
Jenkinsfile, but in a follow-up ticket we can move the source of truth to a 
json file and have Groovy parse that for the profiles.

Note this post-installs the JDKs required.  This might not work well for CI, 
but I didn't want to need a docker push to begin testing.


was (Author: brandon.williams):
The 4.x PRs contain the backports, and the rest contain a utility to run CI 
locally and a couple small bugfixes encountered during the backports.

The way I approached this was to first create the `run-ci-local` script in 5.0 
that runs a test via the docker images just like CI, and then used that to spot 
check a cell of each type and iterate on those on 4.x, where I first copied the 
5.0 CI infra wholesale.  This results in a commit of just the needed changes 
for 4.0 or 4.1 compared to the existing infra, which I hope facilitates review.

Note this post-installs the JDKs required.  This might not work well for CI, 
but I didn't want to need a docker push to begin testing.

Currently `run-ci-local` duplicates what is in the `pipelineProfiles` in the 
Jenkinsfile, but in a follow-up ticket we can move the source of truth to a 
json file and have Groovy parse that for the profiles.

> Modernize the 4.x CI pipeline (Backport CASSANDRA-18145)
> --------------------------------------------------------
>
>                 Key: CASSANDRA-21629
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-21629
>             Project: Apache Cassandra
>          Issue Type: Improvement
>          Components: CI
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>            Priority: Normal
>             Fix For: 4.0.x, 4.1.x, 5.0.x, 6.x, 7.x
>
>          Time Spent: 50m
>  Remaining Estimate: 0h
>
> In CASSANDRA-18145 the k8s-based immutable provisioning and repeatable CI was 
> added and has enjoyed great success.  It was forward-looking and applied to 
> the 5.0 branch and up, but now with a year of hindsight we can see there have 
> been eight 4.x releases and these won't stop until 7.0 is released.  It would 
> be beneficial to the project to have a unified CI process for all branches 
> before that occurs.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to