You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@madlib.apache.org by Dianjin Wang <wa...@gmail.com> on 2025/12/23 05:53:00 UTC
[DISCUSS] Ideas for Apache MADlib Renewal
Hi all,
As a recent contributor to Apache MADlib, I've been working on a few
PRs and observing the project's current state. With Greenplum going
closed source and some long-time contributors becoming less active, I
believe this is a pivotal moment for MADlib.
I'd like to share some thoughts and open them up for discussion. I'm
not a committer or PMC member — just hoping to help MADlib move
forward.
## Project Roadmap
1. PostgreSQL Support
- Add support for PostgreSQL 16/17/18
2. Apache Cloudberry Support
- Merge the in-progress PR [1] (currently blocked by a CI issue; root
cause identified and shared)
- Release MADlib 2.2(?) with official Cloudberry support
3. Greenplum Support
- Freeze Greenplum support at the 2.x line (given it's now closed source)
- Remove Greenplum support in the next major MADlib 3.0(?), unless
maintainers step forward
4. Infrastructure Modernization
- Rebuild test Docker images (current base images are outdated and
unreproducible)
- Optionally migrate CI from Jenkins to GitHub Actions for better
developer experience on GitHub
- Site:
* Modernize the website with a Markdown-based static site generator
instead of editing html manually
* Cleaner branch management (use `master` as working branch,
`asf-site` for publishing)
## Community Growth
1. Expand Visibility across the PG and Cloudberry community
- Register a MADlib community account on postgresql.org
- Set up a MADlib channel in the Cloudberry Slack workspace
- Cross-post release announcements or news to both PG and Cloudberry channels.
- Encourage PMC members, committers, and community users to present
MADlib at meetups and conferences
2. Contribution
* Lower the barrier to get involved and contribute
* Encourage and attract more users/developers to join the contribution team.
~~~
The above are just my thoughts. Appreciate feedback from the community:
* Do these ideas align with where MADlib should go next?
* Would it make sense to move forward with these efforts?
Looking forward to the discussion.
[1] https://github.com/apache/madlib/pull/627
Best,
Dianjin Wang