You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@unomi.apache.org by Serge Huber <sh...@apache.org> on 2025/11/14 09:16:41 UTC
[Minutes] Project Meeting – November 13th
Hi all,
Here are the meeting minutes from our Zoom meeting held yesterday. As
usual if you notice anything wrong or missing please reply to this
thread.
cheers,
Serg.e..
Attendees:
• Jonathan Sinovassin-Naïk (Jahia, Apache Unomi PMC)
• Francois Gerthoffert (Jahia, Apache Unomi PMC)
• Jérôme Blanchard (Jahia, Apache Unomi PMC)
• Serge Huber (Inoyu, Apache Unomi PMC Chair)
Below is the consolidated summary of the discussions held during the meeting.
⸻
1. Status of the 3.0 Release
The team reviewed the state of the Apache Unomi 3.0 [1] release:
• Version 3.0 has been successfully released and appears to be
functioning correctly.
• Jahia’s internal environment migrations are complete, and customer
environment migrations are expected to begin within approximately
fifteen days.
• The Jahia Cloud Team is preparing the remaining infrastructure
needed before customer migrations can start.
• Jahia also noted longer-term internal objectives related to retiring
older intermediate environments, driven partly by Elastic Cloud [2]
constraints and requirements from on-premise security teams.
⸻
2. Infrastructure, Docker Images & Kubernetes Context
Several points regarding deployment and operational practice were discussed:
• At Jahia, the Unomi-based stack runs on Elasticsearch [3] and Elastic Cloud.
• The Jahia Cloud Team currently does not use the project’s standard
Docker images.
• The official Unomi Docker images still run as root, which is
discouraged in Kubernetes [4] contexts.
• Inoyu is standardizing its own deployments on Kubernetes.
• Findings from this work may help identify improvements or issues
relevant to Unomi’s standard Docker images.
• Broader updates to containerization and deployment practices are
expected as part of a medium-term evolution on both sides.
⸻
3. Progress on 3.1 Development
Updates from the development branch included:
• The OpenSearch [5] PR has been fully revised, with non-OpenSearch
components removed.
• Conflicts introduced by earlier release changes have been resolved.
• The OpenSearch branch test suite now passes.
• In parallel, work continues on the unomi-3-dev branch, where not all
tests pass yet and additional fixes are underway.
• Integration updates involving Apache Karaf [6] were merged.
• Several smaller PRs were prepared to simplify reviews, such as:
• Build script improvements
• Adoption of Maven build caching [7]
• A progress bar for integration tests showing remaining time and test results
The OpenSearch PR has been open since January and will require review
once Jahia’s development team has additional availability.
⸻
4. Work on unomi-3-dev: Test Failures & Item ID Prefix Handling
While working on the unomi-3-dev branch and investigating integration
test failures, the team identified an issue:
• A problem involving item ID prefixes caused some Unomi types not to
be found, even though they existed.
• The behaviour appears linked to the code responsible for
synchronizing item IDs and meta IDs.
• This issue is not part of the Elasticsearch 7→9 migration work; it
was discovered during debugging of the unomi-3-dev branch test
failures.
• Investigation and fixes are ongoing as part of the work to get the
full test suite passing on unomi-3-dev.
⸻
5. Backports, Master Branch Merges & Test Infrastructure
The team discussed ongoing development workflow topics:
• Backporting changes from the master branch continues, though some
test failures remain to be resolved.
• Smaller and more isolated backports (e.g., graph-generation
improvements) help accelerate integration.
• Merging the OpenSearch work into the master branch as soon as
feasible was noted as important for unblocking future work.
• Improvements to both the unit test framework and migration-related
tests are ongoing.
⸻
6. Deployment Strategy for the 3.1 Line
The team discussed several options for upcoming releases:
• Various approaches for the 3.1 release were considered, including a
potential 3.1.0 Beta.
• The possibility of splitting major feature tracks across multiple
releases was discussed—for example:
• 3.1 focusing primarily on OpenSearch
• 3.2 focusing on larger features, such as multi-tenancy
• Whether an “innovation branch” would help maintain momentum while
keeping master stable was also considered.
• Reducing PR size remains important for speeding up review cycles.
• A 3.0.x maintenance series is expected as Jahia proceeds with
customer migrations.
These subjects will continue to be evaluated as development progresses.
⸻
References
[1] Apache Unomi — https://unomi.apache.org/
[2] Elastic Cloud — https://www.elastic.co/cloud/
[3] Elasticsearch — https://www.elastic.co/elasticsearch/
[4] Kubernetes — https://kubernetes.io/
[5] OpenSearch — https://opensearch.org/
[6] Apache Karaf — https://karaf.apache.org/
[7] Apache Maven — https://maven.apache.org/