Charmed Apache Spark (revision 6)¶
July 29th, 2026
We’re excited to announce a new stable release for Charmed Apache Spark.
This release most notably brings the support for a Charmed Apache Spark Terraform Module to deliver a seamless, production-ready and fully open-source data lake experience. Moreover, this new release also comes with enhanced security and bug fixes to the various components that makes the Charmed Apache Spark solution.
Charmhub: 3.5/stable | Docs | Deploy guide | System requirements
Features¶
This release includes the following major feature:
Enhancements¶
This release includes general enhancements across the solution, as well as to the individual components, as follows.
General¶
[PRA-322] Components upgrade:
Apache Spark versions: 3.5.8-ubuntu1
Apache Kyuubi versions: 1.10.3-ubuntu4
NVIDIA Spark-RAPIDS version: 26.04.2
General updates of Python dependencies, craft build tools, CI workflows and Github actions
[PRA-63] Run integration tests using spread
[PRA-264] Improve TIOBE workflow reliability
Apache Kyuubi¶
Apache Spark History Server¶
Spark Integration Hub¶
Apache Spark Client snap¶
Component bumps (see General section for versions of various components)
[MISC] Reduce the K8s matrix validation and disable fail-fast feature (#148) (#149)
[PRA-211] Reorder PYTHONPATH to give priority to the snap’s stdlib (#146) (#150)
[PRA-256] Set driver metrics sink to JmxSink for shell entrypoints (#159) (#160)
[MISC] Remove workflow on_spark_update_available (#154)
[MISC] Fix permissions for snap release workflow (#156)
chore: adding scheduled test runs on weekends (#168)
[MISC] Update renovate configuration (#162)
Canonical security maintained OCI Images for Apache Spark¶
Charmed Apache Spark Terraform Module¶
[PRA-101] Remove deprecated mailing and updating lock file (#222)
[MISC] Let users decide whether to use COS in UAT tests (#232)
[PRA-306] Unpin juju-agent-version in Spark K8s Bundle integration tests (#231)
[PRA-7][KF-8066] Enable Spark <> Kubeflow integration with new standard (#233)
[PRA-318] Implement automated OCI getter for our products (#240)
[PRA-324] Split bundle by tracks
[PRA-330] Update Postgresql to latest revision on 14/stable (#251)
[PRA-324] Update renovate configuration (3.5) (#253) (#255) (#256)
[PRA-312] Split TLS private key and admin password secrets (#241)
Bug Fixes¶
This release includes several bug fixes across the solution, which are listed below categorized to individual components.
Apache Kyuubi¶
Apache Spark History Server¶
Spark Integration Hub¶
Canonical security maintained OCI Images for Apache Spark¶
Breaking Changes¶
This release includes the following breaking change:
Charmed Apache Spark Terraform Module¶
Documentation improvements¶
The current release also features the following documentation changes:
docs: FE Feedback fixes (#210)
[PRA-309] Update docs for s3-integrator 2/stable and multi-track awareness (#237)
[PRA-165] Update docs to reflect correct behavior when S3 region is not configured (#257) (#284) (#286)
[PRA-324] Adapt docs content to match Spark version on various tracks (#252) (#258) (#254)
Security¶
The following CVEs have been fixed in the new artifacts:
Component |
Severity |
Fixed |
|---|---|---|
Apache Spark |
High |
CVE-2025-48734, CVE-2025-67721, CVE-2026-24281, CVE-2026-24308, CVE-2025-54920 |
Medium |
CVE-2026-34477, CVE-2026-34478, CVE-2026-34479, CVE-2026-34480 |
|
Apache Kyuubi |
High |
CVE-2025-48734, CVE-2026-33870, CVE-2026-33871, CVE-2026-35554, CVE-2026-42198, CVE-2026-42577, CVE-2026-42579, CVE-2026-42583, CVE-2026-42584, CVE-2026-42587, CVE-2026-44249, CVE-2026-45416, CVE-2026-45674, CVE-2026-47691, CVE-2026-50010 |
Medium |
CVE-2026-33558, CVE-2026-34477, CVE-2026-34478, CVE-2026-34479, CVE-2026-34480, CVE-2026-41417, CVE-2026-42580, CVE-2026-42581, CVE-2026-42585, CVE-2026-45536, CVE-2026-45673, CVE-2026-47244, CVE-2026-48043, CVE-2026-50020, CVE-2026-50560, CVE-2026-6860 |
|
Low |
CVE-2026-42578 |
Compatibility¶
The following table summarize the compatibility matrix of the solution:
Note
Model destruction for controllers above 3.6.18+ may sometimes freeze (see Juju issue #22105). In these cases, we recommend destroying the resources manually.
Acknowledgements¶
We are extremely grateful to the Apache Spark and Apache Kyuubi communities for their continuous work, involvement and engagement with open-source to make technologies that process data at scale available to the broader audience.