Contributor meetings 2025 - trinodb/trino GitHub Wiki

Trino Contributor Call, 26 Nov 2025

📹 Video recording on YouTube with timestamps

Attendees

mosabua (Host and organizer), pettyjamesm, roey ogen, raunaq morarka, wendigo (Mateusz), doug balog laurence brown, omer raifler, blake jones

Topics, notes and action items

00:27 Intro & Trino Gateway Release

Manfred (mosabua) opened the call, noting it is likely the last one of the year. He mentioned that Trino Gateway is working towards a release before the end of the year

00:48 Iceberg Support & Roadmap

Roey Ogen raised concerns about Iceberg at scale (dangling deletes, Z-order, v3 support) and asked if Iceberg is considered a "first-class citizen" in Trino . Manfred confirmed it is definitely a first-class citizen. Raunaq Morarka explained the development challenges: complex features require dedicated time, and contributors often rotate in and out. He noted that Z-order is largely superseded by liquid clustering (driven by Databricks/Delta features), making it less of a priority. Suggestion: Keep contributions small and targeted (e.g., procedures for dangling deletes) to get them reviewed and merged faster

[08:55] Resource Groups & Query Management

Roey asked why Resource Groups don't kill queries that exceed quotas Raunaq clarified that Resource Groups were designed by Facebook purely for admission control, not runtime resource management

Advice: To manage memory or "hogging" queries, use query_max_memory_per_node or Fault Tolerant Execution (FTE), which can kill and retry tasks based on resource constraints

14:48 Iceberg Production Issues (Split/FileSystem Errors)

Doug Balog reported production issues (referencing #26262 and #26853) involving "cannot open split" errors during concurrent operations and snapshot expiratio

Raunaq and Mateusz (wendigo) suggested using OpenTelemetry traces or enabling debug logging on the file system wrapper to trace the chain of events .

James Petty noted that many "S3-compatible" storage systems have subtle incompatibilities that can cause such issues.

23:36 PR: Iceberg Identifier Fields

Laurence Brown discussed PR #27026, which preserves identifier fields in Iceberg metadata—critical for Flink upsert workflows []. The group agreed to ping Martin for a final review.

29:22 PR: Ranger Plugin Jetty Downgrade

Laurence also raised a PR regarding a Jetty version downgrade in the Ranger plugin. Mateusz confirmed the change looks fine since the plugin has its own isolated classpath.

32:46 Other Updates

Omer Raifler asked about his metrics PR. Mateusz believed it had already been merged .

Release Automation: Manfred asked about automating Docker releases for Trino Gateway. Mateusz explained the current state of Airlift's release flow (GitHub Packages) and the need for credentials to push to Docker Hub.

Trino Contributor Call, 22 Oct 2025

📹 Video recording on YouTube with timestamps

Attendees

mosabua (Host and organizer), dprophet, pettyjamesm, xkrogen, vishalya, sajjoseph, Satish Chintahippu (@sc250072), Eden Kiknadze

Topics, notes and action items

00:09 Teradata Connector: PR and Licensing Update

Satish provided an update on the Teradata connector PR. The license is being updated to be similar to the Oracle JDBC driver, which should unblock redistribution. Satish requested help from the Trino team (Dane/David) to configure secrets in the GitHub workflow for the Teradata integration tests. Manfred agreed to start a group chat for provisioning the secrets

05:43 Iceberg Optimize: Scaling Problems and Parallelism

Eden discussed two scaling problems when running OPTIMIZE on large Iceberg tables due to a single-writer bottleneck. PR #26632 was mentioned as a proposed solution to use round-robin file reading to allow more than one writer for parallelism between partitions. Issue #27020 discusses scaling up the number of writers per partition based on data size

12:21 Project Updates: Releases and UI Refactoring\

Releases: The 477 release was discussed; the 478 release is being prepared with plans to publish more binaries. Preview UI: The Trino Query UI is being refactored into a reusable React component for use in the Preview UI.

16:51 Trino Gateway: Resource Groups and Queuing

Vishal proposed a discussion on implementing rate limiting/resource groups in Trino Gateway. The consensus was that queuing should be done at the gateway layer to avoid incorrect routing decisions, with existing implementations from Starburst and others cited as precedent

24:06 Client Drivers: OAuth, New APIs, and Spooling

dprophet presented a detailed proposal for modernizing OAuth support in the Python, JDBC, and ODBC drivers. The plan includes support for Client Credentials (M2M), Device Grant, and Protected Resource Metadata for autodiscovery.

38:28 Wrap Up

Manfred concluded the call to begin reviewing the Trino 478 release notes.

Trino Contributor Call, 24 Sep 2025

📹 Video recording on YouTube with timestamps

Attendees

mosabua, dprophet, dain,
wendigo, RoeyoOgen, pettyjamesm, martint, vagaerg, sajjoseph, raunaqmorarka

Topics, notes and action items

Planned topic

Trino Gateway 16
trino-odbc
trino-opa-tools
Status of 477
Java 25 adoption

0:00 Welcome and Introductions

Manfred kicks off the call.

0:33 Trino Gateway 16 release

Manfred announced the release of Trino Gateway 16, noting that it still uses the old release process and that automation efforts are ongoing.

2:15 Trino Release Notes (477/478)

A major discussion on the difficulty of getting release notes ready for Trino releases (e.g., 477/478). Maintainers are urged to help clarify content in the release notes PR to avoid release delays. The community is still looking for ways to improve this process.

08:42 Trino ODBC driver project

The Trino ODBC driver project, donated by Riley, is now running. Eric is involved in development, which currently focuses on addressing issues like buffer overflows and improving error handling/mapping to Trino natives.

13:10 Client-side caching and security in ODBC

Eric advocated for integrating DuckDB-like client-side caching into the ODBC driver to improve performance, especially for BI tools that "waterboard" the backend with recursive lookups and metadata queries. This led to a discussion with Dain and Martin about the complexities of building cache keys that account for access control (e.g., OPA policies).

19:58 Trino OPA Tools Repository

Manfred and Pablo introduced the new Trino OPA Tools repository as a playground for tooling around OPA integration, including a policy to help users migrate away from Ranger.

22:09 Trino-to-Trino connector

Update Saj and Eric provided an update on the Trino-to-Trino connector, noting that the new spooling protocol has led to very quick results, making it harder for users to realize they are running bad/expensive queries. An idea for a policy to kill bad queries based on spool size was discussed.

26:06 Trino-to-Trino and cloud connectivity

James joined, and the discussion shifted to using the Trino-to-Trino connector to link cloud services like Athena down to on-premises data centers, particularly for environments with strict legal constraints (like GDPR).

28:13 Iceberg maintenance and OPTIMIZE issues

Roy raised an issue where the OPTIMIZE command on large Iceberg partitions can lead to stuck writer tasks because the split generator initially focuses on one partition, blocking the pipeline. Potential solutions discussed included applying table partitioning to the scan part of OPTIMIZE and the possibility of reintroducing scaled writers for optimize.

46:04 Final plea for release notes quality

Martin made a final call for all maintainers to improve the quality of release notes, emphasizing the need to focus on user impact and adopting a non-technical tone. Ronak suggested creating more detailed guidelines to help maintainers meet the high standard.

Trino Contributor Call, 27 Aug 2025

📹 Video recording on YouTube with timestamps

Attendees

mosabua, xkrogen, dprophet, dain,
OmerRaifler wendigo, dain,
vishalya, Praveen2112, losipiuk

Topics, notes and action items

0:00 Trino 477 release ETA and discussion
3:25 Trino Gateway 16 release
4:40 Iceberg v3 support
5:35 Trino history server development
10:00 Trino ODBC driver project
13:30 Cluster stats endpoint with all worker data for Trino Gateway
28:27 Java 25 requirement coming in September
31:10 CatalogHandle removal
31:45 ODBC driver release plans
36:20 End

Trino Contributor Call, 23 Jul 2025

📹 Video recording on YouTube with timestamps

Attendees

mosabua, xkrogen, dprophet, dain,
RoeyoOgen, nineinchnick, Omar Raifler, Others - see video

Topics, notes and action items

0:00 Introduction and release update
6:36 Teradata connector
24:46 Trino history server
- https://github.com/trinodb/trino/issues/26238
40:45 Client protocol security and slug
44:45 Trino views and interoperability with other engines

Trino Contributor Call, 27 Jun 2025

📹 Video recording on YouTube with timestamps

Attendees

mosabua, pettyjamesm, xkrogen, dprophet, martint, wendigo, dain,
RoeyoOgen, vishalya, Carlos Benavides

Topics, notes and action items

0:00 Query history server discussion
11:31 Performance testing, performance regression testing, and benchto considered legacy
17:18 Java 24
21:02 Trino Software Foundation info and support, and plan for component/plugin maintainers
26:40 Trino packaging next steps - aim for smaller tarball and container
30:52 OPA, rego, and more tooling and improvements
33:52 Subquery caching and other performance work
38:30 Lakehouse connector and metastores
41:44 Access control on storage layer, grant/revoke, and passthrough