Release 476-e LTS (2 Sep 2025)#

Dell Data Analytics Engine, powered by Starburst Enterprise platform (SEP) 476-e LTS is the follow up SEP release to the 474-e LTS release.

It contains all improvements from Starburst Enterprise releases since the 474-e LTS release.

It includes all improvements from the following Trino release:

Highlights since 474-e#

  • Added support for data product enrichment using the Starburst AI agent as a private preview.

  • Added support for Apache Iceberg table spec version 3.

  • Added support for Iceberg branching.

  • Updated the look and feel of the data products UI.

  • Introduced a new UX to make it easier to view and manage built-in access control roles and policies.

  • Starburst ODBC driver V3 series is now generally available.

Breaking changes#

  • The hive.s3.storage-class-filter configuration property has been renamed to hive.s3-glacier-filter in the Hive connector. If you use this property in your configuration files, the old names are no longer recognized. Update your configurations to use the new names to avoid unexpected behavior.

  • The query.max-written-data-size query management property has been renamed to query.max-write-physical-size. Update your configuration files to use the new property name or the cluster fails to start.

  • The --sun-misc-unsafe-memory-access=allow JVM option is required when running Trino with the BigQuery or Snowflake connectors. Add this option to your Trino JVM configuration to ensure proper connector operation.

  • The kafka-event-listener.client-config-overrides configuration property has been removed. To configure the Kafka client for the event listener, specify the configuration in a separate file and set the kafka-event-listener.config.resources property to the path of that file.

  • The Kudu catalog configuration property kudu.schema-emulation.enabled has been removed and replaced with the new property kudu.schema-emulation.type. If your configuration relied on the old property, you must update your catalog configuration to use kudu.schema-emulation.type. See Kudu configuration properties for more information about allowed values for kudu.schema-emulation.type.

  • The hive.file-status-cache-tables.excluded configuration property has been renamed to hive.file-status-cache.excluded-tables. Update your configuration files to use the new property name or the cluster fails to start.

476-e initial changes#

General#

  • Updated the default value of the max_import_duration property for materialized views from 30m to 120m.

  • Updated the default refresh interval for data product materialized views from 60m to 125m.

  • Improved execution speed for queries that include joins and aggregations.

Security#

  • Added support for allowing the role receiving a location privilege grant to grant the privilege to others in built-in access control.

  • Updated built-in access control to include ownership information when querying the schemas_authorization, tables_authorization, and functions_authorization metadata tables in the system.metadata schema.

Delta Lake#

  • Fixed an issue that prevented querying Delta Lake tables with timestamp_ntz columns in Unity Catalog.

Hive connector#

  • Added support for view refresh using the ALTER VIEW ... REFRESH statement.

  • Fixed an issue where SELECT * queries failed when AWS Lake Formation access control was enabled, the user lacked SELECT permissions on all columns, and hide_inaccessible_columns=true.

  • Fixed incorrect DATE values when reading Parquet files written by Hive in legacy calendar mode for dates prior to 1970-01-01. SEP now detects and converts these legacy calendar dates to proleptic Gregorian calendar.

Iceberg connector#

  • Added support for branching.

  • Added support for Apache Iceberg table spec version 3.

  • Added support for copy-on-write mode when using the merge_mode table property.

  • Added support for AWS Glue v2 to Iceberg. You can enable it using the iceberg.catalog.type=glue_v2 configuration property.

  • Added support for specifying the time zone to apply when reading from Iceberg tables with the iceberg.time-zone configuration property.

  • Improved query planning performance for materialized views by parallelizing base table freshness retrieval.

  • Fixed an issue where data maintenance jobs could target non-Iceberg catalogs.

Kudu connector#

  • Added support for schema emulation when Kudu is integrated with the Hive metastore. See Kudu configuration properties for information on configuring schema emulation.

Memory connector#

  • Added support for view refresh using the ALTER VIEW ... REFRESH statement.

PostgreSQL connector#

Snowflake connector#

  • Added support for the MERGE SQL operation.

476-e.1 changes (2 Sep 2025)#

  • Updated starburst-schema-registry to version 0.9.1-e.8 to mitigate the following CVEs:

    • CVE-2024-22257

    • CVE-2024-22259

    • CVE-2025-22228

  • Limited access to table branch operations (CREATE BRANCH, ALTER BRANCH, and DROP BRANCH) to sysadmin users with BIAC enabled.

  • Fixed an issue where navigating to query overview from a workload management table caused filter controls to become unresponsive.

  • Fixed an issue that prevented the SEP product version from rendering in the About dialog.

476-e.2 changes (12 Sep 2025)#

  • Added support for reading legacy dates and timestamps written by Hive in Orc and Parquet.

  • Added support for generating multiline SQL queries for all datasets in a data product when using the Query button.

  • Fixed an issue that caused failures when running more complex Python-based UDFs.

  • Fixed query failures on nested fields in mixed-type MongoDB arrays.

  • Fixed failure when reading deletion vectors generated by Databricks in Iceberg connector.

476-e.3 was skipped.

476-e.4 changes (22 Sep 2025)#

  • Manual configuration of the starburst catalog using a starburst.properties file is no longer supported for configuring Starburst AI. You must remove any existing starburst.properties file to prevent startup failure. Configuration values previously placed in starburst.properties must now be included in config.properties. This change is not backward compatible.

  • Added support for VARIANT type in the Iceberg connector when using the Hive Thrift metastore.

  • Added support for Confluent’s JSON Schema Registry in the Kafka connector.

476-e.5 changes (29 Sep 2025)#

  • Updated jinjava to 2.8.1 to mitigate CVE-2025-59340.

  • Fixed an issue where invalid partition filter expressions were accepted. Expressions are now validated as SQL syntax.

  • Fixed authentication issues that prevented access to Iceberg Glue v1 catalogs.

  • Fixed an issue with the refresh interval default value in Data Product materialized views. refresh_interval must be greater than 120 minutes. authentication. You can now use registryCredentials to authenticate with private container registries.

  • Fixed an issue where selecting a catalog with a large number of schemas caused the BIAC page to become unresponsive.

476-e.6 changes (10 Oct 2025)#

  • Fixed an issue where data product deletion failures could occur without a visible error. Failures when deleting a data product now return a clear error message.

  • Fixed failure due to column count mismatch when executing the add_files_from_table procedure in Iceberg.

  • Fixed an issue where the cache service could consume excessive memory, potentially leading to out-of-memory errors.

  • Fixed an issue where missing securityContext caused pod creation to fail. You can now set worker.kedaScaler.securityContext, and it is applied to the keda-trino-scaler container in the coordinator pod.

476-e.7 changes (31 Oct 2025)#

  • Added the -XX:-RewriteBytecodes JVM flag to the default configuration. This change disables JVM bytecode rewriting by default to prevent potential heap corruption issues on ARM64 systems.

  • Added configurable S3 API timeouts and retry settings to reduce latency for interactive queries accessing Iceberg tables in S3.

  • Improved performance of data product search, significantly speeding up text-based filtering.

  • Improved MERGE performance for MarkDistinct in NOT MATCHED clauses.

  • Fixed an issue that prevented querying Delta Lake tables with timestamp_ntz columns in Unity Catalog.

  • Fixed an issue that prevented reading certain DATE and TIMESTAMP values when the internal byte representation was invalid when using the Oracle connector.

  • Fixed Azure filesystem access when proxy is used.

  • Fixed failure when aggregation functions were wrapped in GROUP BY AUTO.

  • Fixed incorrect results for queries with joins using ST_Contains(), ST_Intersects(), or ST_Distance() functions.

  • Fixed an issue where grants provided through groups were incorrectly ignored when using SET SESSION AUTHORIZATION.

  • Fixed incorrect results from geometry_to_bing_tiles() where generated tiles did not fully cover the geometry area. location (catalog.default.location.uri), the catalog now creates a subfolder for each schema (e.g., XXX/YYY) instead of placing all schemas directly under the default location.