Release 474-e LTS (2 Jun 2025)

Dell Data Analytics Engine, powered by Starburst Enterprise platform (SEP) 474-e LTS is the follow up SEP release to the 468-e LTS release.

It contains all improvements from Starburst Enterprise releases since the 468-e LTS release

It includes all improvements from the following Trino releases:

• Trino 469

• Trino 470

• Trino 471

• Trino 472

• Trino 473

• Trino 474

Highlights since 468-e

• As of May 2025, all supported SEP LTS versions run on JDK24.

• Added support for AI functionality, including AI SQL functions, AI Model Management, and AI Access Control as a private preview.

• Added support for Data maintenance jobs as a public preview.

• Added support for automatic refresh of Iceberg materialized views as a public preview.

Breaking changes

• The minimum required Java runtime for the JDBC driver and CLI has been updated to Java 11. If you are using an earlier version, you must upgrade to Java 11 or later.

• The optimize_hash_generation session property and the optimizer.optimize-hash-generation configuration option have been removed. You must remove these properties from your configuration or the cluster fails to start.

• The following Unity catalog metastore configuration properties have been renamed in the Delta Lake connector:

  • delta.metastore.unity.host to hive.metastore.unity.host

  • delta.metastore.unity.catalog-name to hive.metastore.unity.catalog-name

  • delta.metastore.unity.access-token to hive.metastore.unity.token

  Update these properties in your catalog configuration file or the cluster fails to start.

• The Kinesis connector has been removed. Update your dependencies and configurations to use an alternative connector.

• The following HTTP client property prefixes have been renamed:

  • workerInfo to worker-info

  • memoryManager to memory-manager

  If you use these properties in your configuration files, the old names are no longer recognized. Update your configurations to use the new names to avoid unexpected behavior.

• The following Hive metastore HTTP client properties have been removed:

  • hive.metastore.http.client.additional-headers

  • hive.metastore.http.client.authentication.type

  • hive.metastore.http.client.bearer-token

  • hive.metastore.http.client.read-timeout

  If your current configuration includes any of these properties, update your configuration to remove them. If left in place, you may experience connection and authentication issues when interacting with the Hive metastore.

• Effective May 23, 2025, AWS Glue no longer allows modifying the table_type parameter of an existing table. Any attempt to update an existing table’s table_type, such as migration procedures, results in a validation error from AWS Glue. If you wish to change the table_type parameter, you must use a CTAS operation to create a new table with the desired table_type. Rename the new table to match the original table name and adjust dependent resources as needed.

• The Phoenix connector has been removed due to incompatibility with Java 24. If your setup uses this connector, you must update your configuration to use an alternative supported connector.

• The following dynamic filtering properties have been removed:

  • dynamic-filtering.large-partitioned.range-row-limit-per-driver

  • dynamic-filtering.large.range-row-limit-per-driver

  • dynamic-filtering.small-partitioned.range-row-limit-per-driver

  • dynamic-filtering.small.range-row-limit-per-driver

  If your current configuration includes any of these properties, update your configuration to remove them.

• The Custom Resource Definitions (CRDs) for the External Secrets Operator included in the SEP Helm Chart have been updated from v1beta1 to v1. Upgrade your External Secrets Operator to v0.17, which fully supports v1.

474-e initial changes

General

• Added public preview support for data products in Iceberg format.

• Added the username of the logged-in user to the profile menu in the SEP UI.

• Added support for Apache Ozone’s ofs:// protocol.

• Improved performance and stability when executing queries containing large CASE WHEN expressions.

• Increased execution speed for queries with complex inner join expressions.

• Fixed an issue which prevented the query details page from loading for queries executed on older SEP versions.

• Fixed an issue which caused access denied errors when executing queries from roles with appropriate permissions.

Security

• Added support for Cross-site Request Forgery protection when using password, OAuth2, and SAML authentication mechanisms.

• Added support for Keycloak group provider.

• Fixed an issue where user impersonation with DBeaver or Apache Superset did not behave as intended due to incorrect role handling.

ClickHouse connector

• Added support for password authentication.

• Added support for password credential passthrough.

Elasticsearch connector

• Added support for Kerberos authentication using a credential cache.

MongoDB connector

• Added support for Kerberos authentication using a credential cache.

PostgreSQL connector

• Added support for reading the pgvector data type VECTOR, as well as a number of functions for similarity search.

474-e.1 changes (2 Jun 2025)

• Updated Go to 1.23.9 to mitigate CVE-2025-22871.

• Fixed an issue which caused incremental refresh of materialized views to fail when using session credentials.

• Fixed an issue where some menu items were visible to users without the required permissions.

• Fixed incorrect results when using window functions with DISTINCT.

• Fixed potential failures or incorrect results when querying partitioned tables using the OpenX JSON SerDe.

• Fixed incorrect results when reading timestamp values with leading or trailing spaces using the Regex and OpenX JSON table deserializers.

• Fixed potential performance regression when reading ORC data.

• Fixed incorrect memory usage calculation for truncated tables.

474-e.2 changes (18 Jun 2025)

• Fixed failure when querying views without StorageDescriptor on Glue.

• Fixed an issue where statistics computation was skipped for all columns if some column types did not support statistics.

• Fixed query failures with HIVE_CANNOT_OPEN_SPLIT error when reading ORC files with large row counts.

• Fixed failures when deleting rows from a table.

• Fixed an issue where some UI features were incorrectly made visible for certain server configurations.

474-e.3 changes (2 Jul 2025)

• Introduced resource leakage detection in the runtime for Native S3, Azure FS, and GCS.

• Fixed a bug in schema discovery which led to unexpected errors when discovering partitions with special characters in their values if schema-discovery.identifier-constraint=VALID_IN_HIVE_AND_TRINO was set.

• Fixed an issue where the Built-in access control audit log queue would crash or become unresponsive due to errors. The audit log queue now handles errors more gracefully by persisting queued events and logging more meaningful exceptions to the coordinator logs.

• Fixed an access control issue affecting Spark Iceberg write operations when using object storage governance.

474-e.4 changes (15 Jul 2025)

• Added support for cross-region data retrieval when using the S3 native filesystem.

• Improved performance when adding new partitions to Glue by skipping the update of partition statistics for empty partitions.

• Fixed query failures when the session catalog or schema names provided by clients contained capital letters.

474-e.5 changes (5 Aug 2025)

• Fixed failure in optimize_manifests caused by NULL value in first partition.

• Fixed loading of deletion vectors (DVs) for partitions with special characters in Delta Lake.