Release 479-e STS (13 Feb 2026)#

Starburst Enterprise platform (SEP) 479-e is the follow up SEP release to 478-e. It includes all improvements from the following Trino releases:

This release is a short term support (STS) release.

Breaking changes#

  • Starburst data catalog only uses the catalog.default.location.uri as the database location when the database’s catalogid matches the value you configure in catalog.default.catalog.id. If your catalogid does not match catalog.default.catalog.id, you must either specify a locationUri when creating databases or specify a location for each table explicitly, or queries against those tables fail.

  • As of SEP 479-e, integration with Apache Ranger Admin 2.0.0 is no longer supported. You must upgrade your Ranger Admin deployment to version 2.1.0 or higher to continue using Ranger-based authorization.

  • Starburst Warp Speed now requires local SSD storage on the coordinator in addition to the existing local SSDs on worker nodes. You must provision local SSD storage on the coordinator before upgrading.

  • The following Starburst Warp Speed features are removed:

    • Fast warmup

    • Warmup rules

    • Index and cache resiliency

    Associated properties for configuring these features are removed:

    • warp-speed.config.store.type

    • warp-speed.store.path

    • warp-speed.objectstore.*

    • warp-speed.local-store.path

    • warp-speed.enable.import-export

    • warp-speed.use-http-server-port

    • warp-speed.call-home.*

    You must remove these properties from your configuration or the cluster fails to start.

  • The following Delta Lake Unity Catalog managed table configuration properties have been renamed:

    • hive.metastore.unity.catalog-owned-table-enabled to hive.metastore.unity.catalog-managed-table-enabled

    • delta.feature.catalogOwned-preview to delta.feature.catalogManaged

    Update existing configurations to use the new property names.

  • The query editor Run and troubleshoot configuration property troubleshooting.anonymized-query-plan has been removed. Remove this property from your configuration before upgrading.

General#

  • Upgraded all SEP versions to run on Java 25.0.2.

  • Added public preview support for sharing data products across clusters.

  • Added support for configuring the connection pool connection timeout for JDBC connectors using the connection-pool.connection-timeout property.

  • Added support for using secrets with Starburst Portal.

  • Added local file system support for storage functions.

  • Added the ability to republish data products when source table schemas change.

  • Added support for cron expressions when scheduling materialized view refreshes in data products.

  • Added CTE reuse as a public preview feature.

  • Added support for custom startup script initFile and HTTPS port selection in the Starburst cache service Helm chart.

  • Added a new gateway.load-balancing.strategy configuration toggle to Starburst Gateway. The toggle controls the load balancing method within a single routing group. It can be set to RANDOM or QUERY_COUNT_BASED.

  • Updated materialized views queries to skip reanalysis and access control checks on source tables when run_as_invoker is disabled.

  • Extended the alphanumeric character limit from 40 to 100 for data product titles and data product schema names.

  • Improved responsiveness and performance when navigating the access control UI.

  • Improved error handling for data product refresh failures. The system now correctly displays errors when refresh fails to trigger or when materialization fails.

  • Starburst Warp Speed catalogs now use the file system cache for data caching.

  • Fixed an issue where the Glue GetDatabases endpoint returned incorrect target database information in the Starburst data catalog metastore.

  • Fixed an issue where data maintenance jobs failed when the fault-tolerant execution retry policy was set to TASK.

  • Fixed an issue where the Refresh button did not trigger data product materialization refresh on server.

  • Fixed a performance issue in the data product dashboard where the catalog dropdown was temporarily unavailable due to excessive API calls.

  • Fixed an issue where query result caching was incorrectly applied when viewing or editing data products.

  • Fixed date_add function failure when adding large values to dates or timestamps.

  • Fixed a race condition in Starburst Data Catalog where concurrent partition updates and statistics updates could overwrite each other, resulting in missing partition statistics and degraded query performance.

Security#

  • Update KEDA Trino Scaler to 0.1.23 to mitigate the following CVEs:

    • CVE-2025-49796

    • CVE-2025-49794

    • CVE-2025-6965

  • Added REST API support for delegating column masking and row filtering expression management. System administrators can now grant CREATE, ALTER, DROP, and SHOW permissions for COLUMN_MASK_EXPRESSIONS and ROW_FILTER_EXPRESSIONS entity categories to other roles.

  • Added REST API support for delegating permission to apply column mask and row filter expressions to tables. System administrators can grant apply permissions with or without grant option to other roles.

  • Added support for removing all associated privileges from a role when you delete the role.

  • Added http-server.authentication.oauth2.oidc.use-principal-from-id-token configuration property to extract the user principal from the OIDC ID token instead of the access token.

BigQuery connector#

  • Added support for configuring write stream types with the bigquery.write-stream-type property.

Db2 connector#

  • Added support for TIMESTAMP(p) data type at a maximum precision of 12 fractional seconds.

Delta Lake connector#

  • Added support for specifying the time zone used when reading from Delta Lake tables using the delta.time-zone configuration property.

Elasticsearch connector#

  • Fixed an issue where wildcard queries against Elasticsearch indices with mapping conflicts would fail even when the conflicting fields were not referenced. Queries now succeed if the conflicting fields are not used in query projections or predicates.

HBase connector#

  • Fixed an issue where connector initialization failed when the user did not have table creation permissions and the native table store table already existed.

Hive connector#

  • Fixed a user impersonation failure affecting partitioned join operations.

Iceberg connector#

  • Manual and automatic refresh for materialized views are now generally available.

  • Manual refreshes of Iceberg data products now update the dataset status and last refresh timestamp.

  • Reduced memory pressure when querying tables that contain highly nested fields.

MaxCompute connector#

  • Added support for configuring connection and read timeouts.

Redshift connector#

  • Added support for improving performance of reads from Redshift tables using the UNLOAD command.

SingleStore connector#

  • Added support for LIKE predicate pushdown on VARCHAR types when the singlestore.experimental.enable-string-pushdown-with-binary configuration property is enabled.

  • Added support for complex predicate pushdown on VARCHAR and CHAR types.

  • Added support for aggregate pushdown on VARCHAR and CHAR columns for the following functions: count, count(distinct x), min, and max.

  • Added support for NOT predicate pushdown.

  • Fixed an issue where multiple = predicates on the same column or IN predicates were treated as case-insensitive.

Snowflake connector#

SQL Server connector#

  • Fixed an issue where schemas were not being returned when sqlserver.database-prefix-for-schema.enabled was enabled.

Starburst Stargate connector#