Release 420-e STS (28 Jun 2023)#

Starburst Enterprise platform (SEP) 420-e is the follow up release to the 413-e version. It includes all improvements from the following Trino project releases:

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

Breaking changes#

Removed legacy Teradata query pass-through in favor of the query table function. The deprecated.teradata.query-pass-through.enabled property must be removed from the cluster configuration or the cluster does not start. Instead of legacy pass-through, use the query function.
The Trino S3 file system class was renamed from io.trino.plugin.hive.s3.TrinoS3FileSystem to io.trino.hdfs.s3.TrinoS3FileSystem. If you are using Ranger Audit to S3, update the property value in ranger-audit.xml.
On a cluster with built-in access control, the sysadmin role is no longer assumed automatically when querying views in DEFINER mode whose owner has the sysadmin role. For views that are no longer accessible due to this change, the privileges of the view owner must be adjusted to grant access to the underlying tables.

General#

Added the ability to sort columns in the query editor results table.
Improved error messages in the Starburst Enterprise web UI for empty pages on the Saved queries pane.
Improved tooltip text in the Starburst Enterprise web UI and added copy functionality to tooltip text in the query editor.
Added functionality to collapse the query results table in the query editor.
Updated telemetry to capture the values of select, non-sensitive configuration properties.
Replaced some link text with icons in the query editor.
Moved the SEP version, JDK, and environment information from the main screen to an About dialog accessed from the user menu.
Query logger has been renamed. It is now referred to as the SEP backend service. This change acknowledges the vast increase in scope of the functions it supports, which now include built-access control and the cache service. No changes to configuration are required at this time.

Data products#

Added the ability to open a new query editor tab from the Datasets and Usage examples sections of the Data product details screen.

Security#

Added SCIM user and group synchronization with Azure Active Directory as a public preview.
Added read and write support for views in INVOKER security mode when using AWS Lake Formation access control.
Added enforcement for the requisite privileges to execute ALTER … SET AUTHORIZATION to access control systems as follows:
- Built-in access control requires a privilege that grants ALTER execution to the target entity and the ability to grant the privilege to others.
- Global access control with Apache Ranger requires a privilege granting ALTER execution on the entire catalog.
- Hive and Delta Lake access control with Apache Ranger requires a privilege granting ALTER execution on all schemas.
Deprecated the starburst.audit.access-log.enabled and starburst.audit.access-log.queue-size built-in access control audit log configuration properties in favor of starburst.access-control.audit.access-log.enabled and starburst.access-control.audit.access-log.queue-size, respectively.
Updated security behavior to prevent access control policies from restricting access to the system.jdbc schema.
Fixed an issue that caused policies for functions and procedures in Ranger access control to incorrectly deny access.

Greenplum connector#

Added support for Managed statistics as a public preview.

IBM Db2 connector#

Added support for Managed statistics as a public preview.

Netezza connector#

Added support for Managed statistics as a public preview.

Redshift connector#

Added support for fault-tolerant execution.
Added support for predicate pushdown of complex expressions.

Synapse connector#

Added support for Managed statistics as a public preview.

Starburst Stargate#

Added support for Managed statistics as a public preview.
Removed the ability to set session properties on the remote cluster as a requirement for collecting statistics.