Release 332-e LTS (14 Apr 2020)#
Starburst Enterprise Presto 332-e is the follow up release to the 323-e version.
The 332-e release includes all improvements from the Presto project releases since Presto 323:
This release is a long term support (LTS) release.
In addition, numerous improvements and additions for enterprise features are added:
Apache Ranger as a service in both CFT and K8s
IBM Cloud Object Storage support in IBM Cloud Object Storage Support for the Hive Connector
Added support for parallel operations to the Oracle connector
Increased release cadence with long term support LTS and short term support STS releases
The internal cluster communication in earlier Starburst Enterprise Presto supports using Kerberos for authentication. This functionality has been removed in favor of automatically securing internal cluster communication. Find more details in Secure Internal Communication.
The Starburst BigQuery connector is replaced by the open source BigQuery connector contributed by the community. The new connector is more feature-rich and performant. Users of the old connector need to change their catalog files to adapt to the usage of the new connector and its configuration properties, documented in BigQuery Connector.
The usage of
groupsin resource groups as selector rule properties and used for catalog rules as part of the built-in system access controls is deprecated. Replace any usage with lists and regular expressions in the
groupfields. For example, change
users: ["abc", "def", "xyz"]to
The Starburst Secrets support using a Java keystore file containing any secrets is deprecated and replaced with the secrets support from Presto. More information can be found in Secrets.
The query audit feature changed to use a different configuration file and properties.
Configuration of the path to the license for Starburst Enterprise Presto changed to use a default path. As a result, the
license.pathproperty has been removed entirely.
Ranger-based audit as available in prior releases is included from 332-e.2 onwards only
Regular Kerberos support as as Kerberos and password authentication pass passthrough for the Teradata and Teradata Parallel Direct connectors it as available in prior releases is not included, but planned for a future patch version
The preferred runtime for Presto servers changes from Java 8 to Java 11. While using Java 11 is not a hard requirement at this stage, we strongly suggest to update to gain performance improvements and other benefits. Java 11 is planned as a hard requirement for the next release. More information can be found in in the Java requirements section.
Support Delta Lake connector on Azure
Fix PostgreSQL driver access in event logger
Fix initialization of Presto Ranger plugin
Add standalone usage-client to presto-server docker image
Add cumulative wall time to usage metrics
Add cluster statistics resource REST endpoint for Mission Control
Add support for Hive Ranger Audit
No longer start the
Add predicate pushdown for the Oracle
TIMESTAMPdata types in the Oracle connector.
Fix a bug in the Oracle connector, which in rare cases, could cause
INSERTqueries to fail write data to destination table.
Correct closing and configuring of connections for Oracle connection pools
Fix Teradata connector initialization
Remove requirement to Snowflake role when user impersonation is disabled
Add data integrity verification for Presto internal communication
Add Kerberos credentials and password passthrough for the Oracle connector
Suppress Hive views in the Delta Lake connector
Map Delta Lake timestamp to Presto timestamp with time zone
Backport cluster-internal communication
Support only shared secret usage
Remove support for Kerberos for internal communication
Cluster accessed via HTTPS must have
Deny user impersonation usage in catalogs when system level security with Ranger is used
Fail query, instead of silently skipping, if there are bucket files outside of the bucket range
Fix tracking of queued queries in JMX stats
Use Hive column indices for Parquet tuple domain creation when specified
Accept URIs with an equals sign
=in path in Sentry privilege configuration
Fix to prevent JVM crashes when LongArrayList is too large
Allow specifying behavior if data is inserted into existing Hive partition with the
Fix query failure when reading an ORC ACID table with a filter after the table underwent a minor table compaction.
Fix incorrect query results when reading an ORC ACID table that has deleted rows and underwent a minor compaction.
Fix rare failure when recording server stats in T-Digests.
Match the existing user and group of the table or partition when creating new files on HDFS.