Releases: apache/kyuubi
Releases · apache/kyuubi
v1.7.3-rc0
[RELEASE] Bump 1.7.3
v1.7.2
Highlight
- Enhance JDBC Client and Beeline Client, like: Support running Python scripts and codes in Beeline, fix Http authentication exception message is empty, backport HIVE-19048: Initscript errors are ignored, backport HIVE-15820: comment at the head of beeline -e
- Enhance submission engine related code, like: support yarn-client and yarn-cluster for YarnApplicationOperation, different engine submit timeout config, destroy build process when waitCompletion is false and engine is running in cluster mode
- Fix many flaky tests and incorrect documentations
Release Notes
https://kyuubi.apache.org/release/1.7.2.html
Download Link
v1.7.2-rc0
[RELEASE] Bump 1.7.2
v1.7.1
Highlight
- Kyuubi requires Spark 3.1 and above
- Upgrade Hadoop from 3.3.4 to 3.3.5
- A series of improvements for Arrow based serialization
- A series of improvements for batch job of Spark on K8s/YARN
- Bug fix for 'beeline -e' with SQLs before 'source' statement
- Spark Authz plugin fixes privilege checks on source tables of the permanent view in Spark 3.1
Release Notes
https://kyuubi.apache.org/release/1.7.1.html
Download Link
v1.7.1-rc0
[RELEASE] Bump 1.7.1
v1.7.0
Highlight
- Fully tested w/ Spark 3.1, 3.2 and 3.3, initial support oncoming 3.4
- Fully tested w/ Flink 1.14, 1.15 and 1.16
- Introduce Trino frontend (experimental)
- Introduce Spark Lineage plugin (experimental)
- Introduce Spark Hive connector (experimental)
- Spark engine supports arrow-based results serialization
- Spark engine supports executing Python/PySpark scripts(experimental)
- Several Kyuubi Helm Chart & Spark on K8s improvements
- REST and Batch API functions, clients and stability enhancements
- Major Spark Authz plugin improvements of including Apache Iceberg support
- Security enhancements including LDAP authentication, internal encrypted transports, etc.
Release Notes
https://kyuubi.apache.org/release/1.7.0.html
Download Link
v1.7.0-rc1
[RELEASE] Bump 1.7.0
v1.7.0-rc0
v1.6.1-incubating
Highlight
Server
- Fast fail the kyuubi connection if the engine application has been terminated
- Transfer the TGetInfoReq to kyuubi engine side to check the connection valid and keep connection alive
- GetInfo supports returning server/engine information
- Handle zookeeper watch events
Client
- Implement isSigned method of ResultSetMetaData for Kyuubi JDBC driver
- Extend JDBC URL to support catalog
Engine
- [SPARK] Fix scala NPE issue when adding non-local jar URI to class loader
- [SPARK] Kyuubi should set env SPARK_USER_NAME for K8s deployment
- [SPARK] Spark engines running on K8s give preference to IP for service discovery registration
- [FLINK] Flink SQL engine fixes GetTables operation
- [FLINK] Flink SQL engine supports getColumns operation
- [FLINK] Flink SQL engine supports run DDL across versions
Release Notes
https://kyuubi.apache.org/release/1.6.1-incubating.html
Download Link
v1.6.0-incubating
Highlight
Server
- Support Batch(Jar) Job submission
- Support Thrift HTTP mode
- Support Etcd as Discovery Service
- Support JDBC authentication mechanism
- Unify the authentication for Thrift and REST v1 API
- Enhance metrics system
Client
- Introduce REST client for Batch(Jar) Job submission
- Decouple Hive and Hadoop dependencies from builtin JDBC driver
- Builtin JDBC driver supports using keytab for Kerberos authentication
- Enhance beeline to support Spark console progress bar
Engine
- Fully verified with the latest Spark 3.1/3.2/3.3
- Fully verified with the latest Flink 1.14/1.15
- Fully verified with the Trino 363
- Support Hive backend Engine (Beta)
- Support JDBC backend Engine (Beta)
Plugin
- Introduce Spark TPC-DS connector
- Introduce Spark TPC-H connector
- Introduce Spark Authz (Apache Ranger) plugin
Release Notes
https://kyuubi.apache.org/release/1.6.0-incubating.html