Posted On: Oct 14, 2021
Amazon EMR 6.4 release version now supports Apache Spark 3.1.2 and provides runtime improvements with Amazon EMR Runtime for Apache Spark. Amazon EMR 6.4 provides Presto runtime improvements for PrestoDB 0.254, and runtime improvements for Apache Hive 3.1.2 when you use AWS Glue Data Catalog for your metastore.
Amazon EMR 6.4 supports Apache Hudi 0.8.0, Trino 359, PrestoDB 0.254, Apache HBase 2.4.4, Apache Phoenix 5.1.2, Apache Flink 1.13.1, Apache Livy 0.7.1, JupyterHub 1.4.1, Apache Zookeeper 3.5.7 and Apache MXNet 1.8.0. Please see our release guide to learn more.
Starting Amazon EMR release versions 5.30 and 6.1 and later, you can now automatically terminate idle Amazon EMR clusters. This helps you to minimize costs without monitoring cluster activity. To get started, read our documentation here.
Amazon EMR Studio now supports multiple languages in the same Jupyter-based notebook forSpark workloads. Please see our documentation to learn more. You can now authenticate AmazonEMR Studio users using IAM-based authentication or IAM Federation, in addition to AWS SingleSign-On. You can learn more here.
Amazon EMR 6.4 includes Hudi 0.8.0, which allows you to use multiple applications to concurrently write to the same Hudi table. You can find more details on Hudi 0.8.0 features here. You can now report Hudi Metrics to Amazon CloudWatch, and set Hudi configurations at the cluster level using EMR Configurations API and the Reconfiguration feature.
Amazon EMR 6.4 now supports Spark SQL to write to and update Apache Hive metadata tableson Apache Ranger enabled Amazon EMR clusters. Please see our documentation to learn more.
Starting Amazon EMR release versions 5.7 and later, you can now create clusters with multiple custom Amazon Machine Images (AMIs). You can include both AWS Graviton and non-AWS Graviton instances in the same cluster. For more information, please read our documentation. Amazon EMR 6.4 is generally available in all regions where Amazon EMR is available. Please see Regional Availability of Amazon EMR and our release notes for more details.