Big Data Service 3.4.0 new features and updates

Oracle Big Data Service 3.4.0 includes updates across security, certificate management, lifecycle operations, Java runtime support, and cluster recovery.

Big Data Service

• Hive metadata export for cluster migrations: Added an export script for more reliable migration of Hive metadata from the source cluster to the target cluster. The script is available through Oracle Support on request.
• Work request logs: When a work request fails, the log tab in the work-request console now shows the specific log message instead of a generic operation failed message, with sensitive information being masked.

  Example of DP payload:

  14:54:56.847 [main] INFO c.oracle.adw.bdcs.rest.BDSRestClient – {"id":877221196,"operationType":"ADD_WORKER_NODES","payload":{"ambariURL":{},"customerVisibleLogs":["Failed to add all nodes, please check Ambari"],"healthChecks":{}},"rollbackState":true,"status":"SUCCEEDED","timeFinished":"2025-06-18 14:54:40","timeStarted":"2025-06-18 14:22:19"}

  For more information about logs in a work request, see Listing Work Request Logs for a Cluster.

• OCI Certificates Service integration: OCI Certificates Service is supported for managing cluster SSL certificates. The associated SSL, truststore, and LDAP certificate settings can also be managed through Ambari. See Enabling/Renewing SSL for Self-signed Certificates.
• Generic shape support: Added support for a subset of VM.Standard.Generic and VM.DenseIO.Generic configurations in create, add node, change shape, and autoscale operations. See Integrated OCI Services.

ODH

• Ambari upgrade: Upgraded the cluster management framework to Ambari 3.

• Multi-JDK support: In BDS 3.4.0, JDK 11 is the default for Hive (Tez), Flink, Flume, Kafka, Spark 3, and HBase, while JDK 21 is the default for Trino. Added support for selecting JDK versions through Ambari for Hive (Tez), Flink, Flume, Kafka, Spark 3, and HBase. In earlier releases, JDK 8 remains the default. For details, see Big Data Service Release and Update Versions.

  Note
  
  Java 8 is planned for deprecation in a future release, subject to component support for later Java versions.
• Failed node replacement: Added support for replacement of terminated master and utility nodes using available node backups. This release doesn't include support for worker node replacement. See Replacing a node.
• Search and sort in cluster list: Added search and sort options for cluster name, state, version, and time, making clusters easier to find and manage. See Listing Clusters in a Compartment.
• Usersync activation: The Activate Usersync action is now available for active clusters. See Identity Domain Integration.