Big Data Service 3.1.1 has added additional features

To view the ODH version, JDK version, OS version, and Linux version for this release, see Big Data Service Release and Patch Versions.

Big Data Service 3.1.1 has the following new features and updates:

• You can now be automatically notified of available Big Data Service updates by subscribing to automated update announcements. You can also create announcement subscriptions and filters.
• You can now create a Capacity Report using the Big Data Service BDS_CAPACITY_REPORT_CREATE API command and parameters.
• You can now manage updates for Big Data Service from the Cluster details page.
• Added new parameter (isDryRun) in the OS patch API.
• Upgraded Spark JDBC driver to version 2.6.18.2058.
• Avro is upgraded for Sqoop, Hadoop, HBase, Hive, Oozie, Livy, Flume and Spark components to 1.11.4.
• org.apache.hadoop.log.metrics.EventCounter is removed in Hadoop 3.4.1, therefore it has been removed from hdfs-log4j .

• Zookeeper 3.8.4 has moved from log4j to logback.

  Updated Zookeeper-logback config section from Ambari, to change Zookeeper Server log configurations.

  Updated Zkcli-logback config section from Ambari, to change Zookeeper Client log configurations.

• Introduced a new security setting, security.server.api.validate.readonly.update property to Ambari. This property prevents updates to read-only environment template configurations (for example, *-env) by using the API.
• Environment template configurations (for example, *-env) are no longer displayed on the Configurations page of the following components in Ambari. They have been replaced by new dedicated configuration properties to manage service environments.
  • Ambari Metrics
  • Flink
  • Flume
  • HBase
  • Hive
  • Kafka
  • MapReduce
  • Oozie
  • Ranger
  • Spark
  • Sqoop
  • Tez
  • YARN
  • ZooKeeper

Bug Fixes:

• For Big Data Service Versions 3.1.1 and later, error and output logs can be viewed at /bdslogs/oracle/bds/bootstrap/script_<work_request_id>.log.
• Upgraded the OCI HDFS Connector (oci-hdfs-connector) to version 3.4.1.0.0.1.
• Updated bootstrap error and output log file locations.
• The authentication order for SSH connections between cluster nodes has been updated. Previously, Kerberos authentication was attempted first, followed by public key authentication. With this update, public key authentication is now prioritized.

  The PreferredAuthentications publickey,gssapi-keyex,Pisa-with-mic SSH configuration is now set to enforce the new order. This ensures that SSH first attempts public key authentication, followed by Kerberos-based mechanisms if necessary.

  Note
  
  If you manually created SSH configuration files at /home/opc/.ssh/config or /root/.ssh/config, these files aren't modified as part of this update. Custom configurations remain intact.
• Python library patches are provided as part of the OS patch updates.
• Updated removing a node feature to include Force delete when node is gracefully decommissioned and deleted from Ambari.
• Hue HiveServer2 Failover Support (Big Data Service 3.1.1 / ODH 2.1.1).
• If you're running Spark jobs, you must apply the ARU Patch and relaunch Spark applications before applying ODH 2.2.0 patch.