Big Data Service 3.0.21 has added additional features

BDS 3.0.21 release has the following new features implemented:

• Schema Registry included in ODH, which provides a comprehensive way of storing and exchanging metadata information between its clients
• Includes Python script-based bootstrap helper functions to update Ambari configs
• Connecting Trino to BI tools
• Updating the dcli syntax option -C host file names
• Integration of Object Storage
• Enable Hive ACID property in BDS cluster

In addition, the following bugs/enhancements have been made:

• Fixed issue with failure in check for Flink when adding all services to ODH 1.0.
• Fixed issue with HMaster logs being empty (even if both HMasters stuck in standby state).
• Support Hue to work with NameNode HA (Configure Hadoop HttpFs by default).
• Fix to enable Spark dynamic allocation by default.
• Hue Metastore Manager and Hive Editor display tables which user do not have permission.
• Fixes to improve stability when creating Big Data Service clusters.
• Fix for avoiding principal names collision while creating Big Data Service clusters pointing to same Active Directory.
• Fix ranger db schema to allow large no of hosts in certificate common names.
• Ambari mpack SPARK3 service recommender improvements.
• The default Yarn application’s localized file directory configured to number of disks per node.
• Ability to submit Trino jobs through Oozie
• For executing Trino queries from JupyterHub, install Trino package in Python 3.6 or later.