What's New in RapidMiner Radoop 8.2.2?
Released: July 16th, 2018
The following improvements are part of RapidMiner Radoop 8.2.2.
- Extended Python 3.x version compatibility for Spark Script operator
- Added more tests to tabs on Connection Settings dialog: MapReduce test to Hadoop tab, Hive connection test to Hive tab
- Missing Hadoop settings dialog now may show up when clicking on Test or Full Test, not just for OK (e.g. for HDInsight, Impala)
- Service Principal fields are no longer mandatory for a connection, but there is still a warning for the missing ones
- Connection import methods and validation now ignores the case in true and false values when checking certain boolean Hadoop properties
- BUGFIX: Impersonation is taken into account when running a radoop process from Studio (previously only for connection test and Server)
- BUGFIX: Using Static, Manual Configuration in Spark Resource Allocation Policy field no longer leads to a validation error
- BUGFIX: NumberFormatException is no longer thrown when using formatted values (e.g. 2g) for spark.executor.memory setting in Advanced Hive Parameters
- BUGFIX: Explicitly specified value of spark.executor.memory setting in Advanced Hive Parameters no longer counted twice in connection pool size estimation if it is specified in bytes
- BUGFIX: In case of an error in Spark job test, subsequent tries with other Spark versions no longer fail with validation error (this affected connections with Dynamic Resource Allocation setting)
- BUGFIX: Fixed that subsequent tests on Connections Settings dialog did not update the UI (header, button state) properly
- BUGFIX: Hadoop username field is no longer hidden when Enable MapR Security is set, but a non-MapR Hadoop Version is selected