What's New in RapidMiner Radoop 9.4.0?
Released: September 25th, 2019
The following improvements are part of RapidMiner Radoop 9.4.0.
- A step forward towards enterprise security. Radoop Proxy is capable of authenticating against the latest RapidMiner Server running in Enterprise Single Sign On (SSO) mode. Customers can use their Enterprise SSO credentials to open their way through Radoop Proxy to reach a firewalled Hadoop cluster
- Radoop Proxy now uses the recently introduced repository-based connection definition framework which amongst having nicer UI, it also allows parameter injection and sharing connections (to ease onboarding of new users). Existing Radoop Proxy connections now can be easily converted to use the new connection framework without typing
- Ability to select Amazon Elastic MapReduce (EMR) 5.x as Hadoop version in Radoop Connection Editor. At the same time existing EMR 4.4+ connections were renamed to EMR 4.x and got deprecated (see below)
- Added a Quick Fix option to help the migration from a deprecated EMR 4.x Radoop Connection to EMR 5.x, allowing a safe and easy way to migrate existing Radoop Connections. Use this option if and only if the cluster is indeed having an EMR version 5.x and the Radoop Connection originally specified the Amazon Elastic MapReduce (EMR) 4.4+ (since there has been no option to select EMR 5.x previously)
- Updated tooltip on Custom SparkR archive path field
Update on support matrix
- To provide more focus on the latest development in the Hadoop ecosystem, Radoop started a deprecation process of providing connection capabilities to some specific Hadoop vendors and versions. Although no functionality was removed yet, Radoop now marks such Radoop Connections as DEPRECATED and provides RapidMiner Process level warning during Process Execution. We plan to remove associated functionality in the upcoming releases to accommodate some time for the migration. Please reach out to your RapidMiner Support immediately if your business would be affected by the removal of these features. Hadoop versions in scope: Apache Hadoop 2.x, IBM Open Platform, Amazon Elastic MapReduce (EMR) 4.x, Open Data Platform, MapR 5.x, MapR 6.x
- Similarly to Hadoop versions, a deprecation process was started on some older Spark versions as well. Spark versions in scope: Spark 1.4 or below, Spark 1.5