Distribution-specific notes

Each supported Hadoop distribution makes different choices in terms of packaging, versions of the different components of the Hadoop stack, supported ecosystems.

Each distribution bundles its own libraries and backports specific bugs that can modify the behavior of the Hadoop ecosystem components.

Therefore, there are some specificities related to the support of each Hadoop distribution

• Cloudera CDH
  • Security
    • DSS regular security and Sentry
  • Scala notebook
  • S3 datasets and Spark 2
• Hortonworks HDP
  • Security
    • DSS regular security and Ranger
    • DSS multi-user-security and Ranger
• MapR
  • MEP support
  • Security
  • Others
• Amazon Elastic MapReduce
  • Supported versions
  • Security
  • Deployment scenarios
    • Let DSS dynamically manage one or several EMR clusters
    • Connect DSS to an existing EMR cluster
      • DSS running on one of the cluster nodes
      • DSS outside of the cluster
    • Connect DSS to multiple existing EMR clusters
  • Using EMRFS
• Microsoft Azure HDInsight
  • Tested versions
  • Security
  • Connecting Dataiku DSS to Azure HDInsight
    • DSS running on an edge node managed by HDInsight
      • One-click deployment
      • Using an Azure Resource Management (ARM) template
      • Manual DSS installation
    • DSS running outside of HDInsight
  • Using Dataiku DSS on Azure HDInsight
    • Accessing Dataiku DSS on managed HDInsight edge nodes
    • Operating Dataiku DSS when using managed edge nodes
    • Interacting with cluster storage
• Google Cloud Dataproc
  • Security
  • Known limitations
  • Connecting DSS to Cloud Dataproc
    • DSS running on one of the cluster nodes
    • DSS outside of the cluster