Configure for HDInsight

When deployed to Microsoft Azure, the Designer Cloud Powered by Trifacta® platform must be integrated with Microsoft HDInsight, a Hadoop-based platform for data storage and analytics. This section describes the configuration steps required to integrate with a pre-existing HDI cluster.

Limitations

For this release, the following limitations apply:

• This release supports integration with HDI 3.5 and 3.6 only.

• HDI does not support the client-server web sockets configuration used by the platform. This limitation has the following impacts:

  • Diminished suggestions prompted by platform activities

     

Pre-requisites

This section makes the following assumptions:

1. You have installed and configured the Designer Cloud Powered by Trifacta platform onto an Edge node of a pre-existing HDI cluster.
2. You have installed WebWASB on the platform edge node.

Before You Begin

Create Alteryx user account on HDI cluster

The Designer Cloud Powered by Trifacta platform interacts with the cluster through a single system user account.  A user for the platform must be added to the cluster.

UserID:

If possible, please create the user ID ([hdi.user]) as: trifacta.

This user must be created on each data node in the cluster.

This user should belong to the group ([hdi.group]): trifactausers.

User requirements:

• (if integrating with WASB) Access to WASB
• Permission to run YARN jobs on the cluster. 

Steps:

1. You can apply this change through the Admin Settings Page (recommended) or trifacta-conf.json. For more information, see Platform Configuration Methods.

2. Set the cluster paths in the following locations:

   "hdfs.pathsConfig.fileUpload": "/trifacta/uploads",
   "hdfs.pathsConfig.dictionaries": "/trifacta/dictionaries",
   "hdfs.pathsConfig.batchResults": "/trifacta/queryResults",

   Do not use the trifacta/uploads directory. This directory is used for storing uploads and metadata, which may be used by multiple users. Manipulating files outside of the Designer Cloud application can destroy other users' data. Please use the tools provided through the interface for managing uploads to WASB.

   Individual users can configure the output directory where exported results are stored. See User Profile Page.

3. Save your changes.

Acquire cluster configuration files  

You must acquire the configuration files from the HDI cluster and apply them to the Alteryx node.

Steps:

1. On any edge node of the cluster, acquire the .XML files from the following directory:

   /etc/hadoop/conf

   NOTE: If you are integrating with an instance of Hive on the cluster, you must also acquire the Hive configuration file: /etc/hive/conf/hive-site.xml.

2. These files must be copied to the following directory on the Alteryx node:

   /trifacta/conf/hadoop-site

3. Replace any existing files with these files.

Acquire build build number

You must acquire the full version and build number of the underlying Hortonworks distribution. On any of the cluster nodes, navigate to /usr/hdp. The version and build number is referenced as a directory in this location, named in the following form:

A.B.C.D-X

For the rest of the configuration, the sample values for HDI 3.6 are referenced. Use the appropriate values for your distribution.

Configure the HDI Cluster

The following configuration sections must be reviewed and completed. 

Specify Storage Layer

In the Azure console, you must specify and retain the type of storage to apply. In the Storage tab of the cluster specification, the following storage layers are supported.

Specify Protocol

In the Ambari console, you must specify the communication protocol to use in the cluster. 

Steps:

1. In the Ambari console, please migrate to the following location: HDFS > Configs > Advanced > Advanced Core Site > fs.defaultFS.

2. Set the value according to the following table:
   

   Azure Storage Layer	Protocol (fs.defaultFS) value	Designer Cloud Powered by Trifacta platform config value	Link
   Azure Storage	wasbs://<containername>@<accountname>.blob.core.windows.net	"webapp.storageProtocol": "wasbs",	See Set Base Storage Layer.
   Data Lake Store	adl://home	"webapp.storageProtocol": "hdfs",	See Set Base Storage Layer.

3. Save your changes.

Define Script Action for domain-joined clusters

If you are integrating with a domain-joined cluster, you must specify a script action to set some permissions on cluster directories. 

For more information, see https://docs.microsoft.com/en-us/azure/hdinsight/domain-joined/apache-domain-joined-configure-using-azure-adds.

Steps:

1. In the Advanced Settings tab of the cluster specification, click Script actions.

2. In the textbox, insert the following URL:

   https://raw.githubusercontent.com/trifacta/azure-deploy/master/bin/set-key-permissions.sh

3. Save your changes.

Install Software

If you haven't done so already, you can install the Alteryx software into an edge node of the HDI cluster. For more information, see Installation Steps.

Configure the Platform

These changes must be applied after the  Designer Cloud Powered by Trifacta platformhas been installed.

You can apply this change through the Admin Settings Page (recommended) or trifacta-conf.json. For more information, see Platform Configuration Methods.

Specify Alteryx user

Set the Hadoop username for the Designer Cloud Powered by Trifacta platform to use for executing jobs [hadoop.user (default=trifacta)]:  

"hdfs.username": "[hadoop.user]",

Specify location of client distribution bundle JAR

The Designer Cloud Powered by Trifacta platform ships with client bundles supporting a number of major Hadoop distributions.  You must configure the jarfile for the distribution to use.  These distributions are stored in the following directory:

/trifacta/hadoop-deps

Configure the bundle distribution property (hadoopBundleJar):

  "hadoopBundleJar": "hadoop-deps/hdp-2.6/build/libs/hdp-2.6-bundle.jar"

Configure component settings

For each of the following components, please explicitly set the following settings.

1. You can apply this change through the Admin Settings Page (recommended) or trifacta-conf.json. For more information, see Platform Configuration Methods.

2. Configure Batch Job Runner:

     "batch-job-runner": {
      "autoRestart": true,
       ...
       "classpath": "%(topOfTree)s/hadoop-data/build/install/hadoop-data/hadoop-data.jar:%(topOfTree)s/hadoop-data/build/install/hadoop-data/lib/*:%(topOfTree)s/conf/hadoop-site:/usr/hdp/current/hadoop-client/hadoop-azure.jar:/usr/hdp/current/hadoop-client/lib/azure-storage-2.2.0.jar"
     },

3. Configure the following environment variables:

   "env.PATH": "${HOME}/bin:$PATH:/usr/local/bin:/usr/lib/zookeeper/bin",
   "env.TRIFACTA_CONF": "/opt/trifacta/conf"
   "env.JAVA_HOME": "/usr/lib/jvm/java-1.8.0-openjdk-amd64",

4. Configure the following properties for various Alteryx components:

     "ml-service": {
      "autoRestart": true
     },
     "monitor": {
      "autoRestart": true,
       ...
      "port": <your_cluster_monitor_port>
     },
     "proxy": {
      "autoRestart": true
     },
     "udf-service": {
      "autoRestart": true
     },
     "webapp": {
       "autoRestart": true
     },

5. Disable S3 access:

   "aws.s3.enabled": false,

6. Configure the following Spark Job Service properties:

   "spark-job-service.classpath": "%(topOfTree)s/services/spark-job-server/server/build/libs/spark-job-server-bundle.jar:%(topOfTree)s/conf/hadoop-site/:%(topOfTree)s/services/spark-job-server/build/bundle/*:/usr/hdp/current/hadoop-client/hadoop-azure.jar:/usr/hdp/current/hadoop-client/lib/azure-storage-2.2.0.jar",
   "spark-job-service.env.SPARK_DIST_CLASSPATH": "/usr/hdp/current/hadoop-client/*:/usr/hdp/current/hadoop-mapreduce-client/*",

7. Save your changes.

Configure High Availability

If you are integrating the platform the HDI cluster with high availability enabled, please complete the following steps so that the platform is aware of the failover nodes

Steps:

1. You can apply this change through the Admin Settings Page (recommended) or trifacta-conf.json. For more information, see Platform Configuration Methods.

2. Enable high availability feature on the namenode and resourceManager nodes:

   "feature.highAvailability.namenode": true,
   "feature.highAvailability.resourceManager": true,

3. For each YARN resource manager, you must configure its high availability settings. The following are two example node configurations, including the default port numbers for HDI:

   Tip: Host and port settings should be available in the cluster configuration files you copied to the Alteryx node. Or you can acquire the settings through the cluster's admin console.

     "yarn": {
       "highAvailability": {
         "resourceManagers": {
           "rm1": {
             "port": <your_cluster_rm1_port>,
             "schedulerPort": <your_cluster_rm1_scheduler_port>,
             "adminPort": <your_cluster_rm1_admin_port>,
             "webappPort": <your_cluster_rm1_webapp_port>
           },
           "rm2": {
             "port": <your_cluster_rm2_port>,
             "schedulerPort": <your_cluster_rm2_scheduler_port>,
             "adminPort": <your_cluster_rm2_admin_port>,
             "webappPort": <your_cluster_rm2_webapp_port>
           }
         }
       }
     },

4. Configure the high availability namenodes. The following example configures two namenodes (nn1 and nn2), including the default port numbers for HDI: 

   Tip: Host and port settings should be available in the cluster configuration files you copied to the Alteryx node. Or you can acquire the settings through the cluster's admin console.

     "hdfs": {
       ...
       "highAvailability": {
         "namenodes": {
           "nn1": {
             "port": <your_cluster_namenode1_port>
           },
           "nn2": {
             "port": <your_cluster_namenode2_port>
           }
         }
       ...

5. Save your changes.

Create Hive connection

Limitations:

1. The platform only supports HTTP connections to Hive on HDI. TCP connections are not supported.
2. The Hive port must be set to 10001 for HTTP.

For more information, see Create Hive Connections.

Hive integration requires additional configuration.

For more information, see Configure for Hive.

Configure for Spark Profiling

If you are using Spark for profiling, you must add environment properties to your cluster configuration. See Configure for Spark.

Configure for UDFs

If you are using user-defined functions (UDFs) on your HDInsight cluster, additional configuration is required. See Java UDFs.

Configure Storage

Before you begin running jobs, you must specify your base storage layer, which can be WASB or ADLS. For more information, see Set Base Storage Layer.

Additional configuration is required:

• For more information, see Enable WASB Access.
• For more information, see Enable ADLS Access.