pndaproject/example-applications

Example applications for use with PNDA

Example Applications

This repository contains a number of example applications that can be built and run on PNDA. Each application directory contains more detailed information.

Spark Streaming

Examples of consuming data from Kafka and populating both HBase and OpenTSDB with simple Scala based Spark Streaming applications.

• Write to HBase (scala)
• Write to OpenTSDB (scala)
• Count messages (python)

Spark

Example of consuming data ingested by Gobblin on a batch basis and producing Parquet datasets, optimized for consumption by Impala.

• Write to parquet format (scala)
• Write to parquet format (python)

Jupyter

Example of a notebook for manipulating network data.

H2O

Application that runs the H2O data science platform as an application on PNDA.

Flink Streaming

• Count Words (scala) Count the words from Socket.
• Count Words (python) Count the words from input file.
• Flink Windows (java) host-network-data-usage illustrating Flink windows, triggers and event processing.
• Count Hashtags (java) specific word count from input file illustrating metrics, counters and accumulators.

Compound Packages

An example of a package containing multiple application component types, in this case a Spark app and related Jupyter notebook.

example-applications's Issues

Spark Magic Usage

User guide is required for usage of Spark Magic.

Steps regarding usage of sparkmagic in spark kernel will be helpful in order to use various predefined magics. Details regarding livy connectivity to interact with spark can be added in steps which will help in creating and managing sessions.

Maven Build Issue

Ran into this issue trying to build spark-streaming application.

Failed to execute goal org.apache.maven.plugins:maven-assembly-plugin:2.6:single (make-assembly) on project spark-streaming-to-hbase-example-app: Execution make-assembly of goal org.apache.maven.plugins:maven-assembly-plugin:2.6:single failed: group id '1385651864' is too big ( > 2097151 ). Use STAR or POSIX extensions to overcome this limit -> [Help 1]

Fix seems to adding tarLongFileMode in appropriate pom.xml file - app-package/pom.xml in this case.

    ```
<configuration>
           <tarLongFileMode>posix</tarLongFileMode>
           <descriptors>
               <descriptor>src/main/assembly/assembly.xml</descriptor>
           </descriptors>
           <appendAssemblyId>false</appendAssemblyId>
        </configuration>

Should Kafka Version in examples be 0.10.0.1

In examples that are using kafka (eg: spark-streaming) should pom.xml be updated to have 0.10.0.1 version of kafka clients? I believe that is what is being used in platform.