The following instructions guide you through the installation a simple Flink cluster to locally execute Flink programs on your machine.

In order to execute programs on a running Flink instance (rather than from within your IDE), you need to install Flink on your machine. To do so, follow these steps:

  • Download the Apache Flink 1.6.0 release from the download page. Since we won’t use HDFS or YARN, any Hadoop version will work, including the “without bundled hadoop” version.
  • Extract the downloaded archive
  • The resulting folder contains a Flink setup that can be locally executed without any further configuration.

If you have a local Flink installation, you can easily bring up a small cluster for local testing.

On UNIX system you can start a Flink instance as follows:

cd /to/your/flink/installation
./bin/start-cluster.sh

On Windows you have to run the following commands

cd C:\to\your\flink\installation
.\bin\start-cluster.bat

You can validate that a local Flink instance is running by looking at the log files in ./log/ or opening the JobManager’s webinterface at http://localhost:8081.

4. Run an application

The How to do the Exercises page explains how to build and package a Flink program with Maven. You can use the Flink CLI to run applications that have been packaged into a jar file. Here’s an example of running one of the examples, but you’ll probably need to make some adjustments to the paths and version number:

flink run -c \  
    com.dataartisans.flinktraining.examples.datastream_java.basics.RideCount \
    ~/flink-training-exercises/target/flink-training-exercises-2.4.0.jar

Because this jar file contains many applications, we’ve had to specify which class to run, using the -c flag.

Note that the taskmanager(s) will write their output to files in the ./log/ directory.

On UNIX you call

./bin/stop-cluster.sh

On Windows you quit the running process with Ctrl-C.