This documentation is for an unreleased version of Apache Flink. We recommend you use the latest stable version.

First Steps #

Welcome to Apache Flink! This guide will help you get a Flink cluster up and running so you can start exploring Flink’s capabilities.

Prerequisites #

Choose one of the following installation methods:

• Docker: No Java installation needed, includes SQL Client
• Local Installation: Requires Java 11, 17, or 21
• PyFlink: Requires Java and Python 3.9+

Option A: Docker Installation #

The fastest way to get started with Flink. No Java installation required.

Step 1: Get the docker-compose.yml file #

Download docker-compose.yml or create a file named docker-compose.yml with the following content:

services:
  jobmanager:
    image: flink:latest
    ports:
      - "8081:8081"
    command: jobmanager
    environment:
      - |
        FLINK_PROPERTIES=
        jobmanager.rpc.address: jobmanager        

  taskmanager:
    image: flink:latest
    depends_on:
      - jobmanager
    command: taskmanager
    scale: 1
    environment:
      - |
        FLINK_PROPERTIES=
        jobmanager.rpc.address: jobmanager
        taskmanager.numberOfTaskSlots: 2        

  sql-client:
    image: flink:latest
    depends_on:
      - jobmanager
    command: bin/sql-client.sh
    environment:
      - |
        FLINK_PROPERTIES=
        jobmanager.rpc.address: jobmanager
        rest.address: jobmanager        

Step 2: Start the Cluster #

$ docker compose up -d

Step 3: Verify the Cluster is Running #

Open the Flink Web UI at http://localhost:8081 to verify the cluster is running.

Using the SQL Client #

To start an interactive SQL session:

$ docker compose run sql-client

To exit the SQL Client, type exit; and press Enter.

Stopping the Cluster #

$ docker compose down

For more Docker options (scaling, Application Mode), see the Docker deployment guide.

Option B: Local Installation #

If you prefer to run Flink directly on your machine without Docker.

Step 1: Download Flink #

Flink runs on all UNIX-like environments, including Linux, Mac OS X, and Cygwin (for Windows).

First, verify your Java version:

$ java -version

You need Java 11, 17, or 21 installed. Then, download the latest binary release and extract the archive:

$ tar -xzf flink-*.tgz
$ cd flink-*

Step 2: Start the Cluster #

Start a local Flink cluster:

$ ./bin/start-cluster.sh

Step 3: Verify the Cluster is Running #

Open the Flink Web UI at http://localhost:8081 to verify the cluster is running. You should see the Flink dashboard showing one TaskManager with available task slots.

Using the SQL Client #

To start an interactive SQL session:

$ ./bin/sql-client.sh

To exit the SQL Client, type exit; and press Enter.

Stopping the Cluster #

When you’re done, stop the cluster with:

$ ./bin/stop-cluster.sh

Option C: PyFlink Installation #

For Python development with the Table API or DataStream API. No Flink cluster is required—PyFlink runs in local mode during development.

Step 1: Verify Prerequisites #

PyFlink requires Java and Python:

$ java -version
# Java 11, 17, or 21

$ python --version
# Python 3.9, 3.10, 3.11, or 3.12

Step 2: Install PyFlink #

$ python -m pip install apache-flink

Tip: We recommend installing PyFlink in a virtual environment to keep your project dependencies isolated.

Step 3: Verify Installation #

$ python -c "import pyflink; print(pyflink.__version__)"

You’re now ready to follow the Table API Tutorial or DataStream API Tutorial using the Python tabs.

Next Steps #

Choose a tutorial to start learning:

To learn more about Flink’s core concepts, visit the Concepts section.