All Products
Search

This Product

    Realtime Compute for Apache Flink:Quick start with Flink JAR jobs

    Document Center

    Realtime Compute for Apache Flink:Quick start with Flink JAR jobs

    Last Updated:Mar 26, 2026

    This guide walks you through deploying and running a Flink JAR job on Realtime Compute for Apache Flink. You'll complete the workflow for both stream and batch modes, from uploading your JAR to verifying results.

    This guide covers the following steps:

    1. Develop a JAR package

    2. Upload the JAR package and data file

    3. Deploy a JAR job

    4. Start the job and view results

    5. (Optional) Stop the job

    Prerequisites

    Before you begin, ensure that you have:

    Step 1: Develop a JAR package

    The Flink console does not include a development environment. Build and package your code locally, then upload the resulting JAR.

    For guidance on configuring dependencies, using connectors, and reading additional files from Object Storage Service (OSS), see Develop a JAR job.

    Important

    Match the Flink version used in local development with the engine version you select in Step 3. Also verify the scope of your dependency packages.

    To follow along with this guide, download the test resources below. They contain a word-frequency counter that runs in both stream and batch mode.

    Step 2: Upload the JAR package and data file

    1. Log on to the Realtime Compute for Apache Flink console.

    2. In the Actions column for your workspace, click Console.

    3. In the left navigation pane, click File Management.

    4. Click Upload Resource, then upload both the JAR package and the Shakespeare data file. For details on file storage paths, see File management.

    Step 3: Deploy a JAR job

    Stream job

    1. Go to Job O&M > Operation Center, click Deploy Job, and select JAR Job.

    2. Fill in the deployment parameters. The table below lists the key parameters — use the example values to follow this guide. For all other parameters, see Deploy a job.

      Note

      Ververica Runtime (VVR) 8.0.6 and later support access only to the bucket that was attached when you activated the Flink workspace. Access to other buckets is not supported.

      Important

      Session clusters do not support monitoring and alerts, auto-tuning configuration, or production workloads. Use session clusters for development and testing only. See Debug jobs.

      Parameter Description Example
      Deployment mode Select Stream for a streaming job. Stream
      Deployment name A name for this job deployment. flink-streaming-test-jar
      Engine version The Flink engine version to run the job. Select a version tagged Recommended or Stable for higher reliability. See Release notes and Engine versions. vvr-8.0.9-flink-1.17
      JAR URI The JAR file to run. Select the FlinkQuickStart-1.0-SNAPSHOT.jar file you uploaded in Step 2. If the file is already in File Management, select it directly — no re-upload needed. —
      Entry point class The main class of the program. If your JAR does not declare a main class in its manifest, enter the full class name. The test JAR contains both stream and batch code, so you must specify the stream entry point explicitly. org.example.WordCountStreaming
      Entry point main arguments Runtime arguments passed to the main method. Enter the OSS path of the Shakespeare file you uploaded. Copy the full path from File Management. --input oss://<your-attached-OSS-Bucket-name>/artifacts/namespaces/<project-name>/Shakespeare

      --input oss://<your attached OSS bucket name>/artifacts/namespaces/<project name>/Shakespeare

      --output oss://<your attached OSS bucket name>/artifacts/namespaces/<project name>/batch-quickstart-test-output.txt

      You can copy the full path of the Shakespeare file directly in File Management.

      --input oss://<your-attached-OSS-Bucket-name>/artifacts/namespaces/<project-name>/Shakespeare

      You can copy the full path of the Shakespeare file from File Management.
      Deployment target The resource queue or session cluster to run the job on. See Manage resource queues and Create a session cluster. default-queue

      部署jar.jpg

    3. Click Deploy.

    Batch job

    1. Go to Job O&M > Operation Center, click Deploy Job, and select JAR Job.

    2. Fill in the deployment parameters for the batch job. For all other parameters, see Deploy a job.

      Important

      Session clusters do not support monitoring and alerts, auto-tuning configuration, or production workloads. Use session clusters for development and testing only. See Debug jobs.

      Parameter Description Example
      Deployment mode Select Batch for a batch job. Batch
      Deployment name A name for this job deployment. flink-batch-test-jar
      Engine version The Flink engine version to run the job. Select a version tagged Recommended or Stable. See Release notes and Engine versions. vvr-8.0.9-flink-1.17
      JAR URI The JAR file to run. Select the FlinkQuickStart-1.0-SNAPSHOT.jar file you uploaded in Step 2. —
      Entry point class The main class of the program. The test JAR contains both stream and batch code, so specify the batch entry point explicitly. org.example.WordCountBatch
      Entry point main arguments Runtime arguments passed to the main method. Enter the OSS paths for both the input file (Shakespeare) and the output file. Copy the input file path from File Management. You only need to specify the full path and name of the output file — you do not need to create it in the storage service beforehand. Keep the output file in the same directory as the input file. --input oss://<your-attached-OSS-Bucket-name>/artifacts/namespaces/<project-name>/Shakespeare --output oss://<your-attached-OSS-Bucket-name>/artifacts/namespaces/<project-name>/batch-quickstart-test-output.txt
      Deployment target The resource queue or session cluster to run the job on. See Manage resource queues and Create a session cluster. default-queue

      部署jar_批_ch.jpg

    3. Click Deploy.

    Step 4: Start the job and view results

    Stream job

    1. In Job O&M > Operation Center, find your stream job and click Start in the Actions column.

      流作业启动.jpg

    2. Select Stateless Start and click Start. For other start options, see Start a job.

    3. When the job status changes to Running, open the TaskManager log file that ends with .out and search for "shakespeare" to view the word-frequency results.

      image.png

    Batch job

    1. In Job O&M > Operation Center, find your batch job and click Start.

      批作业启动.jpg

    2. In the Start Job dialog box, click Start. For other start options, see Start a job.

    3. When the job status changes to Finished, log on to the OSS console and open the output file at:

      Log on to the OSS console. You can view the file in the oss://<Your attached OSS Bucket name>/artifacts/namespaces/<Project name>/batch-quickstart-test-output.txt path.

      oss://<your-attached-OSS-Bucket-name>/artifacts/namespaces/<project-name>/batch-quickstart-test-output.txt

      批作业结果

    Note

    The TaskManager .out log displays a maximum of 2,000 records. Because of this limit, the stream job and batch job may show different result record counts. See Print for details.

    Step 5: (Optional) Stop the job

    If you modify a job and want the changes to take effect, you must redeploy the job, and then stop and restart it. You need to stop and restart a job when you make changes such as:

    • Modifying the JAR code

    • Adding or removing WITH parameters

    • Changing the job version

    • Starting a completely new job

    • Updating parameters that don't take effect dynamically

    • The job cannot reuse a state

    For instructions, see Stop a job.

    What's next