After you create and publish a job, you must start the job on the Developments page to run the job. If you want to resume a job that is suspended or canceled, you must start the job.

Background information

You can start a job in the following scenarios:
  • Start a job that is published: You must select Start without state for Start Method and specify Start Time for Reading Data.
  • Start a job that is suspended: When you suspend a job, you can choose whether to create a savepoint before the job is suspended based on your business requirements. The process of starting a job that is suspended varies based on whether you select Create a Savepoint before suspending:
    • If you select Create a Savepoint before suspending, the system resumes the job that is suspended from the last savepoint.
      Note If the savepoint is run but the system does not start to suspend the job, checkpointing is not performed.
    • If you do not select Create a Savepoint before suspending, the system resumes the job that is suspended from the last checkpoint.
  • Start a job that is canceled: When you cancel a job, you can choose whether to create a savepoint before the job is canceled based on your business requirements.
    • If you select Create a Savepoint before cancelling, a savepoint is created for the job before the job is canceled. Create a Savepoint before cancelling is selected by default. After the job is canceled, the system automatically deletes the checkpoint information about the job.
    • If you do not select Create a Savepoint before cancelling, the job is directly canceled. After the job is canceled, the system automatically deletes the checkpoint information about the job.

    If you start a job that is canceled, you can select Start without state for Start Method and specify Start Time for Reading Data. You can also resume the job from the savepoint that you select.

Limits

  • Only Realtime Compute for Apache Flink whose engine version is Ververica Runtime (VVR) 4.0.11 or later supports state data compatibility check.
  • Only Realtime Compute for Apache Flink whose engine version is VVR 4.0.8 or later allows you to specify the start offset.
    The Start Time for Reading Data parameter that is specified in the Deployment Starting Configuration dialog box takes precedence over the startTime parameter that is configured in the DDL code of a job. If you configure the Start Time for Reading Data parameter in both DDL code and the Deployment Starting Configuration dialog box, the configuration of the Start Time for Reading Data parameter that you specified in the Deployment Starting Configuration dialog box takes effect.
    Note
    • Not all connector types support the startTime parameter. For more information about whether the parameters in the WITH clause for each type of connector include the startTime parameter, see the related topic of the specific type of connector. For example, if you want to check whether the parameters in the WITH clause for a Log Service connector include the startTime parameter, see Parameters in the WITH clause for a Log Service connector.
    • The startTime parameter takes effect only if a new job is started and the startTime parameter is specified. If you start a job based on a checkpoint or savepoint, the startTime parameter does not take effect even if this parameter is specified.

Precautions

When you start a job, if you choose to start the job from a checkpoint or savepoint, the system checks the compatibility between the job and state data. For more information about the state data compatibility check, see Reference for Flink state data compatibility.

Procedure

  1. Log on to the Realtime Compute for Apache Flink console.
  2. On the Fully Managed Flink tab, find the workspace that you want to manage and click Console in the Actions column.
  3. In the left-side navigation pane, choose Applications > Deployments.
  4. Find the job that you want to start and click Start in the Actions column.
  5. In the Deployment Starting Configuration dialog box, configure Start Method and Start Options.
    Deployment Starting Configuration dialog box in different scenarios:
    • Start a job that is publishedStart a job that is published
    • Start a job that is suspendedStart a job that is suspended
    • Start a job that is canceledStart a new job or a job that is canceled
  6. In the Deployment Starting Configuration dialog box, click Confirm Running.
  7. View the status of the job.
    Job statusThe following table describes the meaning of each job state.
    Type State Description
    Transitioning STARTING The job is being started.
    Stable RUNNING The job is running.
    Note If the system detects a risk for a job that is running, the system displays the risk type and risk level for the job on the right side of RUNNING. The risk type is Unstable, Failing, or ClusterUnreachable, and the risk level is High, Middle, or Low. In this case, you can click the name of the job to go to the job details page. Then, you can click Diagnosis to view the cause of the risk. For more information, see Job diagnostics.
    SUSPENDED If you click Suspend for a job in the Actions column, the job enters the SUSPENDED state.
    CANCELLED If you click Cancel for a job in the Actions column, the job enters the CANCELLED state.
    Failed FAILED The job fails to run.
    Note Jobs in the FAILED state are considered abnormal jobs.