DocsICP FilingConsole
官方文档
首页

Start a deployment

更新时间：
复制 MD 格式
产品详情
我的收藏

After creating a deployment, you must start it on the Deployments page. Starting a deployment is also necessary to resume a stopped deployment or to apply parameter updates that require a restart.

Prerequisites

An existing deployment is required. For more information, see Create a deployment.

Limitations

You can specify startup options only for a streaming deployment.

Precautions

  • If a RAM user, RAM role, or another Alibaba Cloud account starts a deployment, ensure that the corresponding identity has access to the target namespace. For more information, see Authorize access to the fully managed Flink console and Manage permissions.

  • If you start a deployment by using the Latest State or Specific State option, the system performs a state compatibility check. Starting a deployment with detected incompatibilities is risky, as it may fail or produce unexpected results. For more information, see Flink state compatibility reference.

Procedure

  1. Open the Start dialog for the deployment.

    1. Log on to the fully managed Flink console as a user who has the owner role for the namespace.

    2. In the top navigation bar, select the target namespace.

    3. On the O&M > Deployments page, from the drop-down list, select streaming deployment or batch deployment.

  2. Find the target deployment and in the Actions column, click Start.

  3. (Optional) For a streaming deployment, configure the startup options.

    • Initial mode

      Select this mode for a new deployment or when you cannot reuse an existing state. Based on your business requirements, you can also decide whether to enable Autopilot.

      Strategy

      Description

      Specify source's start time

      Select Specify source's start time and enter a specific time.

      You can set a start time in the UI for the following seven connectors: Kafka, Log Service (SLS), DataHub, ApsaraMQ for RocketMQ, Hologres, Paimon, and MySQL.

      The data read time specified on the deployment startup interface takes precedence over the startTime set in the deployment DDL code.

      Note

      • Kafka versions earlier than 0.11 may not be supported due to potential incompatibility with the Kafka client version used by the connector. We recommend that you upgrade your Kafka version.

      • Not all connectors support configuring the startTime parameter. For more information, see the WITH parameters for each connector, such as Log Service (SLS) WITH parameters, to check whether startTime is included.

      • Only when you start a new deployment and specify startTime does startTime take effect. If you start a deployment from a checkpoint or a savepoint, specifying startTime has no effect.

      Configure Autopilot

      Turn on this switch and select a tuning mode:

      • Autopilot Mode: The system automatically scales resources down when usage is low and scales them up when usage reaches a certain threshold. For more information, see Enable and configure Autopilot.

      • Scheduled Mode: You must select a schedule from the drop-down list. A schedule can contain multiple resource configurations for different time points. You can configure resources based on usage patterns during various periods. For more information, see Configure and apply a scheduled tuning plan.

    • Resume mode

      You can select a startup strategy based on your business requirements and decide whether to enable Autopilot.

      Strategy

      Description

      Latest State

      Resumes the deployment from the latest available savepoint or checkpoint. If you select this option for an SQL deployment, the Flink system detects changes in the SQL code, Flink runtime parameters, and engine version.

      If changes are detected, we recommend that you click Click to detect next to State Compatibility to check for compatibility. Then, proceed based on the results. For information about the compatibility results and suggested actions, see Compatibility.

      Specific State

      Select a specific savepoint to resume from. To learn how to create a savepoint, see Manage deployment state.

      Other deployment state

      After you select this option, specify a target deployment and its corresponding savepoint to resume from. Savepoints can be shared between deployments, but the state must be compatible. For more information, see Manage deployment state.

      Allow Non-restored State (AllowNonRestoredState)

      Note

      This option is supported only for JAR deployments.

      By default, Flink attempts to map the entire savepoint to the operators of the new deployment. If modifications to the deployment cause changes to the operator state, the deployment might fail to resume. Enable this option to allow Flink to skip any state that cannot be mapped and proceed with startup. For more information, see Allow Non-Restored State.

      Configure Autopilot

      Turn on this switch and select a tuning mode:

      • Autopilot Mode: The system automatically scales resources down when usage is low and scales them up when usage reaches a certain threshold. For more information, see Enable and configure Autopilot.

      • Scheduled Mode: You must select a schedule from the drop-down list. A schedule can contain multiple resource configurations for different time points. You can configure resources based on usage patterns during various periods. For more information, see Configure and apply a scheduled tuning plan.

  4. Click Start.

    On the O&M > Deployments page, view the status of the deployment. For more information, see View deployment status.

Related documents

Previous：Dynamic scaling and dynamic parameter updatesNext：View and manage job status