After you modify the SQL logic, change the job version, add parameters to the WITH clause, or add job parameters for a job, you must terminate and then start the job to make the changes take effect. This topic describes how to terminate a job.
- You can only terminate a job that is in the Running or Starting state.
- If you terminate a job, its task status is cleared. For example, if the job you terminate is running a COUNT operation, the COUNT operation starts from 0 after you start the job.
- The Terminate (checkpoint) operation is supported in Realtime Compute for Apache Flink V3.5.0 and later. If your Realtime Compute for Apache Flink is earlier than V3.5.0, the following error message is displayed when you try to perform this operation: An error occurred. System error: The BLINK version is abnormal. Error reason: blink version >= blink-3.5 is required, instance blink-3.4.4.
- Log on to the Realtime Compute development platform.
- In the top navigation bar, click Administration.
- On the Administration page, find the job that you want to terminate, and click Terminate in the Actions column.
Note The Terminate (checkpoint) operation under More is different from the Terminate operation. The system triggers a checkpoint when you perform the Terminate (checkpoint) operation to terminate a job. Therefore, the time consumed to terminate a job by performing the Terminate (checkpoint) operation is longer than that by performing the Terminate operation. The job status is cleared after the job is terminated. The Terminate (checkpoint) operation has other functions in some scenarios. For example, if the upstream storage system is Message Queue for Apache Kafka, the system submits an offset each time it triggers a checkpoint. This ensures that the number of offsets submitted to the Kafka server is consistent with the amount of data consumed.