在 AWS EMR 步骤的代码中添加以下内容以确保步骤在出现错误时停止执行：

from __future__ import print_function
import sys
import boto3

# Set up the command to run on the cluster
command = "command-to-run"

# Run the command on the cluster
response = client.run_job_flow(
    Name='job-name',
    ReleaseLabel='emr-release',
    Instances={
        'InstanceGroups': [
            {
                'Name': 'Master nodes',
                'Market': 'ON_DEMAND',
                'InstanceRole': 'MASTER',
                'InstanceType': 'm4.xlarge',
                'InstanceCount': 1
            },
            {
                'Name': 'Slave nodes',
                'Market': 'SPOT',
                'InstanceRole': 'CORE',
                'InstanceType': 'm4.xlarge',
                'InstanceCount': 3,
                'BidPrice': '0.15'
            }
        ],
        'Ec2KeyName': 'key-pair',
        'KeepJobFlowAliveWhenNoSteps': False,  # set to False to stop the cluster when the steps are completed
        'TerminationProtected': False,
        'Ec2SubnetId': 'subnet-id'
    },
    Steps=[
        {
            'Name': 'job-step-name',
            'ActionOnFailure': 'TERMINATE_JOB_FLOW',  # stop the cluster if this step fails
            'HadoopJarStep': {
                'Jar': 'command-runner.jar',
                'Args': [command]
            }
        }
    ],
    VisibleToAllUsers=True,
    JobFlowRole='EMR_EC2_DefaultRole',
    ServiceRole='EMR_DefaultRole'
)

print("Cluster started with step: %s" % response['JobFlowId'])

在上面的代码示例中，我们使用 KeepJobFlowAliveWhenNoSteps 参数将 AWS EMR 群集设置为在步骤完成后停止运行。如果步骤失败，我们还可以使用 ActionOnFailure 参数将 AWS EMR 群集设置为终止。 停止运