def execute(self, context: 'Context') -> int:
emr_hook = EmrHook(aws_conn_id=self.aws_conn_id)
emr = emr_hook.get_conn()
if self.do_xcom_push:
context['ti'].xcom_push(key='cluster_id', value=self.cluster_id)
EmrClusterLink.persist(
context=context,
operator=self,
region_name=emr_hook.conn_region_name,
aws_partition=emr_hook.conn_partition,
job_flow_id=self.cluster_id,
)
self.log.info('Modifying cluster %s', self.cluster_id)
response = emr.modify_cluster(
ClusterId=self.cluster_id,
StepConcurrencyLevel=self.step_concurrency_level)
if response['ResponseMetadata']['HTTPStatusCode'] != 200:
raise AirflowException(f'Modify cluster failed: {response}')
else:
self.log.info('Steps concurrency level %d',
response['StepConcurrencyLevel'])
return response['StepConcurrencyLevel']
def execute(self, context: 'Context') -> str:
emr = EmrHook(aws_conn_id=self.aws_conn_id,
emr_conn_id=self.emr_conn_id,
region_name=self.region_name)
self.log.info(
'Creating JobFlow using aws-conn-id: %s, emr-conn-id: %s',
self.aws_conn_id, self.emr_conn_id)
if isinstance(self.job_flow_overrides, str):
job_flow_overrides: Dict[str, Any] = ast.literal_eval(
self.job_flow_overrides)
self.job_flow_overrides = job_flow_overrides
else:
job_flow_overrides = self.job_flow_overrides
response = emr.create_job_flow(job_flow_overrides)
if not response['ResponseMetadata']['HTTPStatusCode'] == 200:
raise AirflowException(f'JobFlow creation failed: {response}')
else:
job_flow_id = response['JobFlowId']
self.log.info('JobFlow with id %s created', job_flow_id)
EmrClusterLink.persist(
context=context,
operator=self,
region_name=emr.conn_region_name,
aws_partition=emr.conn_partition,
job_flow_id=job_flow_id,
)
return job_flow_id
def execute(self, context: 'Context') -> None:
emr_hook = EmrHook(aws_conn_id=self.aws_conn_id)
emr = emr_hook.get_conn()
EmrClusterLink.persist(
context=context,
operator=self,
region_name=emr_hook.conn_region_name,
aws_partition=emr_hook.conn_partition,
job_flow_id=self.job_flow_id,
)
self.log.info('Terminating JobFlow %s', self.job_flow_id)
response = emr.terminate_job_flows(JobFlowIds=[self.job_flow_id])
if not response['ResponseMetadata']['HTTPStatusCode'] == 200:
raise AirflowException(f'JobFlow termination failed: {response}')
else:
self.log.info('JobFlow with id %s terminated', self.job_flow_id)
def execute(self, context: 'Context') -> List[str]:
emr_hook = EmrHook(aws_conn_id=self.aws_conn_id)
emr = emr_hook.get_conn()
job_flow_id = self.job_flow_id or emr_hook.get_cluster_id_by_name(
str(self.job_flow_name), self.cluster_states)
if not job_flow_id:
raise AirflowException(
f'No cluster found for name: {self.job_flow_name}')
if self.do_xcom_push:
context['ti'].xcom_push(key='job_flow_id', value=job_flow_id)
EmrClusterLink.persist(
context=context,
operator=self,
region_name=emr_hook.conn_region_name,
aws_partition=emr_hook.conn_partition,
job_flow_id=job_flow_id,
)
self.log.info('Adding steps to %s', job_flow_id)
# steps may arrive as a string representing a list
# e.g. if we used XCom or a file then: steps="[{ step1 }, { step2 }]"
steps = self.steps
if isinstance(steps, str):
steps = ast.literal_eval(steps)
response = emr.add_job_flow_steps(JobFlowId=job_flow_id, Steps=steps)
if not response['ResponseMetadata']['HTTPStatusCode'] == 200:
raise AirflowException(f'Adding steps failed: {response}')
else:
self.log.info('Steps %s added to JobFlow', response['StepIds'])
return response['StepIds']