def check_job_status(job_id):
jobs = query_job(job_id=job_id)
for job in jobs:
if job.f_status != JobStatus.RUNNING:
return False
return True
def job_quantity_constraint(self):
if RuntimeConfig.WORK_MODE == WorkMode.CLUSTER:
if self.role == 'host':
running_jobs = job_utils.query_job(status='running', role=self.role)
if len(running_jobs) >= MAX_CONCURRENT_JOB_RUN_HOST:
raise Exception('The job running on the host side exceeds the maximum running amount')
def stop(job_id, end_status=JobStatus.FAILED, component_name=''):
schedule_logger(job_id).info('get {} job {} {} command'.format(
"cancel" if end_status == JobStatus.CANCELED else "stop", job_id,
component_name))
jobs = job_utils.query_job(job_id=job_id, is_initiator=1)
cancel_success = False
is_cancel = (end_status == JobStatus.CANCELED)
if jobs:
initiator_job = jobs[0]
job_info = {'f_job_id': job_id, 'f_status': end_status}
roles = json_loads(initiator_job.f_roles)
job_work_mode = initiator_job.f_work_mode
initiator_party_id = initiator_job.f_party_id
# set status first
if not component_name:
TaskScheduler.sync_job_status(
job_id=job_id,
roles=roles,
initiator_party_id=initiator_party_id,
initiator_role=initiator_job.f_role,
work_mode=job_work_mode,
job_info=job_info)
for role, partys in roles.items():
for party_id in partys:
response = federated_api(
job_id=job_id,
method='POST',
endpoint='/{}/schedule/{}/{}/{}/{}'.format(
API_VERSION, job_id, role, party_id,
"cancel" if is_cancel else "kill"),
src_party_id=initiator_party_id,
dest_party_id=party_id,
src_role=initiator_job.f_role,
json_body={
'job_initiator': {
'party_id': initiator_job.f_party_id,
'role': initiator_job.f_role
},
'timeout': end_status == JobStatus.TIMEOUT,
'component_name': component_name
},
work_mode=job_work_mode)
if response['retcode'] == 0:
cancel_success = True
schedule_logger(job_id).info(
'send {} {} {} job {} {} command successfully'.
format(role, party_id,
"cancel" if is_cancel else "kill", job_id,
component_name))
if is_cancel:
break
else:
schedule_logger(job_id).info(
'send {} {} {} job {} {} command failed: {}'.
format(role, party_id,
"cancel" if is_cancel else "kill", job_id,
component_name, response['retmsg']))
if is_cancel:
return cancel_success
else:
schedule_logger(job_id).info(
'send {} job {} {} command failed'.format(
"cancel" if is_cancel else "kill", job_id, component_name))
raise Exception('can not found job: {}'.format(job_id))
def query_job():
jobs = job_utils.query_job(**request.json)
if not jobs:
return get_json_result(retcode=101, retmsg='find job failed')
return get_json_result(retcode=0, retmsg='success', data=[job.to_json() for job in jobs])
