def cache_job(self):
"""Cache a job."""
workflow = Session.query(Workflow).filter_by(
id_=self.workflow_uuid).one_or_none()
access_times = calculate_file_access_time(workflow.workspace_path)
prepared_job_cache = JobCache()
prepared_job_cache.job_id = self.job_id
prepared_job_cache.access_times = access_times
Session.add(prepared_job_cache)
Session.commit()
def test_workspace_deletion(app, session, default_user,
sample_yadage_workflow_in_db,
tmp_shared_volume_path, workspace, hard_delete):
"""Test workspace deletion."""
workflow = sample_yadage_workflow_in_db
create_workflow_workspace(sample_yadage_workflow_in_db.get_workspace())
absolute_workflow_workspace = os.path.join(tmp_shared_volume_path,
workflow.get_workspace())
# create a job for the workflow
workflow_job = Job(id_=uuid.uuid4(), workflow_uuid=workflow.id_)
job_cache_entry = JobCache(job_id=workflow_job.id_)
session.add(workflow_job)
session.add(job_cache_entry)
session.commit()
# create cached workspace
cache_dir_path = os.path.abspath(
os.path.join(absolute_workflow_workspace, os.pardir, 'archive',
str(workflow_job.id_)))
os.makedirs(cache_dir_path)
# check that the workflow workspace exists
assert os.path.exists(absolute_workflow_workspace)
assert os.path.exists(cache_dir_path)
_delete_workflow(workflow, hard_delete=hard_delete, workspace=workspace)
if hard_delete or workspace:
assert not os.path.exists(absolute_workflow_workspace)
# check that all cache entries for jobs
# of the deleted workflow are removed
cache_entries_after_delete = JobCache.query.filter_by(
job_id=workflow_job.id_).all()
assert not cache_entries_after_delete
assert not os.path.exists(cache_dir_path)
def test_workspace_deletion(app,
session,
default_user,
yadage_workflow_with_name,
tmp_shared_volume_path,
workspace,
hard_delete):
"""Test workspace deletion."""
with app.test_client() as client:
res = client.post(url_for('api.create_workflow'),
query_string={
"user": default_user.id_},
content_type='application/json',
data=json.dumps(yadage_workflow_with_name))
assert res.status_code == 201
response_data = json.loads(res.get_data(as_text=True))
workflow = Workflow.query.filter(
Workflow.id_ == response_data.get('workflow_id')).first()
assert workflow
absolute_workflow_workspace = os.path.join(
tmp_shared_volume_path,
workflow.get_workspace())
# create a job for the workflow
workflow_job = Job(id_=uuid.uuid4(), workflow_uuid=workflow.id_)
job_cache_entry = JobCache(job_id=workflow_job.id_)
session.add(workflow_job)
session.add(job_cache_entry)
session.commit()
# check that the workflow workspace exists
assert os.path.exists(absolute_workflow_workspace)
with app.test_client() as client:
res = client.put(
url_for('api.set_workflow_status',
workflow_id_or_name=workflow.id_),
query_string={
'user': default_user.id_,
'status': 'deleted'
},
content_type='application/json',
data=json.dumps({'hard_delete': hard_delete,
'workspace': workspace}))
if hard_delete or workspace:
assert not os.path.exists(absolute_workflow_workspace)
# check that all cache entries for jobs
# of the deleted workflow are removed
cache_entries_after_delete = JobCache.query.filter_by(
job_id=workflow_job.id_).all()
assert not cache_entries_after_delete
def test_workspace_deletion(
mock_update_user_quota,
mock_update_workflow_quota,
app,
session,
default_user,
sample_yadage_workflow_in_db,
workspace,
):
"""Test workspace deletion."""
workflow = sample_yadage_workflow_in_db
create_workflow_workspace(sample_yadage_workflow_in_db.workspace_path)
# Add file to the worskpace
file_size = 123
file_path = os.path.join(sample_yadage_workflow_in_db.workspace_path, "temp.txt")
with open(file_path, "w") as f:
f.write("A" * file_size)
# Get disk usage
disk_usage = get_disk_usage_or_zero(sample_yadage_workflow_in_db.workspace_path)
assert disk_usage
# Update disk quotas
store_workflow_disk_quota(sample_yadage_workflow_in_db)
update_users_disk_quota(sample_yadage_workflow_in_db.owner)
# create a job for the workflow
workflow_job = Job(id_=uuid.uuid4(), workflow_uuid=workflow.id_)
job_cache_entry = JobCache(job_id=workflow_job.id_)
session.add(workflow_job)
session.commit()
session.add(job_cache_entry)
session.commit()
# create cached workspace
cache_dir_path = os.path.join(
sample_yadage_workflow_in_db.workspace_path,
"..",
"archive",
str(workflow_job.id_),
)
os.makedirs(cache_dir_path)
# check that the workflow workspace exists
assert os.path.exists(sample_yadage_workflow_in_db.workspace_path)
assert os.path.exists(cache_dir_path)
delete_workflow(workflow, workspace=workspace)
if workspace:
assert not os.path.exists(sample_yadage_workflow_in_db.workspace_path)
mock_update_user_quota.assert_called_once_with(
sample_yadage_workflow_in_db.owner,
bytes_to_sum=-disk_usage,
override_policy_checks=True,
)
mock_update_workflow_quota.assert_called_once_with(
sample_yadage_workflow_in_db,
bytes_to_sum=-disk_usage,
override_policy_checks=True,
)
else:
assert not mock_update_user_quota.called
assert not mock_update_workflow_quota.called
# check that all cache entries for jobs
# of the deleted workflow are removed
cache_entries_after_delete = JobCache.query.filter_by(job_id=workflow_job.id_).all()
assert not cache_entries_after_delete
assert not os.path.exists(cache_dir_path)
def create_job(): # noqa
r"""Create a new job.
---
post:
summary: Creates a new job.
description: >-
This resource is expecting JSON data with all the necessary information
of a new job.
operationId: create_job
consumes:
- application/json
produces:
- application/json
parameters:
- name: job
in: body
description: Information needed to instantiate a Job
required: true
schema:
$ref: '#/definitions/JobRequest'
responses:
201:
description: Request succeeded. The job has been launched.
schema:
type: object
properties:
job_id:
type: string
examples:
application/json:
{
"job_id": "cdcf48b1-c2f3-4693-8230-b066e088c6ac"
}
400:
description: >-
Request failed. The incoming data specification seems malformed.
500:
description: >-
Request failed. Internal controller error. The job could probably
not have been allocated.
"""
json_data = request.get_json()
if not json_data:
return jsonify({'message': 'Empty request'}), 400
# Validate and deserialize input
job_request, errors = job_request_schema.load(json_data)
if errors:
return jsonify(errors), 400
job_parameters = dict(
job_id=str(job_request['job_id']),
workflow_workspace=str(job_request['workflow_workspace']),
docker_img=job_request['docker_img'],
cmd=job_request['cmd'],
cvmfs_mounts=job_request['cvmfs_mounts'],
env_vars=job_request['env_vars'],
shared_file_system=job_request['shared_file_system'],
job_type=job_request.get('job_type'))
job_obj = k8s_instantiate_job(**job_parameters)
if job_obj:
job = copy.deepcopy(job_request)
job['status'] = 'started'
job['restart_count'] = 0
job['max_restart_count'] = 3
job['deleted'] = False
job['obj'] = job_obj
JOB_DB[str(job['job_id'])] = job
job_db_entry = JobTable(
id_=job['job_id'],
workflow_uuid=None,
# The workflow_uuid is populated by the workflow-controller
status=job['status'],
job_type=job_request.get('job_type'),
cvmfs_mounts=job_request['cvmfs_mounts'],
shared_file_system=job_request['shared_file_system'],
docker_img=job_request['docker_img'],
experiment=job_request['experiment'],
cmd=job_request['cmd'],
env_vars=json.dumps(job_request['env_vars']),
restart_count=job['restart_count'],
max_restart_count=job['max_restart_count'],
deleted=job['deleted'],
name=job_request['job_name'],
prettified_cmd=job_request['prettified_cmd'])
Session.add(job_db_entry)
Session.commit()
access_times = calculate_file_access_time(
json_data['workflow_workspace'])
prepared_job_cache = JobCache()
prepared_job_cache.job_id = job['job_id']
prepared_job_cache.access_times = access_times
Session.add(prepared_job_cache)
Session.commit()
return jsonify({'job_id': job['job_id']}), 201
else:
return jsonify({'job': 'Could not be allocated'}), 500