def test_set_default_cluster(self):
clusters = ie.current_env().clusters
master_url = 'test-url'
cluster_name = 'test-cluster'
project = 'test-project'
region = 'test-region'
pipelines = ['pid']
dashboard = 'test-dashboard'
cluster_id = obfuscate(project, region, cluster_name)
ie.current_env().inspector._clusters = {
cluster_id: {
'cluster_name': cluster_name,
'project': project,
'region': region,
'master_url': master_url,
'dashboard': dashboard,
'pipelines': pipelines
}
}
clusters.master_urls[master_url] = MasterURLIdentifier(
project, region, cluster_name)
clusters.set_default_cluster(
ie.current_env().inspector.get_cluster_master_url(cluster_id))
self.assertEqual(MasterURLIdentifier(project, region, cluster_name),
clusters.default_cluster_metadata)
def test_clusters_cleanup_skip_on_duplicate(self, mock_master_url):
clusters = ib.Clusters()
project = 'test-project'
region = 'test-region'
p1 = beam.Pipeline(options=PipelineOptions(
project=project,
region=region,
))
p2 = beam.Pipeline(options=PipelineOptions(
project=project,
region=region,
))
cluster_metadata_1 = MasterURLIdentifier(project_id=project,
region=region)
clusters.dataproc_cluster_managers[str(
id(p1))] = DataprocClusterManager(cluster_metadata_1)
clusters.dataproc_cluster_managers[str(id(p1))].master_url = 'test_url'
clusters.master_urls_to_pipelines['test_url'].append(str(id(p1)))
cluster_metadata_2 = MasterURLIdentifier(project_id=project,
region=region)
clusters.dataproc_cluster_managers[str(
id(p1))] = DataprocClusterManager(cluster_metadata_2)
clusters.dataproc_cluster_managers[str(id(p1))].master_url = 'test_url'
clusters.master_urls_to_pipelines['test_url'].append(str(id(p2)))
from apache_beam.runners.interactive.interactive_beam import _LOGGER
with self.assertLogs(_LOGGER, level='WARNING') as context_manager:
clusters.cleanup(p1)
self.assertTrue('skipping deletion' in context_manager.output[0])
def test_delete_cluster(self):
clusters = ie.current_env().clusters
class MockClusterManager:
master_url = 'test-url'
def cleanup(self):
pass
master_url = 'test-url'
cluster_name = 'test-cluster'
project = 'test-project'
region = 'test-region'
metadata = MasterURLIdentifier(project, region, cluster_name)
p = beam.Pipeline(ir.InteractiveRunner())
ie.current_env()._tracked_user_pipelines.add_user_pipeline(p)
clusters.master_urls[master_url] = metadata
clusters.master_urls_to_dashboards[master_url] = 'test-dashboard'
clusters.dataproc_cluster_managers[str(id(p))] = MockClusterManager()
clusters.master_urls_to_pipelines[master_url] = [str(id(p))]
cluster_id = obfuscate(project, region, cluster_name)
ie.current_env().inspector._clusters[cluster_id] = {
'master_url': master_url,
'pipelines': [str(id(p))]
}
clusters.delete_cluster(
ie.current_env().inspector.get_cluster_master_url(cluster_id))
self.assertEqual(clusters.master_urls, {})
self.assertEqual(clusters.master_urls_to_pipelines, {})
def test_list_clusters(self):
master_url = 'test-url'
cluster_name = 'test-cluster'
project = 'test-project'
region = 'test-region'
pipelines = ['pid']
dashboard = 'test-dashboard'
ie.current_env(
).clusters.master_urls[master_url] = MasterURLIdentifier(
project, region, cluster_name)
ie.current_env(
).clusters.master_urls_to_pipelines[master_url] = pipelines
ie.current_env(
).clusters.master_urls_to_dashboards[master_url] = dashboard
ins = inspector.InteractiveEnvironmentInspector()
cluster_id = obfuscate(project, region, cluster_name)
self.assertEqual(
{
cluster_id: {
'cluster_name': cluster_name,
'project': project,
'region': region,
'master_url': master_url,
'dashboard': dashboard,
'pipelines': pipelines
}
}, json.loads(ins.list_clusters()))
def test_get_staging_location_exception(self, mock_cluster_client):
"""
Test to catch when an error is raised inside get_staging_location.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region',
cluster_name='test-cluster')
cluster_manager = DataprocClusterManager(cluster_metadata)
with self.assertRaises(MockException):
cluster_manager.get_staging_location(cluster_metadata)
def test_get_staging_location(self, mock_cluster_client, mock_list):
"""
Test to receive a mock staging location successfully under
get_staging_location.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region',
cluster_name='test-cluster')
cluster_manager = DataprocClusterManager(cluster_metadata)
self.assertEqual(
cluster_manager.get_staging_location(cluster_metadata),
'gs://test-bucket/google-cloud-dataproc-metainfo/')
def test_create_cluster_default_already_exists(self, mock_cluster_client):
"""
Tests that no exception is thrown when a cluster already exists,
but is using ie.current_env().clusters.default_cluster_name.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
from apache_beam.runners.interactive.dataproc.dataproc_cluster_manager import _LOGGER
with self.assertLogs(_LOGGER, level='INFO') as context_manager:
cluster_manager.create_cluster({})
self.assertTrue('already exists' in context_manager.output[0])
def test_create_cluster_permission_denied(self, mock_cluster_client):
"""
Tests that an exception is thrown when a user is trying to write to
a project while having insufficient permissions.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
from apache_beam.runners.interactive.dataproc.dataproc_cluster_manager import _LOGGER
with self.assertLogs(_LOGGER, level='ERROR') as context_manager:
self.assertRaises(ValueError, cluster_manager.create_cluster, {})
self.assertTrue('Due to insufficient project permissions' in
context_manager.output[0])
def test_create_cluster_region_does_not_exist(self, mock_cluster_client):
"""
Tests that an exception is thrown when a user specifies a region
that does not exist.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
from apache_beam.runners.interactive.dataproc.dataproc_cluster_manager import _LOGGER
with self.assertLogs(_LOGGER, level='ERROR') as context_manager:
self.assertRaises(ValueError, cluster_manager.create_cluster, {})
self.assertTrue(
'Invalid region provided' in context_manager.output[0])
def test_cleanup_does_not_exist(self, mock_cluster_client, mock_cleanup):
"""
Tests that an exception is thrown when cleanup attempts to delete
a cluster that does not exist.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
from apache_beam.runners.interactive.dataproc.dataproc_cluster_manager import _LOGGER
with self.assertLogs(_LOGGER, level='ERROR') as context_manager:
self.assertRaises(ValueError, cluster_manager.cleanup)
self.assertTrue(
'Cluster does not exist' in context_manager.output[0])
def test_cleanup_other_exception(self, mock_cluster_client, mock_cleanup):
"""
Tests that an exception is thrown when the exception is not handled by
any other case under cleanup.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
from apache_beam.runners.interactive.dataproc.dataproc_cluster_manager import _LOGGER
with self.assertLogs(_LOGGER, level='ERROR') as context_manager:
self.assertRaises(MockException, cluster_manager.cleanup)
self.assertTrue(
'Failed to delete cluster' in context_manager.output[0])
def test_get_master_url_and_dashboard(self, mock_parse_method):
"""
Tests that get_master_url_and_dashboard detect the line containing the
unique substring which identifies the location of the master_url and
application id of the Flink master.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
cluster_manager._fs = MockFileSystem()
master_url, dashboard = cluster_manager.get_master_url_and_dashboard(
cluster_metadata, 'test-staging-bucket')
self.assertEqual(master_url, 'test-master-url')
self.assertEqual(dashboard, 'test-dashboard-link')
def test_clusters_cleanup_otherwise(self, mock_cleanup):
clusters = ie.current_env().clusters
project = 'test-project'
region = 'test-region'
p = beam.Pipeline(options=PipelineOptions(
project=project,
region=region,
))
cluster_metadata = MasterURLIdentifier(project_id=project,
region=region)
clusters.dataproc_cluster_managers[str(
id(p))] = DataprocClusterManager(cluster_metadata)
clusters.dataproc_cluster_managers[str(id(p))].master_url = 'test_url'
clusters.cleanup(p)
def test_get_cluster_details_permission_denied(self, mock_cluster_client):
"""
Tests that an exception is thrown when a user is trying to get information
for a project without sufficient permissions to do so.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
from apache_beam.runners.interactive.dataproc.dataproc_cluster_manager import _LOGGER
with self.assertLogs(
_LOGGER, level='ERROR') as context_manager, self.assertRaises(
ValueError):
cluster_manager.get_cluster_details(cluster_metadata)
self.assertTrue('Due to insufficient project permissions' in
context_manager.output[0])
def test_cleanup_all_dataproc_clusters(self, mock_cleanup):
env = ie.InteractiveEnvironment()
project = 'test-project'
region = 'test-region'
p = beam.Pipeline(options=PipelineOptions(
project=project,
region=region,
))
cluster_metadata = MasterURLIdentifier(project_id=project,
region=region)
env.clusters.dataproc_cluster_managers[str(
id(p))] = DataprocClusterManager(cluster_metadata)
env._tracked_user_pipelines.add_user_pipeline(p)
env.cleanup()
self.assertEqual(env.clusters.dataproc_cluster_managers, {})
def test_clusters_describe(self):
clusters = ib.Clusters()
project = 'test-project'
region = 'test-region'
p = beam.Pipeline(options=PipelineOptions(
project=project,
region=region,
))
cluster_metadata = MasterURLIdentifier(project_id=project,
region=region)
clusters.dataproc_cluster_managers[str(
id(p))] = DataprocClusterManager(cluster_metadata)
self.assertEqual(
'test-project',
clusters.describe()[str(id(p))]['cluster_metadata'].project_id)
def test_parse_master_url_and_dashboard(self, mock_cluster_details):
"""
Tests that parse_master_url_and_dashboard properly parses the input
string and produces a mock master_url and mock dashboard link.
"""
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region')
cluster_manager = DataprocClusterManager(cluster_metadata)
line = 'test-line Found Web Interface test-master-url' \
' of application \'test-app-id\'.\n'
master_url, dashboard = cluster_manager.parse_master_url_and_dashboard(
cluster_metadata, line)
self.assertEqual('test-master-url', master_url)
self.assertEqual(
'test-resource-manager/gateway/default/yarn/proxy/test-app-id/',
dashboard)
def test_get_master_url_no_flink_master_and_master_url_exists(self, m_env):
clusters = ib.Clusters()
m_env().clusters = clusters
from apache_beam.runners.portability.flink_runner import FlinkRunner
runner = interactive_runner.InteractiveRunner(
underlying_runner=FlinkRunner())
p = beam.Pipeline(options=PipelineOptions(
project='test-project',
region='test-region',
))
cluster_name = clusters.default_cluster_name
cluster_metadata = MasterURLIdentifier(project_id='test-project',
region='test-region',
cluster_name=cluster_name)
clusters.master_urls['test-url'] = cluster_metadata
clusters.master_urls_to_dashboards['test-url'] = 'test-dashboard'
flink_master = runner._get_dataproc_cluster_master_url_if_applicable(p)
self.assertEqual(
clusters.describe(p)['cluster_metadata'].project_id,
'test-project')
self.assertEqual(flink_master, clusters.describe(p)['master_url'])
def _get_dataproc_cluster_master_url_if_applicable(
self, user_pipeline: beam.Pipeline) -> str:
""" Creates a Dataproc cluster if the provided user_pipeline is running
FlinkRunner and no flink_master_url was provided as an option. A cluster
is not created when a flink_master_url is detected.
Example pipeline options to enable automatic Dataproc cluster creation:
options = PipelineOptions([
'--runner=FlinkRunner',
'--project=my-project',
'--region=my-region',
'--environment_type=DOCKER'
])
Example pipeline options to skip automatic Dataproc cluster creation:
options = PipelineOptions([
'--runner=FlinkRunner',
'--flink_master=example.internal:41979',
'--environment_type=DOCKER'
])
"""
from apache_beam.runners.interactive.dataproc.dataproc_cluster_manager import DataprocClusterManager
from apache_beam.runners.portability.flink_runner import FlinkRunner
flink_master = user_pipeline.options.view_as(
FlinkRunnerOptions).flink_master
clusters = ie.current_env().clusters
# Only consider this logic when both below 2 conditions apply.
if isinstance(self._underlying_runner,
FlinkRunner) and clusters.dataproc_cluster_managers.get(
str(id(user_pipeline)), None) is None:
if flink_master == '[auto]':
# The above condition is True when the user has not provided a
# flink_master.
if ie.current_env()._is_in_ipython:
warnings.filterwarnings(
'ignore',
'options is deprecated since First stable release. References to '
'<pipeline>.options will not be supported',
category=DeprecationWarning)
project_id = (
user_pipeline.options.view_as(GoogleCloudOptions).project)
region = (
user_pipeline.options.view_as(GoogleCloudOptions).region)
if not project_id:
# When a Google Cloud project is not specified, we try to set the
# cluster_metadata to be the default value set from the
# 'Manage Clusters' JupyterLab extension. If a value has not been
# specified, this value defaults to None.
cluster_metadata = ie.current_env(
).clusters.default_cluster_metadata
else:
cluster_name = ie.current_env(
).clusters.default_cluster_name
cluster_metadata = MasterURLIdentifier(
project_id=project_id,
region=region,
cluster_name=cluster_name)
else:
cluster_metadata = clusters.master_urls.get(flink_master, None)
# else noop, no need to log anything because we allow a master_url
# (not managed by us) provided by the user.
if cluster_metadata:
# create the cluster_manager and populate dicts in the clusters
# instance if the pipeline is not already mapped to an existing
# cluster_manager.
cluster_manager = DataprocClusterManager(cluster_metadata)
cluster_manager.create_flink_cluster()
clusters.master_urls[
cluster_manager.master_url] = cluster_metadata
clusters.dataproc_cluster_managers[str(
id(user_pipeline))] = cluster_manager
clusters.master_urls_to_pipelines[
cluster_manager.master_url].append(str(id(user_pipeline)))
clusters.master_urls_to_dashboards[
cluster_manager.master_url] = cluster_manager.dashboard
return cluster_manager.master_url
