def teardown(self):
root_pg, _, flow_id = nf.set_environment()
canvas.schedule_process_group(root_pg.id, False)
while True:
failed = False
for controller in canvas.list_all_controllers(root_pg.id):
try:
canvas.schedule_controller(controller, False)
LOG.debug('Controller %s stopped.',
controller.component.name)
except ApiException as exc:
if exc.status == 409 and 'is referenced by' in exc.body:
LOG.debug(
'Controller %s failed to stop. Will retry later.',
controller.component.name)
failed = True
if not failed:
break
nf.delete_all(root_pg)
efm.delete_all(flow_id)
schreg.delete_all_schemas()
reg_client = versioning.get_registry_client('NiFi Registry')
if reg_client:
versioning.delete_registry_client(reg_client)
nifireg.delete_flows('SensorFlows')
kudu.drop_table()
def lab4_nifi_flow(env):
LOG.info("Running step4_nifi_flow")
# Create a bucket in NiFi Registry to save the edge flow versions
env.sensor_bucket = versioning.get_registry_bucket('SensorFlows')
if not env.sensor_bucket:
env.sensor_bucket = versioning.create_registry_bucket('SensorFlows')
# Create NiFi Process Group
env.reg_client = versioning.create_registry_client('NiFi Registry', _NIFIREG_URL, 'The registry...')
env.sensor_pg = canvas.create_process_group(env.root_pg, PG_NAME, (330, 350))
#env.sensor_flow = versioning.save_flow_ver(env.sensor_pg, env.reg_client, env.sensor_bucket, flow_name='SensorProcessGroup', comment='Enabled version control - ' + env.run_id)
env.sensor_flow = save_flow_ver(env.sensor_pg, env.reg_client, env.sensor_bucket, flow_name='SensorProcessGroup', comment='Enabled version control - ' + str(env.run_id))
# Create controller services
env.sr_svc = create_controller(env.sensor_pg, 'org.apache.nifi.schemaregistry.hortonworks.HortonworksSchemaRegistry', {'url': _SCHREG_API_URL}, True)
env.json_reader_svc = create_controller(env.sensor_pg, 'org.apache.nifi.json.JsonTreeReader', {'schema-access-strategy': 'schema-name', 'schema-registry': env.sr_svc.id}, True)
env.json_writer_svc = create_controller(env.sensor_pg, 'org.apache.nifi.json.JsonRecordSetWriter', {'schema-access-strategy': 'schema-name', 'schema-registry': env.sr_svc.id, 'Schema Write Strategy': 'hwx-schema-ref-attributes'}, True)
# Create flow
sensor_port = canvas.create_port(env.sensor_pg.id, 'INPUT_PORT', 'Sensor Data', 'RUNNING', (0, 0))
upd_attr = create_processor(env.sensor_pg, 'Set Schema Name', 'org.apache.nifi.processors.attributes.UpdateAttribute', (0, 100),
{
'properties': {
'schema.name': 'SensorReading',
},
}
)
canvas.create_connection(sensor_port, upd_attr)
pub_kafka = create_processor(env.sensor_pg, 'Publish to Kafka topic: iot', 'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_0', (0, 300),
{
'properties': {
'bootstrap.servers': 'edge2ai-1.dim.local:9092',
'topic': 'iot',
'record-reader': env.json_reader_svc.id,
'record-writer': env.json_writer_svc.id,
'use-transactions': 'false',
'attribute-name-regex': 'schema.*',
'client.id': PRODUCER_CLIENT_ID,
},
'autoTerminatedRelationships': ['success'],
}
)
canvas.create_connection(upd_attr, pub_kafka, ['success'])
fail_funnel = create_funnel(env.sensor_pg.id, (600, 343))
canvas.create_connection(pub_kafka, fail_funnel, ['failure'])
# Commit changes
#versioning.save_flow_ver(env.sensor_pg, env.reg_client, env.sensor_bucket, flow_id=env.sensor_flow.version_control_information.flow_id, comment='First version - ' + env.run_id)
save_flow_ver(env.sensor_pg, env.reg_client, env.sensor_bucket, flow_id=env.sensor_flow.version_control_information.flow_id, comment='First version - ' + str(env.run_id))
# Start flow
canvas.schedule_process_group(env.root_pg.id, True)
# Update "from Gateway" input port to connect to the process group
update_connection(env.from_gw, env.temp_funnel, sensor_port)
def test_schedule_process_group(fixture_processor, fixture_pg):
test_pg = fixture_pg.generate()
_ = fixture_processor.generate(parent_pg=test_pg)
r1 = canvas.schedule_process_group(test_pg.id, 'RUNNING')
assert r1.state == 'RUNNING'
r2 = canvas.schedule_process_group(test_pg.id, 'STOPPED')
assert r2.state == 'STOPPED'
with pytest.raises(ValueError):
_ = canvas.schedule_process_group(test_pg.id, 'BANANA')
def test_delete_process_group(fixture_pg, regress, fixture_processor):
# Delete stopped PG
pg_1 = fixture_pg.generate()
r1 = canvas.delete_process_group(pg_1.id, pg_1.revision)
assert r1.id == pg_1.id
assert r1.status is None
# Test deleting a running PG
pg_2 = fixture_pg.generate()
p_1 = fixture_processor.generate(parent_pg=pg_2)
canvas.schedule_process_group(pg_2.id, 'RUNNING')
with pytest.raises(ValueError):
_ = canvas.delete_process_group(pg_2.id, pg_2.revision)
def setUpClass(cls):
super(Test1ToNTest, cls).setUpClass()
print("Start of tests: preparing nifi objects")
config.nifi_config.host = 'http://192.168.56.5:8080/nifi-api'
flow_name = "Test1ToNTest"
nav = CanvasNavigator()
# Delete all leftovers from previous (failed?) tests
pgs_to_be_deleted = nav.groups(flow_name)
for pg in pgs_to_be_deleted:
canvas.delete_process_group(pg, force=True)
# Create new process group in root
Test1ToNTest.pg_test = canvas.create_process_group(nav.current, flow_name, (0, 0))
# Create simple flow to test
Test1ToNTest.proc_start = canvas.create_processor(
Test1ToNTest.pg_test,
canvas.get_processor_type("GenerateFlowFile"),
CANVAS_CENTER,
"Start")
Test1ToNTest.proc_2 = canvas.create_processor(
Test1ToNTest.pg_test,
canvas.get_processor_type("DebugFlow"),
CANVAS_CENTER,
"Processor 2")
Test1ToNTest.proc_3 = canvas.create_processor(
Test1ToNTest.pg_test,
canvas.get_processor_type("DebugFlow"),
CANVAS_CENTER,
"Processor 3")
Test1ToNTest.proc_end_1 = canvas.create_processor(
Test1ToNTest.pg_test,
canvas.get_processor_type("DebugFlow"),
CANVAS_CENTER,
"End 1")
Test1ToNTest.proc_end_2 = canvas.create_processor(
Test1ToNTest.pg_test,
canvas.get_processor_type("DebugFlow"),
CANVAS_CENTER,
"End 2")
canvas.update_processor(Test1ToNTest.proc_end_1,
nifi.ProcessorConfigDTO(auto_terminated_relationships=["success", "failure"]))
canvas.update_processor(Test1ToNTest.proc_end_2,
nifi.ProcessorConfigDTO(auto_terminated_relationships=["success", "failure"]))
Test1ToNTest.conn_1 = canvas.create_connection(Test1ToNTest.proc_start, Test1ToNTest.proc_2, ["success"])
Test1ToNTest.conn_2 = canvas.create_connection(Test1ToNTest.proc_2, Test1ToNTest.proc_3, ["success", "failure"])
Test1ToNTest.conn_3 = canvas.create_connection(Test1ToNTest.proc_3, Test1ToNTest.proc_end_1, ["success", "failure"])
Test1ToNTest.conn_4 = canvas.create_connection(Test1ToNTest.proc_3, Test1ToNTest.proc_end_2, ["success", "failure"])
canvas.schedule_process_group(Test1ToNTest.pg_test.component.id, scheduled=True)
def test_schedule_process_group(fix_proc, fix_pg):
f_pg = fix_pg.generate()
_ = fix_proc.generate(parent_pg=f_pg)
r1 = canvas.schedule_process_group(f_pg.id, True)
status = canvas.get_process_group(f_pg.id, 'id')
assert r1 is True
assert status.running_count == 1
r2 = canvas.schedule_process_group(f_pg.id, False)
assert r2 is True
status = canvas.get_process_group(f_pg.id, 'id')
assert status.running_count == 0
assert status.stopped_count == 1
with pytest.raises(AssertionError):
_ = canvas.schedule_process_group(f_pg.id, 'BANANA')
def test_delete_process_group(regress_nifi, fix_pg, fix_proc):
# Delete stopped PG
f_pg1 = fix_pg.generate()
r1 = canvas.delete_process_group(f_pg1)
assert r1.id == f_pg1.id
assert r1.status is None
# Test deleting a running PG
pg_2 = fix_pg.generate()
_ = fix_proc.generate(parent_pg=pg_2)
canvas.schedule_process_group(pg_2.id, True)
with pytest.raises(ValueError):
_ = canvas.delete_process_group(pg_2)
# Once more with feeling
r2 = canvas.delete_process_group(pg_2, force=True)
assert r2.status is None
def test_complex_template_versioning(regress_flow_reg, fix_ctv):
# There is a complex bug where a new flow version cannot be switched to
# and generates a NiFi NPE if attempted when create_flow_version is used
# BUG FIXED: issue with variable name found in Swagger definition
# https://github.com/apache/nifi/pull/2479#issuecomment-366582829
# Create a new flow version
vers = versioning.list_flow_versions(fix_ctv.bucket.identifier,
fix_ctv.flow.identifier)
assert vers[0].version == 2
# create_flow_version is the problem
new_ss = versioning.create_flow_version(
flow=fix_ctv.flow,
flow_snapshot=fix_ctv.snapshot_w_template,
refresh=True)
assert isinstance(new_ss, registry.VersionedFlowSnapshot)
vers = versioning.list_flow_versions(fix_ctv.bucket.identifier,
fix_ctv.flow.identifier)
assert vers[0].version == 3
new_ver_info = versioning.get_version_info(fix_ctv.pg)
r1 = versioning.update_flow_ver(fix_ctv.pg)
assert r1.request.complete is True
assert r1.request.failure_reason is None
r2 = canvas.schedule_process_group(fix_ctv.pg.id, True)
status = canvas.get_process_group(fix_ctv.pg.id, 'id')
assert status.running_count >= 5
with pytest.raises(ValueError):
_ = versioning.update_flow_ver(fix_ctv.pg, 'bob')
with pytest.raises(ValueError):
_ = versioning.update_flow_ver(fix_ctv.pg, '9999999')
def global_teardown(run_id=None,
schema_text=None,
cdsw_flag=True,
target_lab=99):
(run_id, root_pg, efm_pg_id, flow_id) = set_environment(run_id)
canvas.schedule_process_group(root_pg.id, False)
while True:
failed = False
for controller in canvas.list_all_controllers(root_pg.id):
try:
canvas.schedule_controller(controller, False)
LOG.debug('Controller %s stopped.', controller.component.name)
except ApiException as exc:
if exc.status == 409 and 'is referenced by' in exc.body:
LOG.debug(
'Controller %s failed to stop. Will retry later.',
controller.component.name)
failed = True
if not failed:
break
nifi_delete_all(root_pg)
class _Env(object):
pass
env = _Env()
env.run_id, env.root_pg, env.efm_pg_id, env.flow_id = set_environment(
run_id)
env.schema_text = schema_text if schema_text is not None else read_in_schema(
)
LOG.info("Using Schema: %s", schema_text)
env.cdsw_flag = cdsw_flag
LOG.info("Build just enough")
lab2_edge_flow(env)
lab3_register_schema(env)
lab6_expand_edge_flow(env)
create_kudu_table()
LOG.info("Created edge, schema, kudu")
def nifi_delete_all(pg):
canvas.schedule_process_group(pg.id, False)
for conn in canvas.list_all_connections(pg.id):
LOG.debug('Connection: ' + conn.id)
canvas.delete_connection(conn, purge=True)
for input_port in canvas.list_all_input_ports(pg.id):
LOG.debug('Input Port: ' + input_port.id)
canvas.delete_port(input_port)
for output_port in canvas.list_all_output_ports(pg.id):
LOG.debug('Output Port: ' + output_port.id)
canvas.delete_port(output_port)
for funnel in canvas.list_all_funnels(pg.id):
LOG.debug('Funnel: ' + funnel.id)
canvas.delete_funnel(funnel)
for processor in canvas.list_all_processors(pg.id):
LOG.debug('Processor: ' + processor.id)
canvas.delete_processor(processor, force=True)
for process_group in canvas.list_all_process_groups(pg.id):
if pg.id == process_group.id:
continue
LOG.debug('Process Group: ' + process_group.id)
nifi_delete_all(process_group)
canvas.delete_process_group(process_group, force=True)
def lab7_rest_and_kudu(env):
LOG.info("Running step7_rest_and_kudu")
# Create controllers
env.json_reader_with_schema_svc = create_controller(
env.sensor_pg,
'org.apache.nifi.json.JsonTreeReader', {
'schema-access-strategy': 'hwx-schema-ref-attributes',
'schema-registry': env.sr_svc.id
},
True,
name='JsonTreeReader - With schema identifier')
props = {
'rest-lookup-url': get_cdsw_altus_api() + '/models/call-model',
'rest-lookup-record-reader': env.json_reader_svc.id,
'rest-lookup-record-path': '/response'
}
if _IS_TLS_ENABLED:
props.update({
'rest-lookup-ssl-context-service': env.ssl_svc.id,
})
rest_lookup_svc = create_controller(
env.sensor_pg, 'org.apache.nifi.lookup.RestLookupService', props, True)
# Build flow
fail_funnel = create_funnel(env.sensor_pg.id, (1400, 340))
props = {
'topic': 'iot',
'topic_type': 'names',
'record-reader': env.json_reader_with_schema_svc.id,
'record-writer': env.json_writer_svc.id,
}
props.update(_get_common_kafka_client_properties(env, 'consumer'))
consume_kafka = create_processor(
env.sensor_pg, 'Consume Kafka iot messages',
'org.apache.nifi.processors.kafka.pubsub.ConsumeKafkaRecord_2_0',
(700, 0), {'properties': props})
canvas.create_connection(consume_kafka, fail_funnel, ['parse.failure'])
predict = create_processor(
env.sensor_pg, 'Predict machine health',
'org.apache.nifi.processors.standard.LookupRecord', (700, 200), {
'properties': {
'record-reader': env.json_reader_with_schema_svc.id,
'record-writer': env.json_writer_svc.id,
'lookup-service': rest_lookup_svc.id,
'result-record-path': '/response',
'routing-strategy': 'route-to-success',
'result-contents': 'insert-entire-record',
'mime.type': "toString('application/json', 'UTF-8')",
'request.body':
"concat('{\"accessKey\":\"', '${cdsw.access.key}', '\",\"request\":{\"feature\":\"', /sensor_0, ', ', /sensor_1, ', ', /sensor_2, ', ', /sensor_3, ', ', /sensor_4, ', ', /sensor_5, ', ', /sensor_6, ', ', /sensor_7, ', ', /sensor_8, ', ', /sensor_9, ', ', /sensor_10, ', ', /sensor_11, '\"}}')",
'request.method': "toString('post', 'UTF-8')",
},
})
canvas.create_connection(predict, fail_funnel, ['failure'])
canvas.create_connection(consume_kafka, predict, ['success'])
update_health = create_processor(
env.sensor_pg, 'Update health flag',
'org.apache.nifi.processors.standard.UpdateRecord', (700, 400), {
'properties': {
'record-reader': env.json_reader_with_schema_svc.id,
'record-writer': env.json_writer_svc.id,
'replacement-value-strategy': 'record-path-value',
'/is_healthy': '/response/result',
},
})
canvas.create_connection(update_health, fail_funnel, ['failure'])
canvas.create_connection(predict, update_health, ['success'])
if get_kudu_version() >= (1, 14):
kudu_table_name = 'default.sensors'
else:
kudu_table_name = 'impala::default.sensors'
write_kudu = create_processor(
env.sensor_pg, 'Write to Kudu',
'org.apache.nifi.processors.kudu.PutKudu', (700, 600), {
'properties': {
'Kudu Masters':
_HOSTNAME + ':7051',
'Table Name':
kudu_table_name,
'record-reader':
env.json_reader_with_schema_svc.id,
'kerberos-credentials-service':
env.keytab_svc.id if _IS_TLS_ENABLED else None,
},
})
canvas.create_connection(write_kudu, fail_funnel, ['failure'])
canvas.create_connection(update_health, write_kudu, ['success'])
props = {
'topic': 'iot_enriched',
'record-reader': env.json_reader_with_schema_svc.id,
'record-writer': env.json_writer_svc.id,
}
props.update(_get_common_kafka_client_properties(env, 'producer'))
pub_kafka_enriched = create_processor(
env.sensor_pg, 'Publish to Kafka topic: iot_enriched',
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_0',
(300, 600), {
'properties': props,
'autoTerminatedRelationships': ['success', 'failure'],
})
canvas.create_connection(update_health, pub_kafka_enriched, ['success'])
props = {
'topic': 'iot_enriched_avro',
'record-reader': env.json_reader_with_schema_svc.id,
'record-writer': env.avro_writer_svc.id,
}
props.update(_get_common_kafka_client_properties(env, 'producer'))
pub_kafka_enriched_avro = create_processor(
env.sensor_pg, 'Publish to Kafka topic: iot_enriched_avro',
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_0',
(-100, 600), {
'properties': props,
'autoTerminatedRelationships': ['success', 'failure'],
})
canvas.create_connection(update_health, pub_kafka_enriched_avro,
['success'])
monitor_activity = create_processor(
env.sensor_pg, 'Monitor Activity',
'org.apache.nifi.processors.standard.MonitorActivity', (700, 800), {
'properties': {
'Threshold Duration': '45 secs',
'Continually Send Messages': 'true',
},
'autoTerminatedRelationships': ['activity.restored', 'success'],
})
canvas.create_connection(monitor_activity, fail_funnel, ['inactive'])
canvas.create_connection(write_kudu, monitor_activity, ['success'])
# Version flow
save_flow_ver(env.sensor_pg,
env.reg_client,
env.sensor_bucket,
flow_id=env.sensor_flow.version_control_information.flow_id,
comment='Second version - ' + str(env.run_id))
# Prepare Impala/Kudu table
create_kudu_table()
# Set the variable with the CDSW access key
if env.cdsw_flag:
canvas.update_variable_registry(
env.sensor_pg, [('cdsw.access.key', get_cdsw_model_access_key())])
# Start everything
canvas.schedule_process_group(env.root_pg.id, True)
def __stop_base(self):
canvas.schedule_process_group(self.base.component.id, False)
def tearDownClass(cls):
print("Tests done: deleting nifi objects")
canvas.schedule_process_group(Test1ToNTest.pg_test.component.id, scheduled=False)
def lab4_rest_and_kudu(self):
# Prepare Impala/Kudu table
kudu.create_table()
kudu_table_name = kudu.get_kudu_table_name('default', 'sensors')
# Set required variables
if not self.context.skip_cdsw:
# Set the variable with the CDSW access key
canvas.update_variable_registry(
self.context.sensor_pg,
[('cdsw.access.key', cdsw.get_model_access_key())])
# Set the variable with the CDSW model API key
canvas.update_variable_registry(
self.context.sensor_pg,
[('cdsw.model.api.key', cdsw.create_model_api_key())])
# Create controllers
self.context.json_reader_with_schema_svc = nf.create_controller(
self.context.sensor_pg,
'org.apache.nifi.json.JsonTreeReader', {
'schema-access-strategy': 'hwx-schema-ref-attributes',
'schema-registry': self.context.sr_svc.id
},
True,
name='JsonTreeReader - With schema identifier')
props = {
'rest-lookup-url': cdsw.get_model_endpoint_url(),
'rest-lookup-record-reader': self.context.json_reader_svc.id,
'rest-lookup-record-path': '/response',
'Authorization': 'Bearer ${cdsw.model.api.key}',
}
if is_tls_enabled():
props.update({
'rest-lookup-ssl-context-service':
self.context.ssl_svc.id,
})
rest_lookup_svc = nf.create_controller(
self.context.sensor_pg, 'org.apache.nifi.lookup.RestLookupService',
props, True)
# Build flow
fail_funnel = nf.create_funnel(self.context.sensor_pg.id, (1400, 340))
props = {
'topic': 'iot',
'topic_type': 'names',
'record-reader': self.context.json_reader_with_schema_svc.id,
'record-writer': self.context.json_writer_svc.id,
}
props.update(
kafka.get_common_client_properties(self.context, 'consumer',
CONSUMER_GROUP_ID,
PRODUCER_CLIENT_ID))
consume_kafka = nf.create_processor(
self.context.sensor_pg, 'Consume Kafka iot messages', [
'org.apache.nifi.processors.kafka.pubsub.ConsumeKafkaRecord_2_6',
'org.apache.nifi.processors.kafka.pubsub.ConsumeKafkaRecord_2_0'
], (700, 0), {'properties': props})
canvas.create_connection(consume_kafka, fail_funnel, ['parse.failure'])
predict = nf.create_processor(
self.context.sensor_pg, 'Predict machine health',
'org.apache.nifi.processors.standard.LookupRecord', (700, 200), {
'properties': {
'record-reader':
self.context.json_reader_with_schema_svc.id,
'record-writer':
self.context.json_writer_svc.id,
'lookup-service':
rest_lookup_svc.id,
'result-record-path':
'/response',
'routing-strategy':
'route-to-success',
'result-contents':
'insert-entire-record',
'mime.type':
"toString('application/json', 'UTF-8')",
'request.body':
"concat('{\"accessKey\":\"', '${cdsw.access.key}', "
"'\",\"request\":{\"feature\":\"', /sensor_0, ', ', "
"/sensor_1, ', ', /sensor_2, ', ', /sensor_3, ', ', "
"/sensor_4, ', ', /sensor_5, ', ', /sensor_6, ', ', "
"/sensor_7, ', ', /sensor_8, ', ', /sensor_9, ', ', "
"/sensor_10, ', ', /sensor_11, '\"}}')",
'request.method':
"toString('post', 'UTF-8')",
},
})
canvas.create_connection(predict, fail_funnel, ['failure'])
canvas.create_connection(consume_kafka, predict, ['success'])
update_health = nf.create_processor(
self.context.sensor_pg, 'Update health flag',
'org.apache.nifi.processors.standard.UpdateRecord', (700, 400), {
'properties': {
'record-reader':
self.context.json_reader_with_schema_svc.id,
'record-writer': self.context.json_writer_svc.id,
'replacement-value-strategy': 'record-path-value',
'/is_healthy': '/response/result',
},
})
canvas.create_connection(update_health, fail_funnel, ['failure'])
canvas.create_connection(predict, update_health, ['success'])
write_kudu = nf.create_processor(
self.context.sensor_pg, 'Write to Kudu',
'org.apache.nifi.processors.kudu.PutKudu', (700, 600), {
'properties': {
'Kudu Masters':
get_hostname() + ':7051',
'Table Name':
kudu_table_name,
'record-reader':
self.context.json_reader_with_schema_svc.id,
'kerberos-credentials-service':
self.context.keytab_svc.id if is_tls_enabled() else None,
},
})
canvas.create_connection(write_kudu, fail_funnel, ['failure'])
canvas.create_connection(update_health, write_kudu, ['success'])
props = {
'topic': 'iot_enriched',
'record-reader': self.context.json_reader_with_schema_svc.id,
'record-writer': self.context.json_writer_svc.id,
}
props.update(
kafka.get_common_client_properties(self.context, 'producer',
CONSUMER_GROUP_ID,
PRODUCER_CLIENT_ID))
pub_kafka_enriched = nf.create_processor(
self.context.sensor_pg, 'Publish to Kafka topic: iot_enriched', [
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_6',
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_0'
], (300, 600), {
'properties': props,
'autoTerminatedRelationships': ['success', 'failure'],
})
canvas.create_connection(update_health, pub_kafka_enriched,
['success'])
props = {
'topic': 'iot_enriched_avro',
'record-reader': self.context.json_reader_with_schema_svc.id,
'record-writer': self.context.avro_writer_svc.id,
}
props.update(
kafka.get_common_client_properties(self.context, 'producer',
CONSUMER_GROUP_ID,
PRODUCER_CLIENT_ID))
pub_kafka_enriched_avro = nf.create_processor(
self.context.sensor_pg,
'Publish to Kafka topic: iot_enriched_avro', [
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_6',
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_0'
], (-100, 600), {
'properties': props,
'autoTerminatedRelationships': ['success', 'failure'],
})
canvas.create_connection(update_health, pub_kafka_enriched_avro,
['success'])
monitor_activity = nf.create_processor(
self.context.sensor_pg, 'Monitor Activity',
'org.apache.nifi.processors.standard.MonitorActivity', (700, 800),
{
'properties': {
'Threshold Duration': '45 secs',
'Continually Send Messages': 'true',
},
'autoTerminatedRelationships':
['activity.restored', 'success'],
})
canvas.create_connection(monitor_activity, fail_funnel, ['inactive'])
canvas.create_connection(write_kudu, monitor_activity, ['success'])
# Version flow
nifireg.save_flow_ver(self.context.sensor_pg,
self.context.reg_client,
self.context.sensor_bucket,
flow_id=self.context.sensor_flow.
version_control_information.flow_id,
comment='Second version - {}'.format(
self.run_id))
# Start everything
canvas.schedule_process_group(self.context.root_pg.id, True)
def lab2_nifi_flow(self):
# Create a bucket in NiFi Registry to save the edge flow versions
self.context.sensor_bucket = versioning.get_registry_bucket(
'SensorFlows')
if not self.context.sensor_bucket:
self.context.sensor_bucket = versioning.create_registry_bucket(
'SensorFlows')
# Create NiFi Process Group
self.context.reg_client = versioning.create_registry_client(
'NiFi Registry', nifireg.get_url(), 'The registry...')
self.context.sensor_pg = canvas.create_process_group(
self.context.root_pg, PG_NAME, (330, 350))
self.context.sensor_flow = nifireg.save_flow_ver(
self.context.sensor_pg,
self.context.reg_client,
self.context.sensor_bucket,
flow_name='SensorProcessGroup',
comment='Enabled version control - {}'.format(self.run_id))
# Update default SSL context controller service
ssl_svc_name = 'Default NiFi SSL Context Service'
if is_tls_enabled():
props = {
'SSL Protocol': 'TLS',
'Truststore Type': 'JKS',
'Truststore Filename':
'/opt/cloudera/security/jks/truststore.jks',
'Truststore Password': get_the_pwd(),
'Keystore Type': 'JKS',
'Keystore Filename': '/opt/cloudera/security/jks/keystore.jks',
'Keystore Password': get_the_pwd(),
'key-password': get_the_pwd(),
}
self.context.ssl_svc = canvas.get_controller(ssl_svc_name, 'name')
if self.context.ssl_svc:
canvas.schedule_controller(self.context.ssl_svc, False)
self.context.ssl_svc = canvas.get_controller(
ssl_svc_name, 'name')
canvas.update_controller(
self.context.ssl_svc,
nifi.ControllerServiceDTO(properties=props))
self.context.ssl_svc = canvas.get_controller(
ssl_svc_name, 'name')
canvas.schedule_controller(self.context.ssl_svc, True)
else:
self.context.keytab_svc = nf.create_controller(
self.context.root_pg,
'org.apache.nifi.ssl.StandardRestrictedSSLContextService',
props,
True,
name=ssl_svc_name)
# Create controller services
if is_tls_enabled():
self.context.ssl_svc = canvas.get_controller(ssl_svc_name, 'name')
props = {
'Kerberos Keytab': '/keytabs/admin.keytab',
'Kerberos Principal': 'admin',
}
self.context.keytab_svc = nf.create_controller(
self.context.sensor_pg,
'org.apache.nifi.kerberos.KeytabCredentialsService', props,
True)
else:
self.context.ssl_svc = None
self.context.keytab_svc = None
props = {
'url': schreg.get_api_url(),
}
if is_tls_enabled():
props.update({
'kerberos-credentials-service':
self.context.keytab_svc.id,
'ssl-context-service':
self.context.ssl_svc.id,
})
self.context.sr_svc = nf.create_controller(
self.context.sensor_pg,
'org.apache.nifi.schemaregistry.hortonworks.HortonworksSchemaRegistry',
props, True)
self.context.json_reader_svc = nf.create_controller(
self.context.sensor_pg, 'org.apache.nifi.json.JsonTreeReader', {
'schema-access-strategy': 'schema-name',
'schema-registry': self.context.sr_svc.id
}, True)
self.context.json_writer_svc = nf.create_controller(
self.context.sensor_pg, 'org.apache.nifi.json.JsonRecordSetWriter',
{
'schema-access-strategy': 'schema-name',
'schema-registry': self.context.sr_svc.id,
'Schema Write Strategy': 'hwx-schema-ref-attributes'
}, True)
self.context.avro_writer_svc = nf.create_controller(
self.context.sensor_pg, 'org.apache.nifi.avro.AvroRecordSetWriter',
{
'schema-access-strategy': 'schema-name',
'schema-registry': self.context.sr_svc.id,
'Schema Write Strategy': 'hwx-content-encoded-schema'
}, True)
# Create flow
sensor_port = canvas.create_port(self.context.sensor_pg.id,
'INPUT_PORT', 'Sensor Data',
'STOPPED', (0, 0))
upd_attr = nf.create_processor(
self.context.sensor_pg, 'Set Schema Name',
'org.apache.nifi.processors.attributes.UpdateAttribute', (0, 100),
{
'properties': {
'schema.name': 'SensorReading',
},
})
canvas.create_connection(sensor_port, upd_attr)
props = {
'topic': 'iot',
'record-reader': self.context.json_reader_svc.id,
'record-writer': self.context.json_writer_svc.id,
}
props.update(
kafka.get_common_client_properties(self.context, 'producer',
CONSUMER_GROUP_ID,
PRODUCER_CLIENT_ID))
pub_kafka = nf.create_processor(
self.context.sensor_pg, 'Publish to Kafka topic: iot', [
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_6',
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_0'
], (0, 300), {
'properties': props,
'autoTerminatedRelationships': ['success'],
})
canvas.create_connection(upd_attr, pub_kafka, ['success'])
fail_funnel = nf.create_funnel(self.context.sensor_pg.id, (600, 343))
canvas.create_connection(pub_kafka, fail_funnel, ['failure'])
# Commit changes
nifireg.save_flow_ver(self.context.sensor_pg,
self.context.reg_client,
self.context.sensor_bucket,
flow_id=self.context.sensor_flow.
version_control_information.flow_id,
comment='First version - {}'.format(self.run_id))
# Start flow
canvas.schedule_process_group(self.context.root_pg.id, True)
# Update "from Gateway" input port to connect to the process group
nf.update_connection(self.context.from_gw, self.context.temp_funnel,
sensor_port)
canvas.schedule_components(self.context.root_pg.id, True,
[sensor_port])
def lab4_nifi_flow(env):
LOG.info("Running step4_nifi_flow")
# Create a bucket in NiFi Registry to save the edge flow versions
env.sensor_bucket = versioning.get_registry_bucket('SensorFlows')
if not env.sensor_bucket:
env.sensor_bucket = versioning.create_registry_bucket('SensorFlows')
# Create NiFi Process Group
env.reg_client = versioning.create_registry_client('NiFi Registry',
_get_nifireg_url(),
'The registry...')
env.sensor_pg = canvas.create_process_group(env.root_pg, PG_NAME,
(330, 350))
#env.sensor_flow = versioning.save_flow_ver(env.sensor_pg, env.reg_client, env.sensor_bucket, flow_name='SensorProcessGroup', comment='Enabled version control - ' + env.run_id)
env.sensor_flow = save_flow_ver(env.sensor_pg,
env.reg_client,
env.sensor_bucket,
flow_name='SensorProcessGroup',
comment='Enabled version control - ' +
str(env.run_id))
# Update default SSL context controller service
ssl_svc_name = 'Default NiFi SSL Context Service'
if _IS_TLS_ENABLED:
props = {
'SSL Protocol': 'TLS',
'Truststore Type': 'JKS',
'Truststore Filename': '/opt/cloudera/security/jks/truststore.jks',
'Truststore Password': _THE_PWD,
'Keystore Type': 'JKS',
'Keystore Filename': '/opt/cloudera/security/jks/keystore.jks',
'Keystore Password': _THE_PWD,
'key-password': _THE_PWD,
}
env.ssl_svc = canvas.get_controller(ssl_svc_name, 'name')
if env.ssl_svc:
canvas.schedule_controller(env.ssl_svc, False)
env.ssl_svc = canvas.get_controller(ssl_svc_name, 'name')
canvas.update_controller(
env.ssl_svc, nifi.ControllerServiceDTO(properties=props))
env.ssl_svc = canvas.get_controller(ssl_svc_name, 'name')
canvas.schedule_controller(env.ssl_svc, True)
else:
env.keytab_svc = create_controller(
env.root_pg,
'org.apache.nifi.ssl.StandardRestrictedSSLContextService',
props,
True,
name=ssl_svc_name)
# Create controller services
if _IS_TLS_ENABLED:
env.ssl_svc = canvas.get_controller(ssl_svc_name, 'name')
props = {
'Kerberos Keytab': '/keytabs/admin.keytab',
'Kerberos Principal': 'admin',
}
env.keytab_svc = create_controller(
env.sensor_pg, 'org.apache.nifi.kerberos.KeytabCredentialsService',
props, True)
else:
env.ssl_svc = None
env.keytab_svc = None
props = {
'url': _get_schreg_api_url(),
}
if _IS_TLS_ENABLED:
props.update({
'kerberos-credentials-service': env.keytab_svc.id,
'ssl-context-service': env.ssl_svc.id,
})
env.sr_svc = create_controller(
env.sensor_pg,
'org.apache.nifi.schemaregistry.hortonworks.HortonworksSchemaRegistry',
props, True)
env.json_reader_svc = create_controller(
env.sensor_pg, 'org.apache.nifi.json.JsonTreeReader', {
'schema-access-strategy': 'schema-name',
'schema-registry': env.sr_svc.id
}, True)
env.json_writer_svc = create_controller(
env.sensor_pg, 'org.apache.nifi.json.JsonRecordSetWriter', {
'schema-access-strategy': 'schema-name',
'schema-registry': env.sr_svc.id,
'Schema Write Strategy': 'hwx-schema-ref-attributes'
}, True)
env.avro_writer_svc = create_controller(
env.sensor_pg, 'org.apache.nifi.avro.AvroRecordSetWriter', {
'schema-access-strategy': 'schema-name',
'schema-registry': env.sr_svc.id,
'Schema Write Strategy': 'hwx-content-encoded-schema'
}, True)
# Create flow
sensor_port = canvas.create_port(env.sensor_pg.id, 'INPUT_PORT',
'Sensor Data', 'RUNNING', (0, 0))
upd_attr = create_processor(
env.sensor_pg, 'Set Schema Name',
'org.apache.nifi.processors.attributes.UpdateAttribute', (0, 100), {
'properties': {
'schema.name': 'SensorReading',
},
})
canvas.create_connection(sensor_port, upd_attr)
props = {
'topic': 'iot',
'record-reader': env.json_reader_svc.id,
'record-writer': env.json_writer_svc.id,
}
props.update(_get_common_kafka_client_properties(env, 'producer'))
pub_kafka = create_processor(
env.sensor_pg, 'Publish to Kafka topic: iot',
'org.apache.nifi.processors.kafka.pubsub.PublishKafkaRecord_2_0',
(0, 300), {
'properties': props,
'autoTerminatedRelationships': ['success'],
})
canvas.create_connection(upd_attr, pub_kafka, ['success'])
fail_funnel = create_funnel(env.sensor_pg.id, (600, 343))
canvas.create_connection(pub_kafka, fail_funnel, ['failure'])
# Commit changes
#versioning.save_flow_ver(env.sensor_pg, env.reg_client, env.sensor_bucket, flow_id=env.sensor_flow.version_control_information.flow_id, comment='First version - ' + env.run_id)
save_flow_ver(env.sensor_pg,
env.reg_client,
env.sensor_bucket,
flow_id=env.sensor_flow.version_control_information.flow_id,
comment='First version - ' + str(env.run_id))
# Start flow
canvas.schedule_process_group(env.root_pg.id, True)
# Update "from Gateway" input port to connect to the process group
update_connection(env.from_gw, env.temp_funnel, sensor_port)
def step7_rest_and_kudu(env):
# Create controllers
json_reader_with_schema_svc = create_controller(
env.sensor_pg,
'org.apache.nifi.json.JsonTreeReader', {
'schema-access-strategy': 'hwx-schema-ref-attributes',
'schema-registry': env.sr_svc.id
},
True,
name='JsonTreeReader - With schema identifier')
rest_lookup_svc = create_controller(
env.sensor_pg, 'org.apache.nifi.lookup.RestLookupService', {
'rest-lookup-url': get_cdsw_altus_api() + '/models/call-model',
'rest-lookup-record-reader': env.json_reader_svc.id,
'rest-lookup-record-path': '/response'
}, True)
# Build flow
fail_funnel = create_funnel(env.sensor_pg.id, (1400, 340))
consume_kafka = create_processor(
env.sensor_pg, 'Consume Kafka iot messages',
'org.apache.nifi.processors.kafka.pubsub.ConsumeKafkaRecord_2_0',
(700, 0), {
'properties': {
'bootstrap.servers': 'edge2ai-1.dim.local:9092',
'topic': 'iot',
'topic_type': 'names',
'record-reader': json_reader_with_schema_svc.id,
'record-writer': env.json_writer_svc.id,
'honor-transactions': 'false',
'group.id': CONSUMER_GROUP_ID,
'auto.offset.reset': 'latest',
'header-name-regex': 'schema.*',
},
})
canvas.create_connection(consume_kafka, fail_funnel, ['parse.failure'])
predict = create_processor(
env.sensor_pg, 'Predict machine health',
'org.apache.nifi.processors.standard.LookupRecord', (700, 200), {
'properties': {
'record-reader': json_reader_with_schema_svc.id,
'record-writer': env.json_writer_svc.id,
'lookup-service': rest_lookup_svc.id,
'result-record-path': '/response',
'routing-strategy': 'route-to-success',
'result-contents': 'insert-entire-record',
'mime.type': "toString('application/json', 'UTF-8')",
'request.body':
"concat('{\"accessKey\":\"', '${cdsw.access.key}', '\",\"request\":{\"feature\":\"', /sensor_0, ', ', /sensor_1, ', ', /sensor_2, ', ', /sensor_3, ', ', /sensor_4, ', ', /sensor_5, ', ', /sensor_6, ', ', /sensor_7, ', ', /sensor_8, ', ', /sensor_9, ', ', /sensor_10, ', ', /sensor_11, '\"}}')",
'request.method': "toString('post', 'UTF-8')",
},
})
canvas.create_connection(predict, fail_funnel, ['failure'])
canvas.create_connection(consume_kafka, predict, ['success'])
update_health = create_processor(
env.sensor_pg, 'Update health flag',
'org.apache.nifi.processors.standard.UpdateRecord', (700, 400), {
'properties': {
'record-reader': json_reader_with_schema_svc.id,
'record-writer': env.json_writer_svc.id,
'replacement-value-strategy': 'record-path-value',
'/is_healthy': '/response/result',
},
})
canvas.create_connection(update_health, fail_funnel, ['failure'])
canvas.create_connection(predict, update_health, ['success'])
write_kudu = create_processor(
env.sensor_pg, 'Write to Kudu',
'org.apache.nifi.processors.kudu.PutKudu', (700, 600), {
'properties': {
'Kudu Masters': 'edge2ai-1.dim.local:7051',
'Table Name': 'impala::default.sensors',
'record-reader': json_reader_with_schema_svc.id,
},
})
canvas.create_connection(write_kudu, fail_funnel, ['failure'])
canvas.create_connection(update_health, write_kudu, ['success'])
monitor_activity = create_processor(
env.sensor_pg, 'Monitor Activity',
'org.apache.nifi.processors.standard.MonitorActivity', (700, 800), {
'properties': {
'Threshold Duration': '45 secs',
'Continually Send Messages': 'true',
},
'autoTerminatedRelationships': ['activity.restored', 'success'],
})
canvas.create_connection(monitor_activity, fail_funnel, ['inactive'])
canvas.create_connection(write_kudu, monitor_activity, ['success'])
# Version flow
#versioning.save_flow_ver(env.sensor_pg, env.reg_client, env.sensor_bucket, flow_id=env.sensor_flow.version_control_information.flow_id, comment='Second version - ' + env.run_id)
save_flow_ver(env.sensor_pg,
env.reg_client,
env.sensor_bucket,
flow_id=env.sensor_flow.version_control_information.flow_id,
comment='Second version - ' + env.run_id)
# Prepare Impala/Kudu table
create_kudu_table()
# Set the variable with the CDSW access key
canvas.update_variable_registry(
env.sensor_pg, [('cdsw.access.key', get_cdsw_model_access_key())])
# Start everything
canvas.schedule_process_group(env.root_pg.id, True)
root_pg = canvas.get_process_group(root_pg_id, 'id')
location_x = 2000
location_y = 2000
location = (location_x, location_y)
template_entity = nipyapi.templates.upload_template(root_pg_id,
'jangowave_demo.xml')
flow = nipyapi.templates.deploy_template(root_pg_id, template_entity.id, 2000,
2000)
jd = canvas.get_process_group('jangowave_demo')
for cs in canvas.list_all_processors(jd.id):
if cs.status.run_status != "ENABLED" and cs.component.name == "RecordIngest":
config_update = nifi.models.processor_config_dto.ProcessorConfigDTO(
properties={
"Accumulo User": "******",
"Instance Name": instance,
"Accumulo Password": "******",
"ZooKeeper Quorum": zookeeper_list
})
canvas.update_processor(cs, config_update)
for cs in canvas.list_all_controllers(jd.id):
canvas.schedule_controller(cs, True)
for cs in canvas.list_all_processors(jd.id):
canvas.schedule_processor(cs, True)
canvas.schedule_process_group(jd.id, True)
def __start_base(self):
canvas.schedule_process_group(self.base.component.id, True)
