def test_minion_perform_deliver_missing_state_invalid_port_failure():
workflow_id = '6666'
app_id = '5001'
job_id = '1'
out_queue = 'queue_50001'
task_id = 'f033f-284ab-28987e-232add'
function_name = 'juicer.spark.transpiler.SparkTranspiler.transpile'
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch(function_name) as mocked_transpile:
with mock.patch('juicer.workflow.workflow.Workflow'
'._build_initial_workflow_graph') as mocked_fn:
mocked_fn.side_effect = lambda: ""
# Setup for mocked_transpile
mocked_transpile.side_effect = get_side_effect(
get_records(), task_id, 0)
redis_conn = mocked_redis()
state_control = StateControlRedis(redis_conn)
data = {
'workflow_id': workflow_id,
'app_id': app_id,
'job_id': job_id,
'type': 'deliver',
'task_id': task_id,
'port': 'port2', # This port is invalid
'output': out_queue,
'workflow': {
"tasks": [],
"flows": []
}
}
state_control.push_app_queue(app_id, json.dumps(data))
minion = SparkMinion(redis_conn=redis_conn,
workflow_id=workflow_id,
app_id=app_id,
config=config)
minion.get_or_create_spark_session = \
dummy_get_or_create_spark_session
minion._emit_event = dummy_emit_event
minion._state = {}
minion._process_message()
# Discard first status message
state_control.pop_app_output_queue(app_id, False)
msg = json.loads(
state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'WARNING', 'Invalid status'
assert msg['code'] == minion.MNN003[0], 'Invalid code'
msg = json.loads(
state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'ERROR', 'Invalid status'
assert msg.get('code') == minion.MNN004[0], 'Invalid code'
result = json.loads(state_control.pop_queue(out_queue, False))
assert not result['sample'], 'Wrong CSV generated'
def test_minion_perform_deliver_success():
workflow_id = '6666'
app_id = '1000'
job_id = '1'
out_queue = 'queue_2000'
sconf = SparkConf()
sc = SparkContext(master='', conf=sconf)
rdd = sc.parallelize(get_records())
df0 = DataFrame(rdd=rdd)
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
redis_conn = mocked_redis()
state_control = StateControlRedis(redis_conn)
data = {
'workflow_id': workflow_id,
'app_id': app_id,
'job_id': job_id,
'type': 'deliver',
'task_id': '033f-284ab-28987e',
'port': 'port0',
'output': out_queue,
'workflow': ''
}
state_control.push_app_queue(app_id, json.dumps(data))
minion = SparkMinion(redis_conn=redis_conn,
workflow_id=workflow_id,
app_id=app_id,
config=config)
minion._emit_event = dummy_emit_event
minion._state = {
data['task_id']: {
'port0': {
'output': df0,
'sample': []
},
'time': 35.92
}
}
minion._process_message()
# Discard first status message
state_control.pop_app_output_queue(app_id, False)
msg = json.loads(state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'SUCCESS', 'Invalid status'
assert msg['code'] == minion.MNN002[0], 'Invalid code'
# CSV data
csv_records = '\n'.join(
map(dataframe_util.convert_to_csv, get_records()))
result = json.loads(state_control.pop_queue(out_queue, False))
assert result['sample'] == csv_records, 'Wrong CSV generated'
def test_minion_generate_output_success():
workflow_id = 6666
app_id = 897987
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
redis_conn = mocked_redis()
minion = SparkMinion(redis_conn=redis_conn,
workflow_id=workflow_id,
app_id=app_id,
config=config)
minion._emit_event = dummy_emit_event
state_control = StateControlRedis(redis_conn)
msgs = ["Message being sent \n{}", {'msg': 'Dictionary being sent'}]
for msg in msgs:
minion._generate_output(msg)
result = json.loads(
state_control.pop_app_output_queue(app_id, False))
assert sorted(result.keys()) == sorted(
['date', 'status', 'message', 'workflow_id', 'app_id', 'code'])
assert result['app_id'] == app_id
assert result['message'] == msg
assert state_control.get_app_output_queue_size(app_id) == 0
def test_minion_generate_invalid_code_failure():
workflow_id = '6666'
app_id = '667788'
job_id = '1'
workflow = {
'workflow_id': workflow_id,
'app_id': app_id,
'job_id': job_id,
'type': 'execute',
'workflow': ''
}
function_name = 'juicer.spark.transpiler.SparkTranspiler.transpile'
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch(function_name) as mocked_transpile:
with mock.patch('juicer.workflow.workflow.Workflow'
'._build_initial_workflow_graph') as mocked_fn:
mocked_fn.side_effect = lambda: ""
# Setup for mocked_transpile
mocked_transpile.side_effect = get_side_effect(None, None, 4)
redis_conn = mocked_redis()
minion = SparkMinion(redis_conn=redis_conn,
workflow_id=workflow_id,
app_id=app_id,
config=config)
minion._emit_event = dummy_emit_event
# Configure mocked redis
with open(
os.path.join(os.path.dirname(__file__),
'fixtures/simple_workflow.json')) as f:
data = json.loads(f.read())
workflow['workflow'] = data
state_control = StateControlRedis(redis_conn)
state_control.push_app_queue(app_id, json.dumps(workflow))
minion._process_message()
assert state_control.get_app_output_queue_size(
app_id) == 2, 'Wrong number of output messages'
# Discards
state_control.pop_app_output_queue(app_id)
msg = json.loads(state_control.pop_app_output_queue(app_id))
assert msg['status'] == 'ERROR'
assert msg['message'][:19] == 'Invalid Python code'
def test_runner_read_start_queue_success():
config = {'juicer': {'servers': {'redis_url': "nonexisting.mock"}}}
app_id = '1'
workflow_id = '1000'
workflow = {
'workflow_id': workflow_id,
'app_id': app_id,
'type': 'execute',
'workflow': {}
}
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch('subprocess.Popen') as mocked_popen:
server = server = JuicerServer(config,
'faked_minions.py',
config_file_path='config.yaml')
mocked_redis_conn = mocked_redis()
state_control = StateControlRedis(mocked_redis_conn)
# Publishes a message to process data
state_control.push_start_queue(json.dumps(workflow))
# Start of testing
server.read_job_start_queue(mocked_redis_conn)
d1 = json.loads(state_control.get_minion_status(app_id))
d2 = {
"port": 36000,
"pid": 1,
}
assert d1 == d2
assert mocked_popen.call_args_list[0][0][0] == [
'nohup', sys.executable, 'faked_minions.py', '-w', workflow_id,
'-a', app_id, '-t', 'spark', '-c', 'config.yaml'
]
assert mocked_popen.called
# Was command removed from the queue?
assert state_control.pop_job_start_queue(False) is None
assert json.loads(state_control.pop_app_queue(app_id)) == workflow
assert state_control.get_workflow_status(
workflow_id) == JuicerServer.STARTED
assert json.loads(state_control.pop_app_output_queue(app_id)) == {
'code': 0,
'message': 'Minion is processing message execute'
}
def test_runner_read_start_queue_minion_already_running_success():
config = {'juicer': {'servers': {'redis_url': "nonexisting.mock"}}}
app_id = 1
workflow_id = 1000
workflow = {
'workflow_id': workflow_id,
'app_id': app_id,
'type': 'execute',
'workflow': {}
}
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch('subprocess.Popen') as mocked_popen:
server = JuicerServer(config, 'faked_minions.py')
mocked_redis_conn = mocked_redis()
state_control = StateControlRedis(mocked_redis_conn)
# Publishes a message to process data
state_control.push_start_queue(json.dumps(workflow))
state_control.set_minion_status(app_id, JuicerServer.STARTED)
# Start of testing
server.read_job_start_queue(mocked_redis_conn)
assert state_control.get_minion_status(
app_id) == JuicerServer.STARTED
assert not mocked_popen.called
# Was command removed from the queue?
assert mocked_redis_conn.lpop('start') is None
assert json.loads(state_control.pop_app_queue(app_id)) == workflow
assert state_control.get_workflow_status(
workflow_id) == JuicerServer.STARTED
assert json.loads(state_control.pop_app_output_queue(app_id)) == {
'code': 0,
'message': 'Minion is processing message execute'
}
def test_runner_read_start_queue_missing_details_failure():
config = {'juicer': {'servers': {'redis_url': "nonexisting.mock"}}}
app_id = 1
workflow_id = 1000
# incorrect key, should raise exception
workflow = {
'workflow_id': workflow_id,
'xapp_id': app_id,
'type': 'execute',
'workflow': {}
}
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch('subprocess.Popen') as mocked_popen:
server = JuicerServer(config, 'faked_minions.py')
mocked_redis_conn = mocked_redis()
# Publishes a message to process data
state_control = StateControlRedis(mocked_redis_conn)
# Publishes a message to process data
state_control.push_start_queue(json.dumps(workflow))
state_control.set_minion_status(app_id, JuicerServer.STARTED)
# Start of testing
server.read_job_start_queue(mocked_redis_conn)
assert state_control.get_minion_status(
app_id) == JuicerServer.STARTED
assert not mocked_popen.called
# Was command removed from the queue?
assert state_control.pop_job_start_queue(block=False) is None
assert state_control.pop_app_queue(app_id, block=False) is None
assert state_control.pop_app_output_queue(app_id,
block=False) is None
assert mocked_redis_conn.hget(workflow_id, 'status') is None
def test_minion_terminate():
try:
from pyspark.sql import SparkSession
except ImportError:
# we will skip this test because pyspark is not installed
return
workflow_id = '6666'
app_id = '897447'
job_id = '1'
function_name = 'juicer.spark.transpiler.SparkTranspiler.transpile'
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch(function_name) as mocked_transpile:
# Setup for mocked_transpile
mocked_transpile.side_effect = get_side_effect(None, 0, 5)
redis_conn = mocked_redis()
minion = SparkMinion(redis_conn=redis_conn,
workflow_id=workflow_id,
app_id=app_id,
config=config)
minion._emit_event = dummy_emit_event
# Configure mocked redis
state_control = StateControlRedis(redis_conn)
with open(
os.path.join(os.path.dirname(__file__),
'fixtures/simple_workflow.json')) as f:
data = json.loads(f.read())
# execute message
state_control.push_app_queue(
app_id,
json.dumps({
'workflow_id': workflow_id,
'app_id': app_id,
'job_id': job_id,
'type': 'execute',
'workflow': data
}))
minion._process_message_nb()
# discard extra message
state_control.pop_app_output_queue(app_id, False)
# job termination
state_control.push_app_queue(
app_id,
json.dumps({
'workflow_id': workflow_id,
'app_id': app_id,
'job_id': job_id,
'type': 'terminate'
}))
minion._process_message()
state_control.pop_app_output_queue(app_id, False)
# first the spark app will throw an exception regarding the job
# canceling
msg = json.loads(state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'ERROR', 'Invalid status'
assert msg['code'] == 1000, 'Invalid code'
# second the minion will report success for the job canceling
# operation
msg = json.loads(state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'SUCCESS', 'Invalid status'
assert msg['code'] == SparkMinion.MNN007[0], 'Invalid code'
# assert app still alive
assert minion.spark_session is not None
assert minion.is_spark_session_available()
# app termination
state_control.push_app_queue(
app_id,
json.dumps({
'workflow_id': workflow_id,
'app_id': app_id,
'type': 'terminate'
}))
minion._process_message()
# discard extra message
state_control.pop_app_output_queue(app_id, False)
msg = json.loads(state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'SUCCESS', 'Invalid status'
assert msg['code'] == SparkMinion.MNN008[0], 'Invalid code'
# assert app still alive
assert minion.spark_session is None
assert not minion.is_spark_session_available()
minion.terminate()
assert not minion.is_spark_session_available()
def test_minion_spark_configuration():
"""
- Start a juicer server
- Instanciate one minion passing specific configurations w.r.t. runtime
environments (app_configs)
- Assert that the spark context within the minion inherited the same configs
that it was supposed to inherit
"""
try:
# noinspection PyUnresolvedReferences
from pyspark.sql import SparkSession
except ImportError:
# we will skip this test because pyspark is not installed
return
workflow_id = '6666'
app_id = '667788'
job_id = '1'
app_configs = {'spark.master': 'local[3]', 'config1': '1', 'config2': '2'}
function_name = 'juicer.spark.transpiler.SparkTranspiler.transpile'
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch(function_name) as mocked_transpile:
with mock.patch('juicer.workflow.workflow.Workflow'
'._build_initial_workflow_graph') as mocked_fn:
mocked_fn.side_effect = lambda: ""
# Setup for mocked_transpile
mocked_transpile.side_effect = get_side_effect(None, 0, 1)
redis_conn = mocked_redis()
minion = SparkMinion(redis_conn=redis_conn,
workflow_id=workflow_id,
app_id=app_id,
config=config)
minion._emit_event = dummy_emit_event
# Configure mocked redis
state_control = StateControlRedis(redis_conn)
with open(
os.path.join(os.path.dirname(__file__),
'fixtures/simple_workflow.json')) as f:
data = json.loads(f.read())
state_control.push_app_queue(
app_id,
json.dumps({
'workflow_id': workflow_id,
'app_id': app_id,
'job_id': job_id,
'type': 'execute',
'app_configs': app_configs,
'workflow': data
}))
minion._process_message()
# check spark session health
assert minion.spark_session is not None
assert minion.is_spark_session_available()
# check configs
ctx_configs = \
minion.spark_session.sparkContext.getConf().getAll()
ctx_configs = {k: v for k, v in ctx_configs}
for k, v in app_configs.items():
assert ctx_configs[k] == v
# check app name
name = minion.spark_session.sparkContext.appName
assert name == u'{}(workflow_id={},app_id={})'.format(
data['name'], workflow_id, app_id)
# check proper termination
minion.terminate()
assert not minion.is_spark_session_available()
state_control.pop_app_output_queue(app_id, False)
msg = json.loads(
state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'SUCCESS', 'Invalid status'
assert msg['code'] == minion.MNN008[0], 'Invalid code'
def test_minion_perform_deliver_missing_state_process_app_with_failure():
workflow_id = '6666'
app_id = '6000'
job_id = '1'
out_queue = 'queue_2000'
task_id = '033f-284ab-28987e'
function_name = 'juicer.spark.transpiler.SparkTranspiler.transpile'
with mock.patch('redis.StrictRedis',
mock_strict_redis_client) as mocked_redis:
with mock.patch(function_name) as mocked_transpile:
with mock.patch('juicer.workflow.workflow.Workflow'
'._build_initial_workflow_graph') as mocked_fn:
mocked_fn.side_effect = lambda: ""
# Setup for mocked_transpile
# Invalid code
mocked_transpile.side_effect = get_side_effect(
get_records(), task_id, 4)
redis_conn = mocked_redis()
state_control = StateControlRedis(redis_conn)
data = {
'workflow_id': workflow_id,
'app_id': app_id,
'job_id': job_id,
'type': 'deliver',
'task_id': task_id,
'port': 'port0',
'output': out_queue,
'workflow': {
"tasks": [],
"flows": []
}
}
state_control.push_app_queue(app_id, json.dumps(data))
minion = SparkMinion(redis_conn=redis_conn,
workflow_id=workflow_id,
app_id=app_id,
config=config)
minion._emit_event = dummy_emit_event
minion._state = {}
minion._process_message()
# Discard first status message
state_control.pop_app_output_queue(app_id, False)
# First message is about missing state
msg = json.loads(
state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'WARNING', 'Invalid status'
assert msg['code'] == minion.MNN003[0], 'Invalid code'
# Second message is about invalid Python code
msg = json.loads(
state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'ERROR', 'Invalid status'
assert msg.get('code') == minion.MNN006[0], 'Invalid code'
# Third message is about unable to read data
msg = json.loads(
state_control.pop_app_output_queue(app_id, False))
assert msg['status'] == 'ERROR', 'Invalid status'
assert msg.get('code') == minion.MNN005[0], 'Invalid code'
assert state_control.get_app_output_queue_size(
app_id) == 0, 'There are messages in app output queue!'
result = json.loads(state_control.pop_queue(out_queue, False))
assert not result['sample'], 'Wrong CSV generated'
