def __init__(self, eggroll_session):
if _EggRoll.instance is not None:
raise EnvironmentError("eggroll should be initialized only once")
host = eggroll_session.get_conf(EGGROLL_ROLL_HOST)
port = eggroll_session.get_conf(EGGROLL_ROLL_PORT)
self.chunk_size = eggroll_session.get_chunk_size()
self.host = host
self.port = port
self.channel = grpc.insecure_channel(
target="{}:{}".format(host, port),
options=[('grpc.max_send_message_length', -1),
('grpc.max_receive_message_length', -1)])
self.session_id = eggroll_session.get_session_id()
self.kv_stub = kv_pb2_grpc.KVServiceStub(self.channel)
self.proc_stub = processor_pb2_grpc.ProcessServiceStub(self.channel)
self.session_stub = node_manager_pb2_grpc.SessionServiceStub(
self.channel)
self.eggroll_session = eggroll_session
_EggRoll.instance = self
self.session_stub.getOrCreateSession(
self.eggroll_session.to_protobuf())
# todo: move to eggrollSession
try:
self.host_name = socket.gethostname()
self.host_ip = socket.gethostbyname(self.host_name)
except socket.gaierror as e:
self.host_name = 'unknown'
self.host_ip = 'unknown'
def init():
if EggRoll.init_flag:
return
config = file_utils.load_json_conf('eggroll/conf/mock_roll.json')
egg_ids = config.get('eggs')
for egg_id in egg_ids:
target = config.get('storage').get(egg_id)
channel = grpc.insecure_channel(
target,
options=[('grpc.max_send_message_length', -1),
('grpc.max_receive_message_length', -1)])
EggRoll.egg_list.append(kv_pb2_grpc.KVServiceStub(channel))
procs = config.get('procs').get(egg_id)
for proc in procs:
_channel = grpc.insecure_channel(
proc,
options=[('grpc.max_send_message_length', -1),
('grpc.max_receive_message_length', -1)])
_stub = processor_pb2_grpc.ProcessServiceStub(_channel)
proc_info = (_channel, _stub)
i = len(EggRoll.proc_list)
EggRoll.proc_egg_map[i] = int(egg_id) - 1
EggRoll.proc_list.append(proc_info)
EggRoll.init_flag = True
def action(_table, host, port, chunked_iter, use_serialize):
_table.set_gc_disable()
_EggRoll.get_instance().channel = grpc.insecure_channel(
target="{}:{}".format(host, port),
options=[('grpc.max_send_message_length', -1),
('grpc.max_receive_message_length', -1)])
with _EggRoll.get_instance().get_channel() as child_channle:
_EggRoll.get_instance().kv_stub = kv_pb2_grpc.KVServiceStub(
child_channle)
_EggRoll.get_instance(
).proc_stub = processor_pb2_grpc.ProcessServiceStub(child_channle)
operand = _EggRoll.get_instance().__generate_operand(
chunked_iter, use_serialize)
_EggRoll.get_instance().kv_stub.putAll(operand,
metadata=_get_meta(_table))
def __init__(self, job_id, host, port, eggroll_context):
if _EggRoll.instance is not None:
raise EnvironmentError("eggroll should be initialized only once")
self.channel = grpc.insecure_channel(target="{}:{}".format(host, port),
options=[('grpc.max_send_message_length', -1),
('grpc.max_receive_message_length', -1)])
self.job_id = job_id
self.kv_stub = kv_pb2_grpc.KVServiceStub(self.channel)
self.proc_stub = processor_pb2_grpc.ProcessServiceStub(self.channel)
self.eggroll_context = eggroll_context
_EggRoll.instance = self
# todo: move to eggrollContext
try:
self.host_name = socket.gethostname()
self.host_ip = socket.gethostbyname(self.host_name)
except socket.gaierror as e:
self.host_name = 'unknown'
self.host_ip = 'unknown'
def put_all(self,
_table,
data: Iterable,
use_serialize=True,
chunk_size=100000,
skip_chunk=0,
include_key=True,
single_process=False):
global gc_tag
gc_tag = False
skipped_chunk = 0
if include_key == True:
kvs = data
else:
kvs = enumerate(data)
chunk_size = self.chunk_size
if chunk_size < CHUNK_SIZE_MIN:
chunk_size = CHUNK_SIZE_DEFAULT
host = self.host
port = self.port
process_pool_size = cpu_count()
if single_process is True:
for chunked_iter in split_every_yield(kvs, chunk_size=chunk_size):
if skipped_chunk < skip_chunk:
skipped_chunk += 1
else:
self.kv_stub.putAll(self.__generate_operand(
chunked_iter, use_serialize=use_serialize),
metadata=_get_meta(_table))
else:
_EggRoll.get_instance().get_channel().close()
with ProcessPoolExecutor(process_pool_size) as executor:
if isinstance(kvs, Sequence): # Sequence
for chunked_iter in split_every_yield(kvs, chunk_size):
if skipped_chunk < skip_chunk:
skipped_chunk += 1
else:
future = executor.submit(_EggRoll.action, _table,
host, port, chunked_iter,
use_serialize)
elif isinstance(kvs, Generator) or isinstance(data, Generator):
index = 0
while True:
chunked_iter = split_every_generator(
kvs, chunk_size, skip_chunk)
if chunked_iter == []:
break
future = executor.submit(_EggRoll.action, _table, host,
port, chunked_iter,
use_serialize)
index += 1
else: # other Iterable types
try:
index = 0
while True:
chunked_iter = split_every(kvs, index, chunk_size,
skip_chunk)
chunked_iter_ = copy.deepcopy(chunked_iter)
next(chunked_iter_)
future = executor.submit(_EggRoll.action, _table,
host, port, chunked_iter,
use_serialize)
index += 1
except StopIteration as e:
LOGGER.debug("StopIteration")
executor.shutdown(wait=True)
_EggRoll.get_instance().channel = grpc.insecure_channel(
target="{}:{}".format(host, port),
options=[('grpc.max_send_message_length', -1),
('grpc.max_receive_message_length', -1)])
_EggRoll.get_instance().kv_stub = kv_pb2_grpc.KVServiceStub(
_EggRoll.get_instance().get_channel())
_EggRoll.get_instance(
).proc_stub = processor_pb2_grpc.ProcessServiceStub(
_EggRoll.get_instance().get_channel())
_EggRoll.get_instance(
).session_stub = node_manager_pb2_grpc.SessionServiceStub(
_EggRoll.get_instance().get_channel())
gc_tag = True
