Develop Reference

Multiprocessing BrokenPipeError

I have a following problem. I am running a parallel task. I am getting this error:
Traceback (most recent call last):
File "/usr/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap
self.run()
File "/usr/lib/python3.8/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "eclat_model.py", line 127, in do_work
function(*args, work_queue, valid_list)
File "eclat_model.py", line 115, in eclat_parallel_helper
valid_list.extend(next_vectors)
File "<string>", line 2, in extend
File "/usr/lib/python3.8/multiprocessing/managers.py", line 834, in _callmethod
conn.send((self._id, methodname, args, kwds))
File "/usr/lib/python3.8/multiprocessing/connection.py", line 206, in send
self._send_bytes(_ForkingPickler.dumps(obj))
File "/usr/lib/python3.8/multiprocessing/connection.py", line 404, in _send_bytes
self._send(header)
File "/usr/lib/python3.8/multiprocessing/connection.py", line 368, in _send
n = write(self._handle, buf)
BrokenPipeError: [Errno 32] Broken pipe
Relevant functions in eclat_model.py look like this:
def eclat_parallel_helper(index, bit_vectors, min_support, work_queue, valid_list):
next_vectors = []
for j in range(index + 1, len(bit_vectors)):
item_vector = bit_vectors[index][0] | bit_vectors[j][0]
transaction_vector = bit_vectors[index][1] & bit_vectors[j][1]
support = get_vector_support(transaction_vector)
if support >= min_support:
next_vectors.append((item_vector, transaction_vector, support))
if len(next_vectors) > 0:
valid_list.extend(next_vectors)
for i in range(len(next_vectors)):
work_queue.put((eclat_parallel_helper, (i, next_vectors, min_support)))
def do_work(work_queue, valid_list, not_done):
# work queue entries have the form (function, args)
while not_done.value:
try:
function, args = work_queue.get_nowait()
except QueueEmptyError:
continue
function(*args, work_queue, valid_list)
work_queue.task_done()
work_queue.close()
EDIT:
Multiprocessing part of the code is as follows: bit_vectors is a list of lists, where each entry is of the form
[items, transactions, support], where items is a bit vector encoding which items appear in the itemset, vector is a bit vector encoding which transactions the itemset appears in, and support is the number of transactions in which the itemset occurs.
from multiprocessing import Process, JoinableQueue, Manager, Value, cpu_count
def eclat_parallel(bit_vectors, min_support):
not_done = Value('i', 1)
manager = Manager()
valid_list = manager.list()
work_queue = JoinableQueue()
for i in range(len(bit_vectors)):
work_queue.put((eclat_parallel_helper, (i, bit_vectors, min_support)))
processes = []
for i in range(cpu_count()):
p = Process(target=do_work, args=(work_queue, valid_list, not_done), daemon=True)
p.start()
processes.append(p)
work_queue.join()
not_done.value = 0
work_queue.close()
valid_itemset_vectors = bit_vectors
for element in valid_list:
valid_itemset_vectors.append(element)
for p in processes:
p.join()
return valid_itemset_vectors
What does this error mean, please? Am I appending too many elements into next_vectors list?

I had the same issue, in my case just added a delay (time.sleep(0.01)) to solve it.
The problem is that the individual processes are too fast on queue that causes the error.

sharing a PyTable across multiprocesses

I create a PyTable object W_hat where processes should share and save the results their instead of returning them.
from multiprocessing import Lock
from multiprocessing import Pool
import tables as tb
def parallel_l21(labels, X, lam, g, W_hat):
g_indxs = np.where(labels == g)[0]
tmp = rfs(X[g_indxs, 1:].T, X[:, :-1].T, gamma=lam, verbose=False).T
tmp[abs(tmp) <= 1e-6] = 0
with lock:
W_hat[:, g_indxs] = np.array(tmp)
def init_child(lock_):
global lock
lock = lock_
#Previous code is omitted.
n_ = X_test.shape[0]
tb.file._open_files.close_all()
f = tb.open_file(path_name + 'dot' + sub_num + str(lam) + '.h5', 'w')
filters = tb.Filters(complevel=5, complib='blosc')
W_hat = f.create_carray(f.root, 'data', tb.Float32Atom(), shape=(n_, n_), filters=filters)
W_hats = []
for i in np.unique(labels):
W_hats.append(W_hat)
lock = Lock()
with Pool(processes=cpu_count, initializer=init_child, initargs=(lock,)) as pool:
print(pool)
pool.starmap(parallel_l21, zip(repeat(labels), repeat(X), repeat(lam), np.unique(labels), W_hats))
Now, when running into starmap, this error shows up:
Traceback (most recent call last):
File "/Applications/PyCharm CE 2.app/Contents/plugins/python-ce/helpers/pydev/_pydevd_bundle/pydevd_exec2.py", line 3, in Exec
exec(exp, global_vars, local_vars)
File "<input>", line 1, in <module>
File "/usr/local/Cellar/python#3.8/3.8.6_1/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/pool.py", line 372, in starmap
return self._map_async(func, iterable, starmapstar, chunksize).get()
File "/usr/local/Cellar/python#3.8/3.8.6_1/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/pool.py", line 771, in get
raise self._value
File "/usr/local/Cellar/python#3.8/3.8.6_1/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/pool.py", line 537, in _handle_tasks
put(task)
File "/usr/local/Cellar/python#3.8/3.8.6_1/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/connection.py", line 206, in send
self._send_bytes(_ForkingPickler.dumps(obj))
File "/usr/local/Cellar/python#3.8/3.8.6_1/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/reduction.py", line 51, in dumps
cls(buf, protocol).dump(obj)
File "stringsource", line 2, in tables.hdf5extension.Array.__reduce_cython__
TypeError: self.dims,self.dims_chunk,self.maxdims cannot be converted to a Python object for pickling
Note: I thought that the code works fine on Python 3.6.8 but it turns out that it is not the case.

Multiprocessing Gremlin "OSError: [Errno 9] Bad file descriptor"

I'm trying to compute a feature for every vertex in my graph using gremlinpython. It's too slow to sequentially iterate over every single vertex. While batching could help to provide a speedup, I thought first I'd try parallizing the query.
Broadly, 1. get the full set of vertices, 2. split them over num_cores=x, 3. iterate over each sub-vertex set in parallel.
But I'm getting the error "OSError: [Errno 9] Bad file descriptor". The below code is my latest attempt at solving this.
import multiprocessing
from gremlin_python.structure.graph import Graph
from gremlin_python.driver.driver_remote_connection import DriverRemoteConnection
from gremlin_python.process.traversal import lt
def create_traversal_object():
graph = Graph()
g = graph.traversal().withRemote(DriverRemoteConnection('ws://localhost:8182/gremlin', 'g'))
return g
g = create_traversal_object()
num_cores = 1
vertex_lsts = np.array_split(g.V().limit(30).id().toList(), num_cores)
class FeatureClass():
def __init__(self, g, vertex_list):
self.g = g
self.vertex_list = vertex_list
def orchestrator(self):
for vertex_id in self.vertex_list:
self.compute_number_of_names(float(vertex_id))
def get_names(self, vertex_id):
return self.g.V(vertex_id).inE().values('benef_nm').dedup().toList()
class Simulation(multiprocessing.Process):
def __init__(self, id, worker, *args, **kwargs):
# must call this before anything else
multiprocessing.Process.__init__(self)
self.id = id
self.worker = worker
self.args = args
self.kwargs = kwargs
sys.stdout.write('[%d] created\n' % (self.id))
def run(self):
sys.stdout.write('[%d] running ... process id: %s\n' % (self.id, os.getpid()))
self.worker.orchestrator()
sys.stdout.write('[%d] completed\n' % (self.id))
list_of_objects = [FeatureClass(create_traversal_object(), vertex_lst) for vertex_lst in vertex_lsts]
list_of_sim = [Simulation(id=k, worker=obj) for k, obj in enumerate(list_of_objects)]
for sim in list_of_sim:
sim.start()
Here's the full stack-trace, looks like it's an issue with tornado, which gremlinpython uses.
Process Simulation-1:
Traceback (most recent call last):
File "/Users/greatora/anaconda3/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "<ipython-input-4-b3177477fabe>", line 42, in run
self.worker.orchestrator()
File "<ipython-input-4-b3177477fabe>", line 23, in orchestrator
self.compute_number_of_names(float(vertex_id))
File "<ipython-input-4-b3177477fabe>", line 26, in compute_number_of_names
print(self.g.V(vertex_id).inE().values('benef_nm').dedup().count().next())
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/process/traversal.py", line 88, in next
return self.__next__()
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/process/traversal.py", line 47, in __next__
self.traversal_strategies.apply_strategies(self)
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/process/traversal.py", line 512, in apply_strategies
traversal_strategy.apply(traversal)
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/driver/remote_connection.py", line 148, in apply
remote_traversal = self.remote_connection.submit(traversal.bytecode)
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/driver/driver_remote_connection.py", line 53, in submit
result_set = self._client.submit(bytecode)
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/driver/client.py", line 108, in submit
return self.submitAsync(message, bindings=bindings).result()
File "/Users/greatora/anaconda3/lib/python3.6/concurrent/futures/_base.py", line 432, in result
return self.__get_result()
File "/Users/greatora/anaconda3/lib/python3.6/concurrent/futures/_base.py", line 384, in __get_result
raise self._exception
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/driver/connection.py", line 63, in cb
f.result()
File "/Users/greatora/anaconda3/lib/python3.6/concurrent/futures/_base.py", line 425, in result
return self.__get_result()
File "/Users/greatora/anaconda3/lib/python3.6/concurrent/futures/_base.py", line 384, in __get_result
raise self._exception
File "/Users/greatora/anaconda3/lib/python3.6/concurrent/futures/thread.py", line 56, in run
result = self.fn(*self.args, **self.kwargs)
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/driver/protocol.py", line 74, in write
self._transport.write(message)
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/gremlin_python/driver/tornado/transport.py", line 37, in write
lambda: self._ws.write_message(message, binary=True))
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/tornado/ioloop.py", line 453, in run_sync
self.start()
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/tornado/ioloop.py", line 863, in start
event_pairs = self._impl.poll(poll_timeout)
File "/Users/greatora/anaconda3/lib/python3.6/site-packages/tornado/platform/kqueue.py", line 66, in poll
kevents = self._kqueue.control(None, 1000, timeout)
OSError: [Errno 9] Bad file descriptor
I'm using Pythton3.7, gremlinpython==3.4.6, MacOS.

I'm still not entirely sure what the issue was, but this works.
import multiprocessing
from multiprocessing import Pool
import itertools
def graph_function(vertex_id_list):
graph = Graph()
g = graph.traversal().withRemote(DriverRemoteConnection('ws://localhost:8182/gremlin', 'g'))
res = []
for vertex_id in vertex_id_list:
res.append(g.V(str(vertex_id)).inE().values('benef_nm').dedup().toList())
return res
num_cores = 4
vertex_lst = g.V().limit(30).id().toList()
vertex_lsts = np.array_split(vertex_lst, num_cores)
with Pool(processes=num_cores) as pool:
results = pool.map(graph_function, vertex_lsts)
results = [*itertools.chain.from_iterable(results)]

How to return a counter dictionary from a function passed to multiprocessing?

I have a list of CSV files. I want to do a set of operations on each of them and then produce a counter dict and i want to cerate a master list containing individual counter dict from all CSV files. I want to parallelize processing each of the csv file and then return the counter dict from each file. I found a similar solution here : How can I recover the return value of a function passed to multiprocessing.Process?
I used the solution suggested by David Cullen. This solution works perfectly for strings, but when I tried to return a counter dict or a normal dict. All the CSV files are processed until the send_end.send(result) and it hangs on there forever when executed and then throws a memory error. I am running this in a Linux server with more than sufficient memory for creating the list of counter dicts.
I used the following code:
import multiprocessing
#get current working directory
cwd = os.getcwd()
#take a list of all files in cwd
files = os.listdir(cwd)
#defining the function that needs to be done on all csv files
def worker(f,send_end):
infile= open(f)
#read liens in csv file
lines = infile.readlines()
#split the lines by "," and store it in a list of lists
master_lst = [line.strip().split(“,”) for line in lines]
#extract the second field in each sublist
counter_lst = [ element[1] for element in master_lst]
print “Total elements in the list” + str(len(counter_lst))
#create a dictionary of count elements
a = Counter(counter_lst)
# return the counter dict
send_end.send(a)
def main():
jobs = []
pipe_list = []
for f in files:
if f.endswith('.csv'):
recv_end, send_end = multiprocessing.Pipe(duplex=False)
p = multiprocessing.Process(target=worker, args=(f, send_end))
jobs.append(p)
pipe_list.append(recv_end)
p.start()
for proc in jobs:
proc.join()
result_list = [x.recv() for x in pipe_list]
print len(result_list)
if __name__ == '__main__':
main()
The error that i get is the following:
Process Process-42:
Traceback (most recent call last):
File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in
_bootstrap
self.run()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 114, in run
self._target(*self._args, **self._kwargs)
File "/home/amm/python/collapse_multiprocessing_return.py", line 32, in
worker
a = Counter(counter_lst)
File "/usr/lib64/python2.7/collections.py", line 444, in __init__
self.update(iterable, **kwds)
File "/usr/lib64/python2.7/collections.py", line 526, in update
self[elem] = self_get(elem, 0) + 1
MemoryError
Process Process-17:
Traceback (most recent call last):
Process Process-6:
Traceback (most recent call last):
File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in
_bootstrap
File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in
_bootstrap
Process Process-8:
Traceback (most recent call last):
File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in
_bootstrap
self.run()
self.run()
self.run()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 114, in run
File "/usr/lib64/python2.7/multiprocessing/process.py", line 114, in run
self._target(*self._args, **self._kwargs)
File "/usr/lib64/python2.7/multiprocessing/process.py", line 114, in run
File "/home/amm/python/collapse_multiprocessing_return.py", line 32, in
worker
self._target(*self._args, **self._kwargs)
self._target(*self._args, **self._kwargs)
File "/home/amm/python/collapse_multiprocessing_return.py", line 32, in
worker
File "/home/amm/python/collapse_multiprocessing_return.py", line 32, in
worker
a = Counter(counter_lst_lst)
a = Counter(counter_lst_lst)
a = Counter(counter_lst_lst)
File "/usr/lib64/python2.7/collections.py", line 444, in __init__
File "/usr/lib64/python2.7/collections.py", line 444, in __init__
File "/usr/lib64/python2.7/collections.py", line 444, in __init__
self.update(iterable, **kwds)
File "/usr/lib64/python2.7/collections.py", line 526, in update
self[elem] = self_get(elem, 0) + 1
MemoryError
self.update(iterable, **kwds)
self.update(iterable, **kwds)
File "/usr/lib64/python2.7/collections.py", line 526, in update
File "/usr/lib64/python2.7/collections.py", line 526, in update
self[elem] = self_get(elem, 0) + 1
self[elem] = self_get(elem, 0) + 1
MemoryError
MemoryError
Process Process-10:
Traceback (most recent call last):
File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in
_bootstrap
self.run()
File "/usr/lib64/python2.7/multiprocessing/process.py", line 114, in run
self._target(*self._args, **self._kwargs)
File "/home/amm/python/collapse_multiprocessing_return.py", line 32, in
worker
a = Counter(counter_lst)
File "/usr/lib64/python2.7/collections.py", line 444, in __init__
self.update(iterable, **kwds)
File "/usr/lib64/python2.7/collections.py", line 526, in update
self[elem] = self_get(elem, 0) + 1
MemoryError
^Z
[18]+ Stopped collapse_multiprocessing_return.py
Now instead of "a" in send_end.send(a) if i replace f, the filename. It prints the number of csv files in the directory (which is what len(result_list) does in this case). But when the counter dict "a" is returned it gets stuck forever, throwing the above error.
I would like to have the code pass the counter dict to receive end without any error/problems. Is there a work around? Could someone please suggest a possible solution?
p.s: I am new to multiprocessing module, sorry if this question sounds naive. Also, i tried the multiprocessing.Manager(), but got a similar error

Your traceback mentions Process Process-42:, so there are at least 42 processes being created. You're creating a process for every CSV file, which is not useful and is probably causing the memory error.
Your problem can be solved much more simply using multiprocessing.Pool.map. The worker function can also be shortened greatly:
def worker(f):
with open(f) as infile:
return Counter(line.strip().split(",")[1]
for line in infile)
def main():
pool = multiprocessing.Pool()
result_list = pool.map(worker, [f for f in files if f.endswith('.csv')])
Passing no arguments to the pool means it'll create as many processes as you have CPU cores. Using more may or may not increase performance.

need help with this Kombu error

Pardon the copy and paste from the python interpreter but I'm trying to play with Kombu but I can't seem to create a consumer. Please help, I'm utterly in the dark here.
>>> from kombu.messaging import Consumer, Producer
>>> from kombu.entity import Exchange, Queue
>>> x = Exchange("stmt",type="topic")
>>> helloQ = Queue("hello", exchange=x, routing_key="stmt.hello")
>>>
>>> from kombu.connection import BrokerConnection
>>> conn = BrokerConnection("scheduledb.lab.compete.com", "clippy", "clippy", "clippy")
>>> channel = conn.channel()
>>> c = Consumer(channel, helloQ)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/messaging.py", line 231, in __init__
self.declare()
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/messaging.py", line 241, in declare
queue.declare()
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/entity.py", line 362, in declare
self.name and self.queue_declare(nowait, passive=False),
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/entity.py", line 380, in queue_declare
nowait=nowait)
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/syn.py", line 14, in blocking
return __sync_current(fun, *args, **kwargs)
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/syn.py", line 30, in __blocking__
return fun(*args, **kwargs)
File "build/bdist.cygwin-1.7.8-i686/egg/amqplib/client_0_8/channel.py", line 1294, in queue_declare
File "build/bdist.cygwin-1.7.8-i686/egg/amqplib/client_0_8/abstract_channel.py", line 89, in wait
File "build/bdist.cygwin-1.7.8-i686/egg/amqplib/client_0_8/connection.py", line 218, in _wait_method
File "build/bdist.cygwin-1.7.8-i686/egg/amqplib/client_0_8/abstract_channel.py", line 105, in wait
File "build/bdist.cygwin-1.7.8-i686/egg/amqplib/client_0_8/connection.py", line 367, in _close
amqplib.client_0_8.exceptions.AMQPConnectionException: (530, u"NOT_ALLOWED - parameters for queue 'hello' in vhost 'clippy' not equivalent", (50, 10), 'Channel.queue_declare')
>>> boundX = x(helloQ)
>>> c = Consumer(channel, helloQ)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/messaging.py", line 231, in __init__
self.declare()
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/messaging.py", line 241, in declare
queue.declare()
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/entity.py", line 361, in declare
return (self.name and self.exchange.declare(nowait),
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/entity.py", line 151, in declare
nowait=nowait)
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/syn.py", line 14, in blocking
return __sync_current(fun, *args, **kwargs)
File "/usr/lib/python2.6/site-packages/kombu-1.0.6-py2.6.egg/kombu/syn.py", line 30, in __blocking__
return fun(*args, **kwargs)
File "build/bdist.cygwin-1.7.8-i686/egg/amqplib/client_0_8/channel.py", line 839, in exchange_declare
File "build/bdist.cygwin-1.7.8-i686/egg/amqplib/client_0_8/abstract_channel.py", line 69, in _send_method
AttributeError: 'NoneType' object has no attribute 'method_writer'

Adding to #asksol's answer. If you are using rabbitmq you can use the rabbitmqctl command to list the queue details and compare those settings with the settings in your own code. Hopefully that gives you enough information to detect the conflict.

Look at the error
amqplib.client_0_8.exceptions.AMQPConnectionException: (530, u"NOT_ALLOWED - parameters for queue 'hello' in vhost 'clippy' not equivalent", (50, 10), 'Channel.queue_declare')
This means the queue has already been declared, but with other parameters than what you
are declaring it with now.