[ https://issues.apache.org/jira/browse/ARROW-10676?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17616526#comment-17616526 ]
Joris Van den Bossche commented on ARROW-10676: ----------------------------------------------- Not an answer to you actual issue, but note that Plasma is deprecated (see ARROW-17860 and the email thread that is linked there) > [Python] pickle error occurs when using pyarrow._plasma.PlasmaClient with > multiprocess on mac (python3.8.5) > ----------------------------------------------------------------------------------------------------------- > > Key: ARROW-10676 > URL: https://issues.apache.org/jira/browse/ARROW-10676 > Project: Apache Arrow > Issue Type: Bug > Components: C++ - Plasma, Python > Affects Versions: 0.17.1, 2.0.0 > Environment: OS: mac os catalina 10.15.6; > Python version: python3.8.5; > Reporter: BinbinLiang > Priority: Major > > * *The environment is:* > ** OS: mac os catalina 10.15.6; > ** {color:#ff0000}Python version: python3.8.5;{color} > * {color:#ff0000}*It's ok, when use python3.7(3.7.2 or 3.7.3) or > python3.6(3.6.8).{color}* > The error occurs only in python3.8(3.8.5). > * *When I use 'pyarrow._plasma.PlasmaClient' in the > 'multiprocessing.context.Process', the error occurs.* > * > ** First, I have a subclass of 'multiprocessing.context.Process', named > Executor; > The defined code like this: 'class Executor(Process):' . > ** And then, I create a 'PlasmaClient' object in the '__init__' function of > 'class Executor' . > ** Finally, I new a 'Executor' object, and call the 'start()' function. > * *The code is:* > {code:java} > // code placeholder > from multiprocessing.context import Process > from time import sleep > from pyarrow import plasma > class Executor(Process): > def __init__(self): > super().__init__() > self._plasma_client = plasma.connect("/tmp/plasma") > if __name__ == "__main__": > executor = Executor() > executor.start() > sleep(10) > {code} > * *The detail informations of traceback are as follows:* > {code:java} > // code placeholder > Traceback (most recent call last): > File > "/Users/liangbinbin/PycharmProjects/waimai_data_cubeinsight/python/taf/tests/plasma_client_with_multiprocessing_test.py", > line 15, in <module> > executor.start() > File > "/Users/liangbinbin/Applications/anaconda3/envs/python3_8/lib/python3.8/multiprocessing/process.py", > line 121, in start > self._popen = self._Popen(self) > File > "/Users/liangbinbin/Applications/anaconda3/envs/python3_8/lib/python3.8/multiprocessing/context.py", > line 224, in _Popen > return _default_context.get_context().Process._Popen(process_obj) > File > "/Users/liangbinbin/Applications/anaconda3/envs/python3_8/lib/python3.8/multiprocessing/context.py", > line 284, in _Popen > return Popen(process_obj) > File > "/Users/liangbinbin/Applications/anaconda3/envs/python3_8/lib/python3.8/multiprocessing/popen_spawn_posix.py", > line 32, in __init__ > super().__init__(process_obj) > File > "/Users/liangbinbin/Applications/anaconda3/envs/python3_8/lib/python3.8/multiprocessing/popen_fork.py", > line 19, in __init__ > self._launch(process_obj) > File > "/Users/liangbinbin/Applications/anaconda3/envs/python3_8/lib/python3.8/multiprocessing/popen_spawn_posix.py", > line 47, in _launch > reduction.dump(process_obj, fp) > File > "/Users/liangbinbin/Applications/anaconda3/envs/python3_8/lib/python3.8/multiprocessing/reduction.py", > line 60, in dump > ForkingPickler(file, protocol).dump(obj) > File "stringsource", line 2, in > pyarrow._plasma.PlasmaClient.__reduce_cython__ > TypeError: no default __reduce__ due to non-trivial __cinit__ > {code} > * And the error does not occur, when I change > 'multiprocessing.context.Process' to > '{color:#de350b}multiprocess.context.Process{color}' which use *dill* to > serialize. > -- This message was sent by Atlassian Jira (v8.20.10#820010)