Cassandra multiprocessing can't pickle _thread.lock objects

问题

I tried to use Cassandra and multiprocessing to insert rows (dummy data) concurrently based on the examples in

http://www.datastax.com/dev/blog/datastax-python-driver-multiprocessing-example-for-improved-bulk-data-throughput

This is my code

class QueryManager(object):

concurrency = 100  # chosen to match the default in execute_concurrent_with_args

def __init__(self, session, process_count=None):
    self.pool = Pool(processes=process_count, initializer=self._setup, initargs=(session,))

@classmethod
def _setup(cls, session):
    cls.session = session
    cls.prepared = cls.session.prepare("""
INSERT INTO test_table (key1, key2, key3, key4, key5) VALUES (?, ?, ?, ?, ?)
""")

def close_pool(self):
    self.pool.close()
    self.pool.join()

def get_results(self, params):
    results = self.pool.map(_multiprocess_write, (params[n:n+self.concurrency] for n in range(0, len(params), self.concurrency)))
    return list(itertools.chain(*results))

@classmethod
def _results_from_concurrent(cls, params):
    return [results[1] for results in execute_concurrent_with_args(cls.session, cls.prepared, params)]


def _multiprocess_write(params):
    return QueryManager._results_from_concurrent(params)


if __name__ == '__main__':

    processes = 2

    # connect cluster
    cluster = Cluster(contact_points=['127.0.0.1'], port=9042)
    session = cluster.connect()

    # database name is a concatenation of client_id and system_id
    keyspace_name = 'unit_test_0'

    # drop keyspace if it already exists in a cluster
    try:
        session.execute("DROP KEYSPACE IF EXISTS " + keyspace_name)
    except:
        pass

    create_keyspace_query = "CREATE KEYSPACE " + keyspace_name \
                        + " WITH replication = {'class': 'SimpleStrategy',    'replication_factor': '1'};"
    session.execute(create_keyspace_query)

    # use a session's keyspace
    session.set_keyspace(keyspace_name)

    # drop table if it already exists in the keyspace
    try:
        session.execute("DROP TABLE IF EXISTS " + "test_table")
    except:
        pass

    # create a table for invoices in the keyspace
    create_test_table = "CREATE TABLE test_table("

    keys = "key1 text,\n" \
           "key2 text,\n" \
           "key3 text,\n" \
           "key4 text,\n" \
           "key5 text,\n"

    create_invoice_table_query += keys
    create_invoice_table_query += "PRIMARY KEY (key1))"
    session.execute(create_test_table)

    qm = QueryManager(session, processes)

    params = list()
    for row in range(100000):
        key = 'test' + str(row)
        params.append([key, 'test', 'test', 'test', 'test'])

    start = time.time()
    rows = qm.get_results(params)
    delta = time.time() - start
    log.info(fm('Cassandra inserts 100k dummy rows for ', delta, ' secs'))

when I executed the code, I got the following error

TypeError: can't pickle _thread.lock objects

which pointed at

self.pool = Pool(processes=process_count, initializer=self._setup, initargs=(session,))

回答1:

That would suggest you're trying to serialize a lock over IPC boundaries. I think it might be because you're supplying a Session object as an argument to the worker initialization function. Make the init function create a new session in each worker process (see the "Session per Process" section in the blog post you cited).

回答2:

I know this already has an answer, but I wanted to highlight a couple of changes in the cassandra-driver package that make this code still not work properly with python 3.7 and the 3.18.0 cassandra-driver package.

If you look at the blog post that is linked. The __init__ function doesn't pass in the session, but it passes a cluster object. Even the cluster cannot be sent as an initarg anymore because it contains a lock. You need to create it inside the def _setup(cls): classmethod.

Second, execute_concurrent_with_args returns a ResultSet now and that also cannot be serialized. The older version of the cassandra-driver package just returned a list of objects.

To fix the above code change these 2 sections:

First, the __init__ and _setup methods

def __init__(self, process_count=None):
    self.pool = Pool(processes=process_count, initializer=self._setup)

@classmethod
def _setup(cls):
    cluster = Cluster()
    cls.session = cluster.connect()
    cls.prepared = cls.session.prepare("""
        INSERT INTO test_table (key1, key2, key3, key4, key5) VALUES (?, ?, ?, ?, ?)
        """)

Second, the _results_from_concurrent method

@classmethod
def _results_from_concurrent(cls, params):
    return [list(results[1]) for results in execute_concurrent_with_args(cls.session, cls.prepared, params)]

Lastly, if you are interested in a gist for the multiprocess_execute.py in the original DataStax blog post that works with python3 and cassandra-driver 3.18.0, you can find that here: https://gist.github.com/jWolo/6127b2e57c7e24740afd7a4254cc00a3

来源：https://stackoverflow.com/questions/37942249/cassandra-multiprocessing-cant-pickle-thread-lock-objects