Skip to content

[SPARK-12717][PYSPARK][BRANCH-1.6] Resolving race condition with pyspark broadcasts when using multiple threads #17722

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

[SPARK-12717][PYSPARK][BRANCH-1.6] Resolving race condition with pyspark broadcasts when using multiple threads #17722

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 13 additions & 9 deletions python/pyspark/rdd.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@
import socket
from subprocess import Popen, PIPE
from tempfile import NamedTemporaryFile
from threading import Thread
from threading import Thread, Lock
from collections import defaultdict
from itertools import chain
from functools import reduce
Expand Down Expand Up @@ -57,6 +57,9 @@

__all__ = ["RDD"]

# Lock which will make sure that dependend broadcast variables are pickled along
# with their PythonRDD wrapped function when using multple threads(SPARK-12717).
_lock = Lock()

def portable_hash(x):
"""
Expand Down Expand Up @@ -2374,14 +2377,15 @@ def _jrdd(self):
else:
profiler = None

command = (self.func, profiler, self._prev_jrdd_deserializer,
self._jrdd_deserializer)
pickled_cmd, bvars, env, includes = _prepare_for_python_RDD(self.ctx, command, self)
python_rdd = self.ctx._jvm.PythonRDD(self._prev_jrdd.rdd(),
bytearray(pickled_cmd),
env, includes, self.preservesPartitioning,
self.ctx.pythonExec, self.ctx.pythonVer,
bvars, self.ctx._javaAccumulator)
with _lock:
command = (self.func, profiler, self._prev_jrdd_deserializer,
self._jrdd_deserializer)
pickled_cmd, bvars, env, includes = _prepare_for_python_RDD(self.ctx, command, self)
python_rdd = self.ctx._jvm.PythonRDD(self._prev_jrdd.rdd(),
bytearray(pickled_cmd),
env, includes, self.preservesPartitioning,
self.ctx.pythonExec, self.ctx.pythonVer,
bvars, self.ctx._javaAccumulator)
self._jrdd_val = python_rdd.asJavaRDD()

if profiler:
Expand Down