Xiao Li created SPARK-28485:
-------------------------------
Summary: PythonBroadcast may delete the broadcast file while a
Python worker still needs it
Key: SPARK-28485
URL: https://issues.apache.org/jira/browse/SPARK-28485
Project: Spark
Issue Type: Improvement
Components: PySpark
Affects Versions: 3.0.0
Reporter: Xiao Li
How to reproduce:
Run "bin/pyspark --master local[1,1] --conf spark.memory.fraction=0.0001" to
start PySpark and run the following codes:
{code:java}
b = sc.broadcast([100])
sc.parallelize([0],1).map(lambda x: 0 if x == 0 else b.value[0]).collect()
sc._jvm.java.lang.System.gc()
import time time.sleep(5)
sc._jvm.java.lang.System.gc()
time.sleep(5)
sc.parallelize([1],1).map(lambda x: 0 if x == 0 else b.value[0]).collect(){code}
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
