pyspark java.net.SocketException: Connection reset by peer
在window、运行pyspark训练模型,报错
Caused by: java.net.SocketException: Connection reset by peer: socket write error
py4j.protocol.Py4JJavaError: An error occurred while calling z:org.apache.spark.api.python.PythonRDD.runJob.
: org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 1.0 failed 1 times, most recent failure: Lost task 0.0 in stage 1.0 (TID 1, localhost, executor driver): java.net.SocketException: Connection reset by peer: socket write error
at java.net.SocketOutputStream.socketWrite0(Native Method)
at java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:109)
at java.net.SocketOutputStream.write(SocketOutputStream.java:153)
...
Caused by: java.net.SocketException: Connection reset by peer: socket write error
at java.net.SocketOutputStream.socketWrite0(Native Method)
at java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:109)
at java.net.SocketOutputStream.write(SocketOutputStream.java:153)
...
解决:修改python环境下的pyspark包中的worker. py 中的process函数
(注:python环境下的pyspark包是从spark目录拷贝过来的)
eg: E:\Anaconda3\envs\py27\Lib\site-packages\pyspark\worker. py
定位到 process 函数:
def process():
iterator = deserializer.load_stream(infile)
serializer.dump_stream(func(split_index, iterator), outfile)
# 添加以下内容
for obj in interator:
pass
refer:
https://issues.apache.org/jira/browse/SPARK-12261
https://stackoverflow.com/questions/26954566/apache-spark-pyspark-crash-for-large-dataset
end
还没有评论,来说两句吧...