有一个问题困扰了我几天。当我使用 spark-submit 运行 Spark 应用程序时,它会发生,如下所示:
09:14:36 Exception in thread "main" java.io.IOException: Cannot run program "python3.7": error=2, No such file or directory
09:14:36 at java.lang.ProcessBuilder.start(ProcessBuilder.java:1048)
09:14:36 at org.apache.spark.deploy.PythonRunner$.main(PythonRunner.scala:100)
09:14:36 at org.apache.spark.deploy.PythonRunner.main(PythonRunner.scala)
09:14:36 at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
09:14:36 at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
09:14:36 at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
09:14:36 at java.lang.reflect.Method.invoke(Method.java:498)
09:14:36 at org.apache.spark.deploy.JavaMainApplication.start(SparkApplication.scala:52)
09:14:36 at org.apache.spark.deploy.SparkSubmit$.org$apache$spark$deploy$SparkSubmit$$runMain(SparkSubmit.scala:894)
09:14:36 at org.apache.spark.deploy.SparkSubmit$.doRunMain$1(SparkSubmit.scala:198)
09:14:36 at org.apache.spark.deploy.SparkSubmit$.submit(SparkSubmit.scala:228)
09:14:36 at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:137)
09:14:36 at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)
09:14:36 Caused by: java.io.IOException: error=2, No such file or directory
09:14:36 at java.lang.UNIXProcess.forkAndExec(Native Method)
09:14:36 at java.lang.UNIXProcess.<init>(UNIXProcess.java:247)
09:14:36 at java.lang.ProcessImpl.start(ProcessImpl.java:134)
09:14:36 at java.lang.ProcessBuilder.start(ProcessBuilder.java:1029)
09:14:36 ... 12 more
我已经在~/.bashrc
和/etc/profile
中配置了所有环境变量。我绝对确定所有环境变量都是正确的。我可以在我的 Ubuntu 中运行python3.7
或 PySpark。
export JAVA_HOME=/usr/lib/jvm/java-8-oracle
export PATH=${PATH}:${JAVA_HOME}
export PATH=${PATH}:/mwdata/python3.7/bin
export SPARK_HOME=/srv/spark
export PATH=${PATH}:${SPARK_HOME}/bin
export PYSPARK_PYTHON=python3.7
export PYSPARK_DRIVER_PYTHON=python3.7
那么为什么找不到python3.7
呢?
PySpark的核心依赖于Py4J(当前版本0.10.7(
不幸的是,当前版本的py4j 0.10.7与Python 3.7不兼容。