温馨提示×

温馨提示×

您好,登录后才能下订单哦!

密码登录×
登录注册×
其他方式登录
点击 登录注册 即表示同意《亿速云用户服务条款》

Java spark中的bug分析

发布时间:2021-11-24 14:38:04 来源:亿速云 阅读:201 作者:iii 栏目:大数据

这篇文章主要介绍“Java spark中的bug分析”,在日常操作中,相信很多人在Java spark中的bug分析问题上存在疑惑,小编查阅了各式资料,整理出简单好用的操作方法,希望对大家解答”Java spark中的bug分析”的疑惑有所帮助!接下来,请跟着小编一起来学习吧!

在spark 中存在一个bug,该bug的详细信息如下:

None.get
java.util.NoSuchElementException: None.get
scala.None$.get(Option.scala:529)
scala.None$.get(Option.scala:527)
org.apache.spark.sql.execution.FileSourceScanExec.needsUnsafeRowConversion$lzycompute(DataSourceScanExec.scala:178)
org.apache.spark.sql.execution.FileSourceScanExec.needsUnsafeRowConversion(DataSourceScanExec.scala:176)
org.apache.spark.sql.execution.FileSourceScanExec.doExecute(DataSourceScanExec.scala:463)
org.apache.spark.sql.execution.SparkPlan.$anonfun$execute$1(SparkPlan.scala:175)
org.apache.spark.sql.execution.SparkPlan.$anonfun$executeQuery$1(SparkPlan.scala:213)
org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:151)
org.apache.spark.sql.execution.SparkPlan.executeQuery(SparkPlan.scala:210)
org.apache.spark.sql.execution.SparkPlan.execute(SparkPlan.scala:171)
org.apache.spark.sql.execution.InputAdapter.inputRDD(WholeStageCodegenExec.scala:525)
org.apache.spark.sql.execution.InputRDDCodegen.inputRDDs(WholeStageCodegenExec.scala:453)
org.apache.spark.sql.execution.InputRDDCodegen.inputRDDs$(WholeStageCodegenExec.scala:452)
org.apache.spark.sql.execution.InputAdapter.inputRDDs(WholeStageCodegenExec.scala:496)
org.apache.spark.sql.execution.FilterExec.inputRDDs(basicPhysicalOperators.scala:133)
org.apache.spark.sql.execution.ProjectExec.inputRDDs(basicPhysicalOperators.scala:47)
org.apache.spark.sql.execution.WholeStageCodegenExec.doExecute(WholeStageCodegenExec.scala:720)
org.apache.spark.sql.execution.SparkPlan.$anonfun$execute$1(SparkPlan.scala:175)
org.apache.spark.sql.execution.SparkPlan.$anonfun$executeQuery$1(SparkPlan.scala:213)
org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:151)
org.apache.spark.sql.execution.SparkPlan.executeQuery(SparkPlan.scala:210)
org.apache.spark.sql.execution.SparkPlan.execute(SparkPlan.scala:171)
org.apache.spark.sql.execution.DeserializeToObjectExec.doExecute(objects.scala:96)
org.apache.spark.sql.execution.SparkPlan.$anonfun$execute$1(SparkPlan.scala:175)
org.apache.spark.sql.execution.SparkPlan.$anonfun$executeQuery$1(SparkPlan.scala:213)
org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:151)
org.apache.spark.sql.execution.SparkPlan.executeQuery(SparkPlan.scala:210)
org.apache.spark.sql.execution.SparkPlan.execute(SparkPlan.scala:171)
org.apache.spark.sql.execution.QueryExecution.toRdd$lzycompute(QueryExecution.scala:122)
org.apache.spark.sql.execution.QueryExecution.toRdd(QueryExecution.scala:121)
org.apache.spark.sql.Dataset.rdd$lzycompute(Dataset.scala:3200)
org.apache.spark.sql.Dataset.rdd(Dataset.scala:3198)

根据源码定位FileSourceScanExec,定位到如下位置:

 SparkSession.getActiveSession.get.sessionState.conf.parquetVectorizedReaderEnabled

SparkSession.getActiveSession.get的内容如下:

/**
   * Returns the active SparkSession for the current thread, returned by the builder.
   *
   * @note Return None, when calling this function on executors
   *
   * @since 2.2.0
   */
  def getActiveSession: Option[SparkSession] = {
    if (TaskContext.get != null) {
      // Return None when running on executors.
      None
    } else {
      Option(activeThreadSession.get)
    }
  }

正如注释所写的一样,当在executors端获取SparkSession的时候,直接返回None。 为什么直接返回none,可以参考spark-pr-21436
当然这个问题,已经有人发现了并且提交了pr-29667,所以拿到commitID(37a660866342f2d64ad2990a5596e67cfdf044c0)直接cherry-pick就ok了,

分析一下原因: 其实该原因就是同一个jvm中,两个不同的线程同步调用,就如unit test所示:

test("SPARK-32813: Table scan should work in different thread") {
    val executor1 = Executors.newSingleThreadExecutor()
    val executor2 = Executors.newSingleThreadExecutor()
    var session: SparkSession = null
    SparkSession.cleanupAnyExistingSession()

    withTempDir { tempDir =>
      try {
        val tablePath = tempDir.toString + "/table"
        val df = ThreadUtils.awaitResult(Future {
          session = SparkSession.builder().appName("test").master("local[*]").getOrCreate()

          session.createDataFrame(
            session.sparkContext.parallelize(Row(Array(1, 2, 3)) :: Nil),
            StructType(Seq(
              StructField("a", ArrayType(IntegerType, containsNull = false), nullable = false))))
            .write.parquet(tablePath)

          session.read.parquet(tablePath)
        }(ExecutionContext.fromExecutorService(executor1)), 1.minute)

        ThreadUtils.awaitResult(Future {
          assert(df.rdd.collect()(0) === Row(Seq(1, 2, 3)))
        }(ExecutionContext.fromExecutorService(executor2)), 1.minute)
      } finally {
        executor1.shutdown()
        executor2.shutdown()
        session.stop()
      }
    }
  }

到此,关于“Java spark中的bug分析”的学习就结束了,希望能够解决大家的疑惑。理论与实践的搭配能更好的帮助大家学习,快去试试吧!若想继续学习更多相关知识,请继续关注亿速云网站,小编会继续努力为大家带来更多实用的文章!

向AI问一下细节

免责声明:本站发布的内容(图片、视频和文字)以原创、转载和分享为主,文章观点不代表本网站立场,如果涉及侵权请联系站长邮箱:is@yisu.com进行举报,并提供相关证据,一经查实,将立刻删除涉嫌侵权内容。

AI