Class BarrierTaskContext

Object
org.apache.spark.TaskContext
org.apache.spark.BarrierTaskContext
All Implemented Interfaces:
Serializable, org.apache.spark.internal.Logging

public class BarrierTaskContext extends TaskContext implements org.apache.spark.internal.Logging
:: Experimental :: A TaskContext with extra contextual info and tooling for tasks in a barrier stage. Use get() to obtain the barrier context for a running barrier task.
See Also:
  • Nested Class Summary

    Nested classes/interfaces inherited from interface org.apache.spark.internal.Logging

    org.apache.spark.internal.Logging.LogStringContext, org.apache.spark.internal.Logging.SparkShellLoggingFilter
  • Method Summary

    Modifier and Type
    Method
    Description
    Adds a (Java friendly) listener to be executed on task completion.
    Adds a listener to be executed on task failure (which includes completion listener failure, if the task body did not already fail).
    allGather(String message)
    :: Experimental :: Blocks until all tasks in the same stage have reached this routine.
    int
    How many times this task has been attempted.
    void
    :: Experimental :: Sets a global barrier and waits until all tasks in this stage hit this barrier.
    int
    CPUs allocated to the task.
    get()
    :: Experimental :: Returns the currently active BarrierTaskContext.
    Get a local property set upstream in the driver, or null if it is missing.
    scala.collection.immutable.Seq<Source>
    ::DeveloperApi:: Returns all metrics sources with the given name which are associated with the instance which runs the task.
    :: Experimental :: Returns BarrierTaskInfo for all tasks in this barrier stage, ordered by partition ID.
    boolean
    Returns true if the task has completed.
    boolean
    Returns true if the task has failed.
    boolean
    Returns true if the task has been killed.
    int
    Total number of partitions in the stage that this task belongs to.
    int
    The ID of the RDD partition that is computed by this task.
    scala.collection.immutable.Map<String,ResourceInformation>
    Resources allocated to the task.
    (java-specific) Resources allocated to the task.
    int
    How many times the stage that this task belongs to has been attempted.
    int
    The ID of the stage that this task belong to.
    long
    An ID that is unique to this task attempt (within the same SparkContext, no two task attempts will share the same attempt ID).
    org.apache.spark.executor.TaskMetrics
     

    Methods inherited from class org.apache.spark.TaskContext

    addTaskCompletionListener, addTaskFailureListener, getPartitionId

    Methods inherited from class java.lang.Object

    equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    Methods inherited from interface org.apache.spark.internal.Logging

    initializeForcefully, initializeLogIfNecessary, initializeLogIfNecessary, initializeLogIfNecessary$default$2, isTraceEnabled, log, logDebug, logDebug, logDebug, logDebug, logError, logError, logError, logError, logInfo, logInfo, logInfo, logInfo, logName, LogStringContext, logTrace, logTrace, logTrace, logTrace, logWarning, logWarning, logWarning, logWarning, org$apache$spark$internal$Logging$$log_, org$apache$spark$internal$Logging$$log__$eq, withLogContext
  • Method Details

    • get

      public static BarrierTaskContext get()
      :: Experimental :: Returns the currently active BarrierTaskContext. This can be called inside of user functions to access contextual information about running barrier tasks.
      Returns:
      (undocumented)
    • barrier

      public void barrier()
      :: Experimental :: Sets a global barrier and waits until all tasks in this stage hit this barrier. Similar to MPI_Barrier function in MPI, the barrier() function call blocks until all tasks in the same stage have reached this routine.

      CAUTION! In a barrier stage, each task must have the same number of barrier() calls, in all possible code branches. Otherwise, you may get the job hanging or a SparkException after timeout. Some examples of '''misuses''' are listed below: 1. Only call barrier() function on a subset of all the tasks in the same barrier stage, it shall lead to timeout of the function call.

      
         rdd.barrier().mapPartitions { iter =>
             val context = BarrierTaskContext.get()
             if (context.partitionId() == 0) {
                 // Do nothing.
             } else {
                 context.barrier()
             }
             iter
         }
       

      2. Include barrier() function in a try-catch code block, this may lead to timeout of the second function call.

      
         rdd.barrier().mapPartitions { iter =>
             val context = BarrierTaskContext.get()
             try {
                 // Do something that might throw an Exception.
                 doSomething()
                 context.barrier()
             } catch {
                 case e: Exception => logWarning("...", e)
             }
             context.barrier()
             iter
         }
       
    • allGather

      public String[] allGather(String message)
      :: Experimental :: Blocks until all tasks in the same stage have reached this routine. Each task passes in a message and returns with a list of all the messages passed in by each of those tasks.

      CAUTION! The allGather method requires the same precautions as the barrier method

      The message is type String rather than Array[Byte] because it is more convenient for the user at the cost of worse performance.

      Parameters:
      message - (undocumented)
      Returns:
      (undocumented)
    • getTaskInfos

      public BarrierTaskInfo[] getTaskInfos()
      :: Experimental :: Returns BarrierTaskInfo for all tasks in this barrier stage, ordered by partition ID.
      Returns:
      (undocumented)
    • isCompleted

      public boolean isCompleted()
      Description copied from class: TaskContext
      Returns true if the task has completed.
      Specified by:
      isCompleted in class TaskContext
      Returns:
      (undocumented)
    • isFailed

      public boolean isFailed()
      Description copied from class: TaskContext
      Returns true if the task has failed.
      Specified by:
      isFailed in class TaskContext
      Returns:
      (undocumented)
    • isInterrupted

      public boolean isInterrupted()
      Description copied from class: TaskContext
      Returns true if the task has been killed.
      Specified by:
      isInterrupted in class TaskContext
      Returns:
      (undocumented)
    • addTaskCompletionListener

      public BarrierTaskContext addTaskCompletionListener(TaskCompletionListener listener)
      Description copied from class: TaskContext
      Adds a (Java friendly) listener to be executed on task completion. This will be called in all situations - success, failure, or cancellation. Adding a listener to an already completed task will result in that listener being called immediately.

      Two listeners registered in the same thread will be invoked in reverse order of registration if the task completes after both are registered. There are no ordering guarantees for listeners registered in different threads, or for listeners registered after the task completes. Listeners are guaranteed to execute sequentially.

      An example use is for HadoopRDD to register a callback to close the input stream.

      Exceptions thrown by the listener will result in failure of the task.

      Specified by:
      addTaskCompletionListener in class TaskContext
      Parameters:
      listener - (undocumented)
      Returns:
      (undocumented)
    • addTaskFailureListener

      public BarrierTaskContext addTaskFailureListener(TaskFailureListener listener)
      Description copied from class: TaskContext
      Adds a listener to be executed on task failure (which includes completion listener failure, if the task body did not already fail). Adding a listener to an already failed task will result in that listener being called immediately.

      Note: Prior to Spark 3.4.0, failure listeners were only invoked if the main task body failed.

      Specified by:
      addTaskFailureListener in class TaskContext
      Parameters:
      listener - (undocumented)
      Returns:
      (undocumented)
    • stageId

      public int stageId()
      Description copied from class: TaskContext
      The ID of the stage that this task belong to.
      Specified by:
      stageId in class TaskContext
      Returns:
      (undocumented)
    • stageAttemptNumber

      public int stageAttemptNumber()
      Description copied from class: TaskContext
      How many times the stage that this task belongs to has been attempted. The first stage attempt will be assigned stageAttemptNumber = 0, and subsequent attempts will have increasing attempt numbers.
      Specified by:
      stageAttemptNumber in class TaskContext
      Returns:
      (undocumented)
    • partitionId

      public int partitionId()
      Description copied from class: TaskContext
      The ID of the RDD partition that is computed by this task.
      Specified by:
      partitionId in class TaskContext
      Returns:
      (undocumented)
    • numPartitions

      public int numPartitions()
      Description copied from class: TaskContext
      Total number of partitions in the stage that this task belongs to.
      Specified by:
      numPartitions in class TaskContext
      Returns:
      (undocumented)
    • attemptNumber

      public int attemptNumber()
      Description copied from class: TaskContext
      How many times this task has been attempted. The first task attempt will be assigned attemptNumber = 0, and subsequent attempts will have increasing attempt numbers.
      Specified by:
      attemptNumber in class TaskContext
      Returns:
      (undocumented)
    • taskAttemptId

      public long taskAttemptId()
      Description copied from class: TaskContext
      An ID that is unique to this task attempt (within the same SparkContext, no two task attempts will share the same attempt ID). This is roughly equivalent to Hadoop's TaskAttemptID.
      Specified by:
      taskAttemptId in class TaskContext
      Returns:
      (undocumented)
    • getLocalProperty

      public String getLocalProperty(String key)
      Description copied from class: TaskContext
      Get a local property set upstream in the driver, or null if it is missing. See also org.apache.spark.SparkContext.setLocalProperty.
      Specified by:
      getLocalProperty in class TaskContext
      Parameters:
      key - (undocumented)
      Returns:
      (undocumented)
    • taskMetrics

      public org.apache.spark.executor.TaskMetrics taskMetrics()
      Specified by:
      taskMetrics in class TaskContext
    • getMetricsSources

      public scala.collection.immutable.Seq<Source> getMetricsSources(String sourceName)
      Description copied from class: TaskContext
      ::DeveloperApi:: Returns all metrics sources with the given name which are associated with the instance which runs the task. For more information see org.apache.spark.metrics.MetricsSystem.
      Specified by:
      getMetricsSources in class TaskContext
      Parameters:
      sourceName - (undocumented)
      Returns:
      (undocumented)
    • cpus

      public int cpus()
      Description copied from class: TaskContext
      CPUs allocated to the task.
      Specified by:
      cpus in class TaskContext
      Returns:
      (undocumented)
    • resources

      public scala.collection.immutable.Map<String,ResourceInformation> resources()
      Description copied from class: TaskContext
      Resources allocated to the task. The key is the resource name and the value is information about the resource. Please refer to ResourceInformation for specifics.
      Specified by:
      resources in class TaskContext
      Returns:
      (undocumented)
    • resourcesJMap

      public Map<String,ResourceInformation> resourcesJMap()
      Description copied from class: TaskContext
      (java-specific) Resources allocated to the task. The key is the resource name and the value is information about the resource. Please refer to ResourceInformation for specifics.
      Specified by:
      resourcesJMap in class TaskContext
      Returns:
      (undocumented)