Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
274 changes: 274 additions & 0 deletions core/src/main/java/org/apache/spark/TaskContext.java
Original file line number Diff line number Diff line change
@@ -0,0 +1,274 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/

package org.apache.spark;

import java.io.Serializable;
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;

import scala.Function0;
import scala.Function1;
import scala.Unit;
import scala.collection.JavaConversions;

import org.apache.spark.annotation.DeveloperApi;
import org.apache.spark.executor.TaskMetrics;
import org.apache.spark.util.TaskCompletionListener;
import org.apache.spark.util.TaskCompletionListenerException;

/**
* :: DeveloperApi ::
* Contextual information about a task which can be read or mutated during execution.
*/
@DeveloperApi
public class TaskContext implements Serializable {

private int stageId;
private int partitionId;
private long attemptId;
private boolean runningLocally;
private TaskMetrics taskMetrics;

/**
* :: DeveloperApi ::
* Contextual information about a task which can be read or mutated during execution.
*
* @param stageId stage id
* @param partitionId index of the partition
* @param attemptId the number of attempts to execute this task
* @param runningLocally whether the task is running locally in the driver JVM
* @param taskMetrics performance metrics of the task
*/
@DeveloperApi
public TaskContext(Integer stageId, Integer partitionId, Long attemptId, Boolean runningLocally,
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

fyi all these should be primitives rather than boxed Integers/Longs/Booleans. I'm going to fix it in a separate pr.

TaskMetrics taskMetrics) {
this.attemptId = attemptId;
this.partitionId = partitionId;
this.runningLocally = runningLocally;
this.stageId = stageId;
this.taskMetrics = taskMetrics;
}


Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

extra line

/**
* :: DeveloperApi ::
* Contextual information about a task which can be read or mutated during execution.
*
* @param stageId stage id
* @param partitionId index of the partition
* @param attemptId the number of attempts to execute this task
* @param runningLocally whether the task is running locally in the driver JVM
*/
@DeveloperApi
public TaskContext(Integer stageId, Integer partitionId, Long attemptId,
Boolean runningLocally) {
this.attemptId = attemptId;
this.partitionId = partitionId;
this.runningLocally = runningLocally;
this.stageId = stageId;
this.taskMetrics = TaskMetrics.empty();
}


/**
* :: DeveloperApi ::
* Contextual information about a task which can be read or mutated during execution.
*
* @param stageId stage id
* @param partitionId index of the partition
* @param attemptId the number of attempts to execute this task
*/
@DeveloperApi
public TaskContext(Integer stageId, Integer partitionId, Long attemptId) {
this.attemptId = attemptId;
this.partitionId = partitionId;
this.runningLocally = false;
this.stageId = stageId;
this.taskMetrics = TaskMetrics.empty();
}

private static ThreadLocal<TaskContext> taskContext =
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would remove this and the below function. I think it's better if a TaskContext instance is not itself aware of being stored as a thread local. We can keep all the logic related to thread locals in the static code.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But then its already static code ? + Where else should I put it ? I

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think Patrick meant putting it in executor or in Task's run method

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

alright

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But then how would someone access it ? I hope he wants to retain atleast the get method inside TaskContext as static.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes, get must be static. he's only thinking we should move the set to somewhere else

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

assuming i'm reading his mind correctly :)

new ThreadLocal<TaskContext>();

/**
* :: Internal API ::
* This is spark internal API, not intended to be called from user programs.
*/
public static void setTaskContext(TaskContext tc) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you add a warning in the javadoc that explains this should never be called by user programs? It would be nice if this could be private[spark] but unfortunately we are in java land :P

taskContext.set(tc);
}

public static TaskContext get() {
return taskContext.get();
}

/**
* :: Internal API ::
*/
public static void remove() {
taskContext.remove();
}

// List of callback functions to execute when the task completes.
private transient List<TaskCompletionListener> onCompleteCallbacks =
new ArrayList<TaskCompletionListener>();

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can u go through the file and use only one line to separate functions

// Whether the corresponding task has been killed.
private volatile Boolean interrupted = false;

// Whether the task has completed.
private volatile Boolean completed = false;

/**
* Checks whether the task has completed.
*/
public Boolean isCompleted() {
return completed;
}

/**
* Checks whether the task has been killed.
*/
public Boolean isInterrupted() {
return interrupted;
}

/**
* Add a (Java friendly) listener to be executed on task completion.
* This will be called in all situation - success, failure, or cancellation.
* <p/>
* An example use is for HadoopRDD to register a callback to close the input stream.
*/
public TaskContext addTaskCompletionListener(TaskCompletionListener listener) {
onCompleteCallbacks.add(listener);
return this;
}

/**
* Add a listener in the form of a Scala closure to be executed on task completion.
* This will be called in all situations - success, failure, or cancellation.
* <p/>
* An example use is for HadoopRDD to register a callback to close the input stream.
*/
public TaskContext addTaskCompletionListener(final Function1<TaskContext, Unit> f) {
onCompleteCallbacks.add(new TaskCompletionListener() {
@Override
public void onTaskCompletion(TaskContext context) {
f.apply(context);
}
});
return this;
}

/**
* Add a callback function to be executed on task completion. An example use
* is for HadoopRDD to register a callback to close the input stream.
* Will be called in any situation - success, failure, or cancellation.
*
* Deprecated: use addTaskCompletionListener
*
* @param f Callback function.
*/
@Deprecated
public void addOnCompleteCallback(final Function0<Unit> f) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We don't have the deprecation instructions any more - can you add them to the Javadoc? We should direct users to addTaskCompletionListener

onCompleteCallbacks.add(new TaskCompletionListener() {
@Override
public void onTaskCompletion(TaskContext context) {
f.apply();
}
});
}

/**
* ::Internal API::
* Marks the task as completed and triggers the listeners.
*/
public void markTaskCompleted() throws TaskCompletionListenerException {
completed = true;
List<String> errorMsgs = new ArrayList<String>(2);
// Process complete callbacks in the reverse order of registration
List<TaskCompletionListener> revlist =
new ArrayList<TaskCompletionListener>(onCompleteCallbacks);
Collections.reverse(revlist);
for (TaskCompletionListener tcl: revlist) {
try {
tcl.onTaskCompletion(this);
} catch (Throwable e) {
errorMsgs.add(e.getMessage());
}
}

if (!errorMsgs.isEmpty()) {
throw new TaskCompletionListenerException(JavaConversions.asScalaBuffer(errorMsgs));
}
}

/**
* ::Internal API::
* Marks the task for interruption, i.e. cancellation.
*/
public void markInterrupted() {
interrupted = true;
}

@Deprecated
/** Deprecated: use getStageId() */
public int stageId() {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In the future I think we'll use standard "Java Bean" getters for most classes we expose like this. I think it might be good to add these (getStageId, getPartitionId). We can keep the current ones for compatiblity (stageId) and deprecate them in favor of the new ones.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i.e.

@Deprecated
/** Deprecated: use getStageId() */
public int stageId() {

}

return stageId;
}

@Deprecated
/** Deprecated: use getPartitionId() */
public int partitionId() {
return partitionId;
}

@Deprecated
/** Deprecated: use getAttemptId() */
public long attemptId() {
return attemptId;
}

@Deprecated
/** Deprecated: use getRunningLocally() */
public boolean runningLocally() {
return runningLocally;
}

public boolean getRunningLocally() {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this should be isRunningLocally

return runningLocally;
}

public int getStageId() {
return stageId;
}

public int getPartitionId() {
return partitionId;
}

public long getAttemptId() {
return attemptId;
}

/** ::Internal API:: */
public TaskMetrics taskMetrics() {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can u try remove the public modifier and see if it still compiles? I think in that case it defaults to package private, which is the same as the Scala one

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It doesn't, because java is too strict about package private. The package name should be exact same does not expose it even in sub-packages.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok then we should document clearly that it is not a public api

return taskMetrics;
}
}
Loading