From 813e96e6faee44079eb52acbdc6c8aa58fb8d191 Mon Sep 17 00:00:00 2001 From: Sameer Agarwal Date: Fri, 8 Apr 2016 17:23:32 -0700 Subject: [SPARK-14454] Better exception handling while marking tasks as failed ## What changes were proposed in this pull request? This patch adds support for better handling of exceptions inside catch blocks if the code within the block throws an exception. For instance here is the code in a catch block before this change in `WriterContainer.scala`: ```scala logError("Aborting task.", cause) // call failure callbacks first, so we could have a chance to cleanup the writer. TaskContext.get().asInstanceOf[TaskContextImpl].markTaskFailed(cause) if (currentWriter != null) { currentWriter.close() } abortTask() throw new SparkException("Task failed while writing rows.", cause) ``` If `markTaskFailed` or `currentWriter.close` throws an exception, we currently lose the original cause. This PR fixes this problem by implementing a utility function `Utils.tryWithSafeCatch` that suppresses (`Throwable.addSuppressed`) the exception that are thrown within the catch block and rethrowing the original exception. ## How was this patch tested? No new functionality added Author: Sameer Agarwal Closes #12234 from sameeragarwal/fix-exception. --- .../org/apache/spark/rdd/PairRDDFunctions.scala | 8 ++---- .../scala/org/apache/spark/scheduler/Task.scala | 14 ++++++++--- .../main/scala/org/apache/spark/util/Utils.scala | 29 ++++++++++++++-------- 3 files changed, 31 insertions(+), 20 deletions(-) (limited to 'core') diff --git a/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala b/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala index 296179b75b..085829af6e 100644 --- a/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala +++ b/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala @@ -1111,9 +1111,7 @@ class PairRDDFunctions[K, V](self: RDD[(K, V)]) maybeUpdateOutputMetrics(outputMetricsAndBytesWrittenCallback, recordsWritten) recordsWritten += 1 } - } { - writer.close(hadoopContext) - } + }(finallyBlock = writer.close(hadoopContext)) committer.commitTask(hadoopContext) outputMetricsAndBytesWrittenCallback.foreach { case (om, callback) => om.setBytesWritten(callback()) @@ -1200,9 +1198,7 @@ class PairRDDFunctions[K, V](self: RDD[(K, V)]) maybeUpdateOutputMetrics(outputMetricsAndBytesWrittenCallback, recordsWritten) recordsWritten += 1 } - } { - writer.close() - } + }(finallyBlock = writer.close()) writer.commit() outputMetricsAndBytesWrittenCallback.foreach { case (om, callback) => om.setBytesWritten(callback()) diff --git a/core/src/main/scala/org/apache/spark/scheduler/Task.scala b/core/src/main/scala/org/apache/spark/scheduler/Task.scala index 46c64f61de..c91d8fbfc4 100644 --- a/core/src/main/scala/org/apache/spark/scheduler/Task.scala +++ b/core/src/main/scala/org/apache/spark/scheduler/Task.scala @@ -80,10 +80,16 @@ private[spark] abstract class Task[T]( } try { runTask(context) - } catch { case e: Throwable => - // Catch all errors; run task failure callbacks, and rethrow the exception. - context.markTaskFailed(e) - throw e + } catch { + case e: Throwable => + // Catch all errors; run task failure callbacks, and rethrow the exception. + try { + context.markTaskFailed(e) + } catch { + case t: Throwable => + e.addSuppressed(t) + } + throw e } finally { // Call the task completion callbacks. context.markTaskCompleted() diff --git a/core/src/main/scala/org/apache/spark/util/Utils.scala b/core/src/main/scala/org/apache/spark/util/Utils.scala index c304629bcd..78e164cff7 100644 --- a/core/src/main/scala/org/apache/spark/util/Utils.scala +++ b/core/src/main/scala/org/apache/spark/util/Utils.scala @@ -1260,26 +1260,35 @@ private[spark] object Utils extends Logging { } /** - * Execute a block of code, call the failure callbacks before finally block if there is any - * exceptions happen. But if exceptions happen in the finally block, do not suppress the original - * exception. + * Execute a block of code and call the failure callbacks in the catch block. If exceptions occur + * in either the catch or the finally block, they are appended to the list of suppressed + * exceptions in original exception which is then rethrown. * - * This is primarily an issue with `finally { out.close() }` blocks, where - * close needs to be called to clean up `out`, but if an exception happened - * in `out.write`, it's likely `out` may be corrupted and `out.close` will + * This is primarily an issue with `catch { abort() }` or `finally { out.close() }` blocks, + * where the abort/close needs to be called to clean up `out`, but if an exception happened + * in `out.write`, it's likely `out` may be corrupted and `abort` or `out.close` will * fail as well. This would then suppress the original/likely more meaningful * exception from the original `out.write` call. */ - def tryWithSafeFinallyAndFailureCallbacks[T](block: => T)(finallyBlock: => Unit): T = { + def tryWithSafeFinallyAndFailureCallbacks[T](block: => T) + (catchBlock: => Unit = (), finallyBlock: => Unit = ()): T = { var originalThrowable: Throwable = null try { block } catch { - case t: Throwable => + case cause: Throwable => // Purposefully not using NonFatal, because even fatal exceptions // we don't want to have our finallyBlock suppress - originalThrowable = t - TaskContext.get().asInstanceOf[TaskContextImpl].markTaskFailed(t) + originalThrowable = cause + try { + logError("Aborting task", originalThrowable) + TaskContext.get().asInstanceOf[TaskContextImpl].markTaskFailed(originalThrowable) + catchBlock + } catch { + case t: Throwable => + originalThrowable.addSuppressed(t) + logWarning(s"Suppressing exception in catch: " + t.getMessage, t) + } throw originalThrowable } finally { try { -- cgit v1.2.3