From 5bdd9e353287c2a326138b25f80ee255d15942b0 Mon Sep 17 00:00:00 2001 From: Xingbo Jiang Date: Thu, 23 Aug 2018 14:22:56 -0700 Subject: [PATCH 1/2] [SPARK-23207][SPARK-22905][SPARK-24564][SPARK-25114][SQL][BACKPORT-2.1] Shuffle+Repartition on a DataFrame could lead to incorrect answers ## What changes were proposed in this pull request? Back port of #20393 and #22079. Currently shuffle repartition uses RoundRobinPartitioning, the generated result is nondeterministic since the sequence of input rows are not determined. The bug can be triggered when there is a repartition call following a shuffle (which would lead to non-deterministic row ordering), as the pattern shows below: upstream stage -> repartition stage -> result stage (-> indicate a shuffle) When one of the executors process goes down, some tasks on the repartition stage will be retried and generate inconsistent ordering, and some tasks of the result stage will be retried generating different data. The following code returns 931532, instead of 1000000: ``` import scala.sys.process._ import org.apache.spark.TaskContext val res = spark.range(0, 1000 * 1000, 1).repartition(200).map { x => x }.repartition(200).map { x => if (TaskContext.get.attemptNumber == 0 && TaskContext.get.partitionId < 2) { throw new Exception("pkill -f java".!!) } x } res.distinct().count() ``` In this PR, we propose a most straight-forward way to fix this problem by performing a local sort before partitioning, after we make the input row ordering deterministic, the function from rows to partitions is fully deterministic too. The downside of the approach is that with extra local sort inserted, the performance of repartition() will go down, so we add a new config named `spark.sql.execution.sortBeforeRepartition` to control whether this patch is applied. The patch is default enabled to be safe-by-default, but user may choose to manually turn it off to avoid performance regression. This patch also changes the output rows ordering of repartition(), that leads to a bunch of test cases failure because they are comparing the results directly. Add unit test in ExchangeSuite. With this patch(and `spark.sql.execution.sortBeforeRepartition` set to true), the following query returns 1000000: ``` import scala.sys.process._ import org.apache.spark.TaskContext spark.conf.set("spark.sql.execution.sortBeforeRepartition", "true") val res = spark.range(0, 1000 * 1000, 1).repartition(200).map { x => x }.repartition(200).map { x => if (TaskContext.get.attemptNumber == 0 && TaskContext.get.partitionId < 2) { throw new Exception("pkill -f java".!!) } x } res.distinct().count() res7: Long = 1000000 ``` Author: Xingbo Jiang ## How was this patch tested? Ran all SBT unit tests for org.apache.spark.sql.*. Ran pyspark tests for module pyspark-sql. Closes #22079 from bersprockets/SPARK-23207. Lead-authored-by: Xingbo Jiang Co-authored-by: Bruce Robbins Co-authored-by: Zheng RuiFeng Signed-off-by: Xiao Li --- .../unsafe/sort/RecordComparator.java | 4 +- .../unsafe/sort/UnsafeInMemorySorter.java | 7 +- .../unsafe/sort/UnsafeSorterSpillMerger.java | 4 +- .../main/scala/org/apache/spark/rdd/RDD.scala | 2 + .../spark/memory/TestMemoryConsumer.java | 10 + .../sort/UnsafeExternalSorterSuite.java | 4 +- .../sort/UnsafeInMemorySorterSuite.java | 4 +- .../clustering/GaussianMixtureModel.scala | 2 +- .../spark/mllib/feature/ChiSqSelector.scala | 2 +- .../spark/ml/feature/Word2VecSuite.scala | 3 +- .../sql/execution/RecordBinaryComparator.java | 74 ++++ .../execution/UnsafeExternalRowSorter.java | 49 ++- .../apache/spark/sql/internal/SQLConf.scala | 14 + .../sql/execution/UnsafeKVExternalSorter.java | 10 +- .../apache/spark/sql/execution/SortExec.scala | 2 +- .../execution/exchange/ShuffleExchange.scala | 55 ++- .../sort/RecordBinaryComparatorSuite.java | 322 ++++++++++++++++++ .../spark/sql/execution/ExchangeSuite.scala | 26 +- .../datasources/parquet/ParquetIOSuite.scala | 6 +- .../streaming/ForeachSinkSuite.scala | 4 +- 20 files changed, 574 insertions(+), 30 deletions(-) create mode 100644 sql/catalyst/src/main/java/org/apache/spark/sql/execution/RecordBinaryComparator.java create mode 100644 sql/core/src/test/java/test/org/apache/spark/sql/execution/sort/RecordBinaryComparatorSuite.java diff --git a/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/RecordComparator.java b/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/RecordComparator.java index 09e4258792204..02b5de8e128c9 100644 --- a/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/RecordComparator.java +++ b/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/RecordComparator.java @@ -32,6 +32,8 @@ public abstract class RecordComparator { public abstract int compare( Object leftBaseObject, long leftBaseOffset, + int leftBaseLength, Object rightBaseObject, - long rightBaseOffset); + long rightBaseOffset, + int rightBaseLength); } diff --git a/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorter.java b/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorter.java index 5b42843717e90..87bd186b41529 100644 --- a/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorter.java +++ b/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorter.java @@ -62,12 +62,13 @@ public int compare(RecordPointerAndKeyPrefix r1, RecordPointerAndKeyPrefix r2) { int uaoSize = UnsafeAlignedOffset.getUaoSize(); if (prefixComparisonResult == 0) { final Object baseObject1 = memoryManager.getPage(r1.recordPointer); - // skip length final long baseOffset1 = memoryManager.getOffsetInPage(r1.recordPointer) + uaoSize; + final int baseLength1 = UnsafeAlignedOffset.getSize(baseObject1, baseOffset1 - uaoSize); final Object baseObject2 = memoryManager.getPage(r2.recordPointer); - // skip length final long baseOffset2 = memoryManager.getOffsetInPage(r2.recordPointer) + uaoSize; - return recordComparator.compare(baseObject1, baseOffset1, baseObject2, baseOffset2); + final int baseLength2 = UnsafeAlignedOffset.getSize(baseObject2, baseOffset2 - uaoSize); + return recordComparator.compare(baseObject1, baseOffset1, baseLength1, baseObject2, + baseOffset2, baseLength2); } else { return prefixComparisonResult; } diff --git a/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeSorterSpillMerger.java b/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeSorterSpillMerger.java index 01aed95878cf6..f17d79e50b052 100644 --- a/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeSorterSpillMerger.java +++ b/core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeSorterSpillMerger.java @@ -38,8 +38,8 @@ public int compare(UnsafeSorterIterator left, UnsafeSorterIterator right) { prefixComparator.compare(left.getKeyPrefix(), right.getKeyPrefix()); if (prefixComparisonResult == 0) { return recordComparator.compare( - left.getBaseObject(), left.getBaseOffset(), - right.getBaseObject(), right.getBaseOffset()); + left.getBaseObject(), left.getBaseOffset(), left.getRecordLength(), + right.getBaseObject(), right.getBaseOffset(), right.getRecordLength()); } else { return prefixComparisonResult; } diff --git a/core/src/main/scala/org/apache/spark/rdd/RDD.scala b/core/src/main/scala/org/apache/spark/rdd/RDD.scala index 199a3770ac4ad..283545b7272da 100644 --- a/core/src/main/scala/org/apache/spark/rdd/RDD.scala +++ b/core/src/main/scala/org/apache/spark/rdd/RDD.scala @@ -413,6 +413,8 @@ abstract class RDD[T: ClassTag]( * * If you are decreasing the number of partitions in this RDD, consider using `coalesce`, * which can avoid performing a shuffle. + * + * TODO Fix the Shuffle+Repartition data loss issue described in SPARK-23207. */ def repartition(numPartitions: Int)(implicit ord: Ordering[T] = null): RDD[T] = withScope { coalesce(numPartitions, shuffle = true) diff --git a/core/src/test/java/org/apache/spark/memory/TestMemoryConsumer.java b/core/src/test/java/org/apache/spark/memory/TestMemoryConsumer.java index db91329c94cb6..0bbaea6b834b8 100644 --- a/core/src/test/java/org/apache/spark/memory/TestMemoryConsumer.java +++ b/core/src/test/java/org/apache/spark/memory/TestMemoryConsumer.java @@ -17,6 +17,10 @@ package org.apache.spark.memory; +import com.google.common.annotations.VisibleForTesting; + +import org.apache.spark.unsafe.memory.MemoryBlock; + import java.io.IOException; public class TestMemoryConsumer extends MemoryConsumer { @@ -43,6 +47,12 @@ void free(long size) { used -= size; taskMemoryManager.releaseExecutionMemory(size, this); } + + @VisibleForTesting + public void freePage(MemoryBlock page) { + used -= page.size(); + taskMemoryManager.freePage(page, this); + } } diff --git a/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeExternalSorterSuite.java b/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeExternalSorterSuite.java index fbbe530a132e1..ec890d1ba96db 100644 --- a/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeExternalSorterSuite.java +++ b/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeExternalSorterSuite.java @@ -73,8 +73,10 @@ public class UnsafeExternalSorterSuite { public int compare( Object leftBaseObject, long leftBaseOffset, + int leftBaseLength, Object rightBaseObject, - long rightBaseOffset) { + long rightBaseOffset, + int rightBaseLength) { return 0; } }; diff --git a/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorterSuite.java b/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorterSuite.java index bd89085aa9a14..377d4a36ce22c 100644 --- a/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorterSuite.java +++ b/core/src/test/java/org/apache/spark/util/collection/unsafe/sort/UnsafeInMemorySorterSuite.java @@ -97,8 +97,10 @@ public void testSortingOnlyByIntegerPrefix() throws Exception { public int compare( Object leftBaseObject, long leftBaseOffset, + int leftBaseLength, Object rightBaseObject, - long rightBaseOffset) { + long rightBaseOffset, + int rightBaseLength) { return 0; } }; diff --git a/mllib/src/main/scala/org/apache/spark/mllib/clustering/GaussianMixtureModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/clustering/GaussianMixtureModel.scala index afbe4f978b286..1933d5499c3bf 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/clustering/GaussianMixtureModel.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/clustering/GaussianMixtureModel.scala @@ -154,7 +154,7 @@ object GaussianMixtureModel extends Loader[GaussianMixtureModel] { val dataArray = Array.tabulate(weights.length) { i => Data(weights(i), gaussians(i).mu, gaussians(i).sigma) } - spark.createDataFrame(dataArray).repartition(1).write.parquet(Loader.dataPath(path)) + spark.createDataFrame(sc.makeRDD(dataArray, 1)).write.parquet(Loader.dataPath(path)) } def load(sc: SparkContext, path: String): GaussianMixtureModel = { diff --git a/mllib/src/main/scala/org/apache/spark/mllib/feature/ChiSqSelector.scala b/mllib/src/main/scala/org/apache/spark/mllib/feature/ChiSqSelector.scala index 7ef2a95b96f2d..95ae1547a9021 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/feature/ChiSqSelector.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/feature/ChiSqSelector.scala @@ -144,7 +144,7 @@ object ChiSqSelectorModel extends Loader[ChiSqSelectorModel] { val dataArray = Array.tabulate(model.selectedFeatures.length) { i => Data(model.selectedFeatures(i)) } - spark.createDataFrame(dataArray).repartition(1).write.parquet(Loader.dataPath(path)) + spark.createDataFrame(sc.makeRDD(dataArray, 1)).write.parquet(Loader.dataPath(path)) } def load(sc: SparkContext, path: String): ChiSqSelectorModel = { diff --git a/mllib/src/test/scala/org/apache/spark/ml/feature/Word2VecSuite.scala b/mllib/src/test/scala/org/apache/spark/ml/feature/Word2VecSuite.scala index 613cc3d60b227..78efa0d6772b0 100644 --- a/mllib/src/test/scala/org/apache/spark/ml/feature/Word2VecSuite.scala +++ b/mllib/src/test/scala/org/apache/spark/ml/feature/Word2VecSuite.scala @@ -205,7 +205,8 @@ class Word2VecSuite extends SparkFunSuite with MLlibTestSparkContext with Defaul val oldModel = new OldWord2VecModel(word2VecMap) val instance = new Word2VecModel("myWord2VecModel", oldModel) val newInstance = testDefaultReadWrite(instance) - assert(newInstance.getVectors.collect() === instance.getVectors.collect()) + assert(newInstance.getVectors.collect().sortBy(_.getString(0)) === + instance.getVectors.collect().sortBy(_.getString(0))) } test("Word2Vec works with input that is non-nullable (NGram)") { diff --git a/sql/catalyst/src/main/java/org/apache/spark/sql/execution/RecordBinaryComparator.java b/sql/catalyst/src/main/java/org/apache/spark/sql/execution/RecordBinaryComparator.java new file mode 100644 index 0000000000000..40c2cc806e87a --- /dev/null +++ b/sql/catalyst/src/main/java/org/apache/spark/sql/execution/RecordBinaryComparator.java @@ -0,0 +1,74 @@ +/* + * 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.sql.execution; + +import org.apache.spark.unsafe.Platform; +import org.apache.spark.util.collection.unsafe.sort.RecordComparator; + +public final class RecordBinaryComparator extends RecordComparator { + + @Override + public int compare( + Object leftObj, long leftOff, int leftLen, Object rightObj, long rightOff, int rightLen) { + int i = 0; + + // If the arrays have different length, the longer one is larger. + if (leftLen != rightLen) { + return leftLen - rightLen; + } + + // The following logic uses `leftLen` as the length for both `leftObj` and `rightObj`, since + // we have guaranteed `leftLen` == `rightLen`. + + // check if stars align and we can get both offsets to be aligned + if ((leftOff % 8) == (rightOff % 8)) { + while ((leftOff + i) % 8 != 0 && i < leftLen) { + final int v1 = Platform.getByte(leftObj, leftOff + i) & 0xff; + final int v2 = Platform.getByte(rightObj, rightOff + i) & 0xff; + if (v1 != v2) { + return v1 > v2 ? 1 : -1; + } + i += 1; + } + } + // for architectures that support unaligned accesses, chew it up 8 bytes at a time + if (Platform.unaligned() || (((leftOff + i) % 8 == 0) && ((rightOff + i) % 8 == 0))) { + while (i <= leftLen - 8) { + final long v1 = Platform.getLong(leftObj, leftOff + i); + final long v2 = Platform.getLong(rightObj, rightOff + i); + if (v1 != v2) { + return v1 > v2 ? 1 : -1; + } + i += 8; + } + } + // this will finish off the unaligned comparisons, or do the entire aligned comparison + // whichever is needed. + while (i < leftLen) { + final int v1 = Platform.getByte(leftObj, leftOff + i) & 0xff; + final int v2 = Platform.getByte(rightObj, rightOff + i) & 0xff; + if (v1 != v2) { + return v1 > v2 ? 1 : -1; + } + i += 1; + } + + // The two arrays are equal. + return 0; + } +} diff --git a/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java b/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java index c29b002a998ca..51baa34d894bc 100644 --- a/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java +++ b/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java @@ -18,6 +18,7 @@ package org.apache.spark.sql.execution; import java.io.IOException; +import java.util.function.Supplier; import scala.collection.Iterator; import scala.math.Ordering; @@ -55,22 +56,47 @@ public abstract static class PrefixComputer { public static class Prefix { /** Key prefix value, or the null prefix value if isNull = true. **/ - long value; + public long value; /** Whether the key is null. */ - boolean isNull; + public boolean isNull; } /** * Computes prefix for the given row. For efficiency, the returned object may be reused in * further calls to a given PrefixComputer. */ - abstract Prefix computePrefix(InternalRow row); + public abstract Prefix computePrefix(InternalRow row); } - public UnsafeExternalRowSorter( + public static UnsafeExternalRowSorter createWithRecordComparator( StructType schema, - Ordering ordering, + Supplier recordComparatorSupplier, + PrefixComparator prefixComparator, + PrefixComputer prefixComputer, + long pageSizeBytes, + boolean canUseRadixSort) throws IOException { + return new UnsafeExternalRowSorter(schema, recordComparatorSupplier, prefixComparator, + prefixComputer, pageSizeBytes, canUseRadixSort); + } + + public static UnsafeExternalRowSorter create( + final StructType schema, + final Ordering ordering, + PrefixComparator prefixComparator, + PrefixComputer prefixComputer, + long pageSizeBytes, + boolean canUseRadixSort) throws IOException { + Supplier recordComparatorSupplier = new Supplier() { + public RecordComparator get() { return new RowComparator(ordering, schema.length()); } + }; + return new UnsafeExternalRowSorter(schema, recordComparatorSupplier, prefixComparator, + prefixComputer, pageSizeBytes, canUseRadixSort); + } + + private UnsafeExternalRowSorter( + StructType schema, + Supplier recordComparatorSupplier, PrefixComparator prefixComparator, PrefixComputer prefixComputer, long pageSizeBytes, @@ -84,7 +110,7 @@ public UnsafeExternalRowSorter( sparkEnv.blockManager(), sparkEnv.serializerManager(), taskContext, - new RowComparator(ordering, schema.length()), + recordComparatorSupplier.get(), prefixComparator, sparkEnv.conf().getInt("spark.shuffle.sort.initialBufferSize", DEFAULT_INITIAL_SORT_BUFFER_SIZE), @@ -207,8 +233,15 @@ private static final class RowComparator extends RecordComparator { } @Override - public int compare(Object baseObj1, long baseOff1, Object baseObj2, long baseOff2) { - // TODO: Why are the sizes -1? + public int compare( + Object baseObj1, + long baseOff1, + int baseLen1, + Object baseObj2, + long baseOff2, + int baseLen2) { + // Note that since ordering doesn't need the total length of the record, we just pass -1 + // into the row. row1.pointTo(baseObj1, baseOff1, -1); row2.pointTo(baseObj2, baseOff2, -1); return ordering.compare(row1, row2); diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala index ed61ff0292db7..b8a32f290e785 100644 --- a/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala +++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala @@ -641,6 +641,18 @@ object SQLConf { .booleanConf .createWithDefault(false) + val SORT_BEFORE_REPARTITION = + SQLConfigBuilder("spark.sql.execution.sortBeforeRepartition") + .internal() + .doc("When perform a repartition following a shuffle, the output row ordering would be " + + "nondeterministic. If some downstream stages fail and some tasks of the repartition " + + "stage retry, these tasks may generate different data, and that can lead to correctness " + + "issues. Turn on this config to insert a local sort before actually doing repartition " + + "to generate consistent repartition results. The performance of repartition() may go " + + "down since we insert extra local sort before it.") + .booleanConf + .createWithDefault(true) + object Deprecated { val MAPRED_REDUCE_TASKS = "mapred.reduce.tasks" } @@ -758,6 +770,8 @@ class SQLConf extends Serializable with Logging { def caseSensitiveAnalysis: Boolean = getConf(SQLConf.CASE_SENSITIVE) + def sortBeforeRepartition: Boolean = getConf(SORT_BEFORE_REPARTITION) + /** * Returns the [[Resolver]] for the current configuration, which can be used to determine if two * identifiers are equal. diff --git a/sql/core/src/main/java/org/apache/spark/sql/execution/UnsafeKVExternalSorter.java b/sql/core/src/main/java/org/apache/spark/sql/execution/UnsafeKVExternalSorter.java index 0d51dc9ff8a85..2d97a2a0dd57b 100644 --- a/sql/core/src/main/java/org/apache/spark/sql/execution/UnsafeKVExternalSorter.java +++ b/sql/core/src/main/java/org/apache/spark/sql/execution/UnsafeKVExternalSorter.java @@ -235,8 +235,14 @@ private static final class KVComparator extends RecordComparator { } @Override - public int compare(Object baseObj1, long baseOff1, Object baseObj2, long baseOff2) { - // Note that since ordering doesn't need the total length of the record, we just pass -1 + public int compare( + Object baseObj1, + long baseOff1, + int baseLen1, + Object baseObj2, + long baseOff2, + int baseLen2) { + // Note that since ordering doesn't need the total length of the record, we just pass -1 // into the row. row1.pointTo(baseObj1, baseOff1 + 4, -1); row2.pointTo(baseObj2, baseOff2 + 4, -1); diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/SortExec.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/SortExec.scala index cc576bbc4c802..197ab24de6b76 100644 --- a/sql/core/src/main/scala/org/apache/spark/sql/execution/SortExec.scala +++ b/sql/core/src/main/scala/org/apache/spark/sql/execution/SortExec.scala @@ -84,7 +84,7 @@ case class SortExec( } val pageSize = SparkEnv.get.memoryManager.pageSizeBytes - val sorter = new UnsafeExternalRowSorter( + val sorter = UnsafeExternalRowSorter.create( schema, ordering, prefixComparator, prefixComputer, pageSize, canUseRadixSort) if (testSpillFrequency > 0) { diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala index 125a4930c6528..ac069087b0c49 100644 --- a/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala +++ b/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala @@ -18,6 +18,7 @@ package org.apache.spark.sql.execution.exchange import java.util.Random +import java.util.function.Supplier import org.apache.spark._ import org.apache.spark.rdd.RDD @@ -30,7 +31,10 @@ import org.apache.spark.sql.catalyst.expressions.codegen.LazilyGeneratedOrdering import org.apache.spark.sql.catalyst.plans.physical._ import org.apache.spark.sql.execution._ import org.apache.spark.sql.execution.metric.SQLMetrics +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types.StructType import org.apache.spark.util.MutablePair +import org.apache.spark.util.collection.unsafe.sort.{PrefixComparators, RecordComparator} /** * Performs a shuffle that will result in the desired `newPartitioning`. @@ -239,14 +243,61 @@ object ShuffleExchange { case RangePartitioning(_, _) | SinglePartition => identity case _ => sys.error(s"Exchange not implemented for $newPartitioning") } + val rddWithPartitionIds: RDD[Product2[Int, InternalRow]] = { - if (needToCopyObjectsBeforeShuffle(part, serializer)) { + // [SPARK-23207] Have to make sure the generated RoundRobinPartitioning is deterministic, + // otherwise a retry task may output different rows and thus lead to data loss. + // + // Currently we following the most straight-forward way that perform a local sort before + // partitioning. + // + // Note that we don't perform local sort if the new partitioning has only 1 partition, under + // that case all output rows go to the same partition. + val newRdd = if (SparkEnv.get.conf.get(SQLConf.SORT_BEFORE_REPARTITION) && + newPartitioning.numPartitions > 1 && + newPartitioning.isInstanceOf[RoundRobinPartitioning]) { rdd.mapPartitionsInternal { iter => + val recordComparatorSupplier = new Supplier[RecordComparator] { + override def get: RecordComparator = new RecordBinaryComparator() + } + // The comparator for comparing row hashcode, which should always be Integer. + val prefixComparator = PrefixComparators.LONG + val canUseRadixSort = SparkEnv.get.conf.get(SQLConf.RADIX_SORT_ENABLED) + // The prefix computer generates row hashcode as the prefix, so we may decrease the + // probability that the prefixes are equal when input rows choose column values from a + // limited range. + val prefixComputer = new UnsafeExternalRowSorter.PrefixComputer { + private val result = new UnsafeExternalRowSorter.PrefixComputer.Prefix + override def computePrefix(row: InternalRow): + UnsafeExternalRowSorter.PrefixComputer.Prefix = { + // The hashcode generated from the binary form of a [[UnsafeRow]] should not be null. + result.isNull = false + result.value = row.hashCode() + result + } + } + val pageSize = SparkEnv.get.memoryManager.pageSizeBytes + + val sorter = UnsafeExternalRowSorter.createWithRecordComparator( + StructType.fromAttributes(outputAttributes), + recordComparatorSupplier, + prefixComparator, + prefixComputer, + pageSize, + canUseRadixSort) + sorter.sort(iter.asInstanceOf[Iterator[UnsafeRow]]) + } + } else { + rdd + } + + if (needToCopyObjectsBeforeShuffle(part, serializer)) { + newRdd.mapPartitionsInternal { iter => val getPartitionKey = getPartitionKeyExtractor() iter.map { row => (part.getPartition(getPartitionKey(row)), row.copy()) } } } else { - rdd.mapPartitionsInternal { iter => + newRdd.mapPartitionsInternal { iter => val getPartitionKey = getPartitionKeyExtractor() val mutablePair = new MutablePair[Int, InternalRow]() iter.map { row => mutablePair.update(part.getPartition(getPartitionKey(row)), row) } diff --git a/sql/core/src/test/java/test/org/apache/spark/sql/execution/sort/RecordBinaryComparatorSuite.java b/sql/core/src/test/java/test/org/apache/spark/sql/execution/sort/RecordBinaryComparatorSuite.java new file mode 100644 index 0000000000000..97f3dc588ecc5 --- /dev/null +++ b/sql/core/src/test/java/test/org/apache/spark/sql/execution/sort/RecordBinaryComparatorSuite.java @@ -0,0 +1,322 @@ +/* + * 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 test.org.apache.spark.sql.execution.sort; + +import org.apache.spark.SparkConf; +import org.apache.spark.memory.TaskMemoryManager; +import org.apache.spark.memory.TestMemoryConsumer; +import org.apache.spark.memory.TestMemoryManager; +import org.apache.spark.sql.catalyst.expressions.UnsafeArrayData; +import org.apache.spark.sql.catalyst.expressions.UnsafeRow; +import org.apache.spark.sql.execution.RecordBinaryComparator; +import org.apache.spark.unsafe.Platform; +import org.apache.spark.unsafe.UnsafeAlignedOffset; +import org.apache.spark.unsafe.array.LongArray; +import org.apache.spark.unsafe.memory.MemoryBlock; +import org.apache.spark.unsafe.types.UTF8String; +import org.apache.spark.util.collection.unsafe.sort.*; + +import org.junit.After; +import org.junit.Before; +import org.junit.Test; + +/** + * Test the RecordBinaryComparator, which compares two UnsafeRows by their binary form. + */ +public class RecordBinaryComparatorSuite { + + private final TaskMemoryManager memoryManager = new TaskMemoryManager( + new TestMemoryManager(new SparkConf().set("spark.memory.offHeap.enabled", "false")), 0); + private final TestMemoryConsumer consumer = new TestMemoryConsumer(memoryManager); + + private final int uaoSize = UnsafeAlignedOffset.getUaoSize(); + + private MemoryBlock dataPage; + private long pageCursor; + + private LongArray array; + private int pos; + + @Before + public void beforeEach() { + // Only compare between two input rows. + array = consumer.allocateArray(2); + pos = 0; + + dataPage = memoryManager.allocatePage(4096, consumer); + pageCursor = dataPage.getBaseOffset(); + } + + @After + public void afterEach() { + consumer.freePage(dataPage); + dataPage = null; + pageCursor = 0; + + consumer.freeArray(array); + array = null; + pos = 0; + } + + private void insertRow(UnsafeRow row) { + Object recordBase = row.getBaseObject(); + long recordOffset = row.getBaseOffset(); + int recordLength = row.getSizeInBytes(); + + Object baseObject = dataPage.getBaseObject(); + assert(pageCursor + recordLength <= dataPage.getBaseOffset() + dataPage.size()); + long recordAddress = memoryManager.encodePageNumberAndOffset(dataPage, pageCursor); + UnsafeAlignedOffset.putSize(baseObject, pageCursor, recordLength); + pageCursor += uaoSize; + Platform.copyMemory(recordBase, recordOffset, baseObject, pageCursor, recordLength); + pageCursor += recordLength; + + assert(pos < 2); + array.set(pos, recordAddress); + pos++; + } + + private int compare(int index1, int index2) { + Object baseObject = dataPage.getBaseObject(); + + long recordAddress1 = array.get(index1); + long baseOffset1 = memoryManager.getOffsetInPage(recordAddress1) + uaoSize; + int recordLength1 = UnsafeAlignedOffset.getSize(baseObject, baseOffset1 - uaoSize); + + long recordAddress2 = array.get(index2); + long baseOffset2 = memoryManager.getOffsetInPage(recordAddress2) + uaoSize; + int recordLength2 = UnsafeAlignedOffset.getSize(baseObject, baseOffset2 - uaoSize); + + return binaryComparator.compare(baseObject, baseOffset1, recordLength1, baseObject, + baseOffset2, recordLength2); + } + + private final RecordComparator binaryComparator = new RecordBinaryComparator(); + + // Compute the most compact size for UnsafeRow's backing data. + private int computeSizeInBytes(int originalSize) { + // All the UnsafeRows in this suite contains less than 64 columns, so the bitSetSize shall + // always be 8. + return 8 + (originalSize + 7) / 8 * 8; + } + + // Compute the relative offset of variable-length values. + private long relativeOffset(int numFields) { + // All the UnsafeRows in this suite contains less than 64 columns, so the bitSetSize shall + // always be 8. + return 8 + numFields * 8L; + } + + @Test + public void testBinaryComparatorForSingleColumnRow() throws Exception { + int numFields = 1; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + row1.pointTo(data1, computeSizeInBytes(numFields * 8)); + row1.setInt(0, 11); + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + row2.pointTo(data2, computeSizeInBytes(numFields * 8)); + row2.setInt(0, 42); + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 0) == 0); + assert(compare(0, 1) < 0); + } + + @Test + public void testBinaryComparatorForMultipleColumnRow() throws Exception { + int numFields = 5; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + row1.pointTo(data1, computeSizeInBytes(numFields * 8)); + for (int i = 0; i < numFields; i++) { + row1.setDouble(i, i * 3.14); + } + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + row2.pointTo(data2, computeSizeInBytes(numFields * 8)); + for (int i = 0; i < numFields; i++) { + row2.setDouble(i, 198.7 / (i + 1)); + } + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 0) == 0); + assert(compare(0, 1) < 0); + } + + @Test + public void testBinaryComparatorForArrayColumn() throws Exception { + int numFields = 1; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + UnsafeArrayData arrayData1 = UnsafeArrayData.fromPrimitiveArray(new int[]{11, 42, -1}); + row1.pointTo(data1, computeSizeInBytes(numFields * 8 + arrayData1.getSizeInBytes())); + row1.setLong(0, (relativeOffset(numFields) << 32) | (long) arrayData1.getSizeInBytes()); + Platform.copyMemory(arrayData1.getBaseObject(), arrayData1.getBaseOffset(), data1, + row1.getBaseOffset() + relativeOffset(numFields), arrayData1.getSizeInBytes()); + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + UnsafeArrayData arrayData2 = UnsafeArrayData.fromPrimitiveArray(new int[]{22}); + row2.pointTo(data2, computeSizeInBytes(numFields * 8 + arrayData2.getSizeInBytes())); + row2.setLong(0, (relativeOffset(numFields) << 32) | (long) arrayData2.getSizeInBytes()); + Platform.copyMemory(arrayData2.getBaseObject(), arrayData2.getBaseOffset(), data2, + row2.getBaseOffset() + relativeOffset(numFields), arrayData2.getSizeInBytes()); + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 0) == 0); + assert(compare(0, 1) > 0); + } + + @Test + public void testBinaryComparatorForMixedColumns() throws Exception { + int numFields = 4; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + UTF8String str1 = UTF8String.fromString("Milk tea"); + row1.pointTo(data1, computeSizeInBytes(numFields * 8 + str1.numBytes())); + row1.setInt(0, 11); + row1.setDouble(1, 3.14); + row1.setInt(2, -1); + row1.setLong(3, (relativeOffset(numFields) << 32) | (long) str1.numBytes()); + Platform.copyMemory(str1.getBaseObject(), str1.getBaseOffset(), data1, + row1.getBaseOffset() + relativeOffset(numFields), str1.numBytes()); + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + UTF8String str2 = UTF8String.fromString("Java"); + row2.pointTo(data2, computeSizeInBytes(numFields * 8 + str2.numBytes())); + row2.setInt(0, 11); + row2.setDouble(1, 3.14); + row2.setInt(2, -1); + row2.setLong(3, (relativeOffset(numFields) << 32) | (long) str2.numBytes()); + Platform.copyMemory(str2.getBaseObject(), str2.getBaseOffset(), data2, + row2.getBaseOffset() + relativeOffset(numFields), str2.numBytes()); + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 0) == 0); + assert(compare(0, 1) > 0); + } + + @Test + public void testBinaryComparatorForNullColumns() throws Exception { + int numFields = 3; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + row1.pointTo(data1, computeSizeInBytes(numFields * 8)); + for (int i = 0; i < numFields; i++) { + row1.setNullAt(i); + } + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + row2.pointTo(data2, computeSizeInBytes(numFields * 8)); + for (int i = 0; i < numFields - 1; i++) { + row2.setNullAt(i); + } + row2.setDouble(numFields - 1, 3.14); + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 0) == 0); + assert(compare(0, 1) > 0); + } + + @Test + public void testBinaryComparatorWhenSubtractionIsDivisibleByMaxIntValue() throws Exception { + int numFields = 1; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + row1.pointTo(data1, computeSizeInBytes(numFields * 8)); + row1.setLong(0, 11); + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + row2.pointTo(data2, computeSizeInBytes(numFields * 8)); + row2.setLong(0, 11L + Integer.MAX_VALUE); + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 1) < 0); + } + + @Test + public void testBinaryComparatorWhenSubtractionCanOverflowLongValue() throws Exception { + int numFields = 1; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + row1.pointTo(data1, computeSizeInBytes(numFields * 8)); + row1.setLong(0, Long.MIN_VALUE); + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + row2.pointTo(data2, computeSizeInBytes(numFields * 8)); + row2.setLong(0, 1); + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 1) < 0); + } + + @Test + public void testBinaryComparatorWhenOnlyTheLastColumnDiffers() throws Exception { + int numFields = 4; + + UnsafeRow row1 = new UnsafeRow(numFields); + byte[] data1 = new byte[100]; + row1.pointTo(data1, computeSizeInBytes(numFields * 8)); + row1.setInt(0, 11); + row1.setDouble(1, 3.14); + row1.setInt(2, -1); + row1.setLong(3, 0); + + UnsafeRow row2 = new UnsafeRow(numFields); + byte[] data2 = new byte[100]; + row2.pointTo(data2, computeSizeInBytes(numFields * 8)); + row2.setInt(0, 11); + row2.setDouble(1, 3.14); + row2.setInt(2, -1); + row2.setLong(3, 1); + + insertRow(row1); + insertRow(row2); + + assert(compare(0, 1) < 0); + } +} diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/ExchangeSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/ExchangeSuite.scala index 36cde3233dce8..2fdd8025fa6d5 100644 --- a/sql/core/src/test/scala/org/apache/spark/sql/execution/ExchangeSuite.scala +++ b/sql/core/src/test/scala/org/apache/spark/sql/execution/ExchangeSuite.scala @@ -17,11 +17,14 @@ package org.apache.spark.sql.execution -import org.apache.spark.sql.Row +import scala.util.Random + +import org.apache.spark.sql.{Dataset, Row} import org.apache.spark.sql.catalyst.expressions.{Alias, Literal} import org.apache.spark.sql.catalyst.plans.physical.{HashPartitioning, IdentityBroadcastMode, SinglePartition} import org.apache.spark.sql.execution.exchange.{BroadcastExchangeExec, ReusedExchangeExec, ShuffleExchange} import org.apache.spark.sql.execution.joins.HashedRelationBroadcastMode +import org.apache.spark.sql.internal.SQLConf import org.apache.spark.sql.test.SharedSQLContext class ExchangeSuite extends SparkPlanTest with SharedSQLContext { @@ -101,4 +104,25 @@ class ExchangeSuite extends SparkPlanTest with SharedSQLContext { assert(!exchange4.sameResult(exchange5)) assert(exchange5 sameResult exchange4) } + + test("SPARK-23207: Make repartition() generate consistent output") { + def assertConsistency(ds: Dataset[java.lang.Long]): Unit = { + ds.persist() + + val exchange = ds.mapPartitions { iter => + Random.shuffle(iter) + }.repartition(111) + val exchange2 = ds.repartition(111) + + assert(exchange.rdd.collectPartitions() === exchange2.rdd.collectPartitions()) + } + + withSQLConf(SQLConf.SORT_BEFORE_REPARTITION.key -> "true") { + // repartition() should generate consistent output. + assertConsistency(spark.range(10000)) + + // case when input contains duplicated rows. + assertConsistency(spark.range(10000).map(i => Random.nextInt(1000).toLong)) + } + } } diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetIOSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetIOSuite.scala index acdadb3103c8a..44139f66abea0 100644 --- a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetIOSuite.scala +++ b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetIOSuite.scala @@ -635,7 +635,7 @@ class ParquetIOSuite extends QueryTest with ParquetTest with SharedSQLContext { val v = (row.getInt(0), row.getString(1)) result += v } - assert(data == result) + assert(data.toSet == result.toSet) } finally { reader.close() } @@ -651,7 +651,7 @@ class ParquetIOSuite extends QueryTest with ParquetTest with SharedSQLContext { val row = reader.getCurrentValue.asInstanceOf[InternalRow] result += row.getString(0) } - assert(data.map(_._2) == result) + assert(data.map(_._2).toSet == result.toSet) } finally { reader.close() } @@ -668,7 +668,7 @@ class ParquetIOSuite extends QueryTest with ParquetTest with SharedSQLContext { val v = (row.getString(0), row.getInt(1)) result += v } - assert(data.map { x => (x._2, x._1) } == result) + assert(data.map { x => (x._2, x._1) }.toSet == result.toSet) } finally { reader.close() } diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/streaming/ForeachSinkSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/streaming/ForeachSinkSuite.scala index 9137d650e906b..41434e6d8b974 100644 --- a/sql/core/src/test/scala/org/apache/spark/sql/execution/streaming/ForeachSinkSuite.scala +++ b/sql/core/src/test/scala/org/apache/spark/sql/execution/streaming/ForeachSinkSuite.scala @@ -52,13 +52,13 @@ class ForeachSinkSuite extends StreamTest with SharedSQLContext with BeforeAndAf var expectedEventsForPartition0 = Seq( ForeachSinkSuite.Open(partition = 0, version = 0), - ForeachSinkSuite.Process(value = 1), + ForeachSinkSuite.Process(value = 2), ForeachSinkSuite.Process(value = 3), ForeachSinkSuite.Close(None) ) var expectedEventsForPartition1 = Seq( ForeachSinkSuite.Open(partition = 1, version = 0), - ForeachSinkSuite.Process(value = 2), + ForeachSinkSuite.Process(value = 1), ForeachSinkSuite.Process(value = 4), ForeachSinkSuite.Close(None) ) From d26901516f414d16f8eb38183053d22dc3b056e3 Mon Sep 17 00:00:00 2001 From: Henry Robinson Date: Thu, 23 Aug 2018 16:51:55 -0700 Subject: [PATCH 2/2] Remove dependency on java 8's Supplier entirely --- .../spark/sql/execution/UnsafeExternalRowSorter.java | 11 +++++++---- .../sql/execution/exchange/ShuffleExchange.scala | 3 +-- 2 files changed, 8 insertions(+), 6 deletions(-) diff --git a/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java b/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java index 51baa34d894bc..402f7a7845d2a 100644 --- a/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java +++ b/sql/catalyst/src/main/java/org/apache/spark/sql/execution/UnsafeExternalRowSorter.java @@ -18,7 +18,6 @@ package org.apache.spark.sql.execution; import java.io.IOException; -import java.util.function.Supplier; import scala.collection.Iterator; import scala.math.Ordering; @@ -52,6 +51,10 @@ public final class UnsafeExternalRowSorter { private final PrefixComputer prefixComputer; private final UnsafeExternalSorter sorter; + public static interface RecordComparatorSupplier { + public RecordComparator get(); + } + public abstract static class PrefixComputer { public static class Prefix { @@ -71,7 +74,7 @@ public static class Prefix { public static UnsafeExternalRowSorter createWithRecordComparator( StructType schema, - Supplier recordComparatorSupplier, + RecordComparatorSupplier recordComparatorSupplier, PrefixComparator prefixComparator, PrefixComputer prefixComputer, long pageSizeBytes, @@ -87,7 +90,7 @@ public static UnsafeExternalRowSorter create( PrefixComputer prefixComputer, long pageSizeBytes, boolean canUseRadixSort) throws IOException { - Supplier recordComparatorSupplier = new Supplier() { + RecordComparatorSupplier recordComparatorSupplier = new RecordComparatorSupplier() { public RecordComparator get() { return new RowComparator(ordering, schema.length()); } }; return new UnsafeExternalRowSorter(schema, recordComparatorSupplier, prefixComparator, @@ -96,7 +99,7 @@ public static UnsafeExternalRowSorter create( private UnsafeExternalRowSorter( StructType schema, - Supplier recordComparatorSupplier, + RecordComparatorSupplier recordComparatorSupplier, PrefixComparator prefixComparator, PrefixComputer prefixComputer, long pageSizeBytes, diff --git a/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala b/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala index ac069087b0c49..5d3fdf2bddc10 100644 --- a/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala +++ b/sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchange.scala @@ -18,7 +18,6 @@ package org.apache.spark.sql.execution.exchange import java.util.Random -import java.util.function.Supplier import org.apache.spark._ import org.apache.spark.rdd.RDD @@ -257,7 +256,7 @@ object ShuffleExchange { newPartitioning.numPartitions > 1 && newPartitioning.isInstanceOf[RoundRobinPartitioning]) { rdd.mapPartitionsInternal { iter => - val recordComparatorSupplier = new Supplier[RecordComparator] { + val recordComparatorSupplier = new UnsafeExternalRowSorter.RecordComparatorSupplier { override def get: RecordComparator = new RecordBinaryComparator() } // The comparator for comparing row hashcode, which should always be Integer.