ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

Address Patrick's and Reynold's comments

Browse source 
Aside from trivial formatting changes, use nulls instead of Options for
DiskMapIterator, and add documentation for spark.shuffle.externalSorting
and spark.shuffle.memoryFraction.

Also, set spark.shuffle.memoryFraction to 0.3, and spark.storage.memoryFraction = 0.6.
This commit is contained in:
Andrew Or 2014-01-10 15:09:51 -08:00
parent 372a533a6c
commit e4c51d2113
5 changed files with 73 additions and 47 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
core/src/main/scala/org/apache/spark/Aggregator.scala
View file

@ -32,7 +32,7 @@ case class Aggregator[K, V, C] (
mergeCombiners: (C, C) => C) {
private val sparkConf = SparkEnv.get.conf
private val externalSorting = sparkConf.get("spark.shuffle.externalSorting", "false").toBoolean
private val externalSorting = sparkConf.getBoolean("spark.shuffle.externalSorting", true)
def combineValuesByKey(iter: Iterator[_ <: Product2[K, V]]) : Iterator[(K, C)] = {
if (!externalSorting) {

3
core/src/main/scala/org/apache/spark/rdd/CoGroupedRDD.scala
View file

@ -106,8 +106,7 @@ class CoGroupedRDD[K](@transient var rdds: Seq[RDD[_ <: Product2[K, _]]], part:
override val partitioner = Some(part)
override def compute(s: Partition, context: TaskContext): Iterator[(K, CoGroupCombiner)] = {
val externalSorting = sparkConf.get("spark.shuffle.externalSorting", "false").toBoolean
val externalSorting = sparkConf.getBoolean("spark.shuffle.externalSorting", true)
val split = s.asInstanceOf[CoGroupPartition]
val numRdds = split.deps.size

2
core/src/main/scala/org/apache/spark/storage/BlockManager.scala
View file

@ -864,7 +864,7 @@ private[spark] object BlockManager extends Logging {
val ID_GENERATOR = new IdGenerator
def getMaxMemory(conf: SparkConf): Long = {
val memoryFraction = conf.getDouble("spark.storage.memoryFraction", 0.66)
val memoryFraction = conf.getDouble("spark.storage.memoryFraction", 0.6)
(Runtime.getRuntime.maxMemory * memoryFraction).toLong
}

85
core/src/main/scala/org/apache/spark/util/collection/ExternalAppendOnlyMap.scala
View file

@ -71,21 +71,24 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
// Collective memory threshold shared across all running tasks
private val maxMemoryThreshold = {
val memoryFraction = sparkConf.getDouble("spark.shuffle.memoryFraction", 0.75)
val memoryFraction = sparkConf.getDouble("spark.shuffle.memoryFraction", 0.3)
val safetyFraction = sparkConf.getDouble("spark.shuffle.safetyFraction", 0.8)
(Runtime.getRuntime.maxMemory * memoryFraction * safetyFraction).toLong
}
// How many inserts into this map before tracking its shuffle memory usage
private val initialInsertThreshold =
sparkConf.getLong("spark.shuffle.initialInsertThreshold", 1000)
// Number of pairs in the in-memory map
private var numPairsInMemory = 0
// Number of in-memory pairs inserted before tracking the map's shuffle memory usage
private val trackMemoryThreshold = 1000
// How many times we have spilled so far
private var spillCount = 0
private val fileBufferSize = sparkConf.getInt("spark.shuffle.file.buffer.kb", 100) * 1024
private val syncWrites = sparkConf.get("spark.shuffle.sync", "false").toBoolean
private val syncWrites = sparkConf.getBoolean("spark.shuffle.sync", false)
private val comparator = new KCComparator[K, C]
private val ser = serializer.newInstance()
private var insertCount = 0
private var spillCount = 0
/**
* Insert the given key and value into the map.
@ -94,14 +97,13 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
* enough room for this to happen. If so, allocate the memory required to grow the map;
* otherwise, spill the in-memory map to disk.
*
* The shuffle memory usage of the first initialInsertThreshold entries is not tracked.
* The shuffle memory usage of the first trackMemoryThreshold entries is not tracked.
*/
def insert(key: K, value: V) {
insertCount += 1
val update: (Boolean, C) => C = (hadVal, oldVal) => {
if (hadVal) mergeValue(oldVal, value) else createCombiner(value)
}
if (insertCount > initialInsertThreshold && currentMap.atGrowThreshold) {
if (numPairsInMemory > trackMemoryThreshold && currentMap.atGrowThreshold) {
val mapSize = currentMap.estimateSize()
var shouldSpill = false
val shuffleMemoryMap = SparkEnv.get.shuffleMemoryMap
@ -114,7 +116,7 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
val availableMemory = maxMemoryThreshold -
(shuffleMemoryMap.values.sum - previouslyOccupiedMemory.getOrElse(0L))
// Assume map grow factor is 2x
// Assume map growth factor is 2x
shouldSpill = availableMemory < mapSize * 2
if (!shouldSpill) {
shuffleMemoryMap(threadId) = mapSize * 2
@ -126,6 +128,7 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
}
}
currentMap.changeValue(key, update)
numPairsInMemory += 1
}
/**
@ -133,7 +136,7 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
*/
private def spill(mapSize: Long) {
spillCount += 1
logWarning("* Spilling in-memory map of %d MB to disk (%d time%s so far)"
logWarning("Spilling in-memory map of %d MB to disk (%d time%s so far)"
.format(mapSize / (1024 * 1024), spillCount, if (spillCount > 1) "s" else ""))
val (blockId, file) = diskBlockManager.createTempBlock()
val writer =
@ -157,9 +160,13 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
shuffleMemoryMap.synchronized {
shuffleMemoryMap(Thread.currentThread().getId) = 0
}
insertCount = 0
numPairsInMemory = 0
}
/**
* Return an iterator that merges the in-memory map with the spilled maps.
* If no spill has occurred, simply return the in-memory map's iterator.
*/
override def iterator: Iterator[(K, C)] = {
if (spilledMaps.isEmpty) {
currentMap.iterator
@ -168,7 +175,9 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
}
}
/** An iterator that sort-merges (K, C) pairs from the in-memory and on-disk maps */
/**
* An iterator that sort-merges (K, C) pairs from the in-memory map and the spilled maps
*/
private class ExternalIterator extends Iterator[(K, C)] {
// A fixed-size queue that maintains a buffer for each stream we are currently merging
@ -187,6 +196,7 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
/**
* Fetch from the given iterator until a key of different hash is retrieved. In the
* event of key hash collisions, this ensures no pairs are hidden from being merged.
* Assume the given iterator is in sorted order.
*/
def getMorePairs(it: Iterator[(K, C)]): ArrayBuffer[(K, C)] = {
val kcPairs = new ArrayBuffer[(K, C)]
@ -219,17 +229,16 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
baseCombiner
}
override def hasNext: Boolean = {
mergeHeap.foreach{ buffer =>
if (!buffer.pairs.isEmpty) {
return true
}
}
false
}
/**
* Return true if there exists an input stream that still has unvisited pairs
*/
override def hasNext: Boolean = mergeHeap.exists(!_.pairs.isEmpty)
/**
* Select a key with the minimum hash, then combine all values with the same key from all input streams.
*/
override def next(): (K, C) = {
// Select a return key from the StreamBuffer that holds the lowest key hash
// Select a key from the StreamBuffer that holds the lowest key hash
val minBuffer = mergeHeap.dequeue()
val (minPairs, minHash) = (minBuffer.pairs, minBuffer.minKeyHash)
if (minPairs.length == 0) {
@ -285,45 +294,43 @@ private[spark] class ExternalAppendOnlyMap[K, V, C](
}
}
// Iterate through (K, C) pairs in sorted order from an on-disk map
/**
* An iterator that returns (K, C) pairs in sorted order from an on-disk map
*/
private class DiskMapIterator(file: File) extends Iterator[(K, C)] {
val fileStream = new FileInputStream(file)
val bufferedStream = new FastBufferedInputStream(fileStream)
val deserializeStream = ser.deserializeStream(bufferedStream)
var nextItem: Option[(K, C)] = None
var nextItem: (K, C) = null
var eof = false
def readNextItem(): Option[(K, C)] = {
def readNextItem(): (K, C) = {
if (!eof) {
try {
return Some(deserializeStream.readObject().asInstanceOf[(K, C)])
return deserializeStream.readObject().asInstanceOf[(K, C)]
} catch {
case e: EOFException =>
eof = true
cleanup()
}
}
None
null
}
override def hasNext: Boolean = {
nextItem match {
case Some(item) => true
case None =>
if (nextItem == null) {
nextItem = readNextItem()
nextItem.isDefined
}
nextItem != null
}
override def next(): (K, C) = {
nextItem match {
case Some(item) =>
nextItem = None
item
case None =>
val item = readNextItem()
item.getOrElse(throw new NoSuchElementException)
val item = if (nextItem == null) readNextItem() else nextItem
if (item == null) {
throw new NoSuchElementException
}
nextItem = null
item
}
// TODO: Ensure this gets called even if the iterator isn't drained.

24
docs/configuration.md
View file

@ -104,13 +104,24 @@ Apart from these, the following properties are also available, and may be useful
</tr>
<tr>
<td>spark.storage.memoryFraction</td>
<td>0.66</td>
<td>0.6</td>
<td>
Fraction of Java heap to use for Spark's memory cache. This should not be larger than the "old"
generation of objects in the JVM, which by default is given 2/3 of the heap, but you can increase
generation of objects in the JVM, which by default is given 0.6 of the heap, but you can increase
it if you configure your own old generation size.
</td>
</tr>
<tr>
<td>spark.shuffle.memoryFraction</td>
<td>0.3</td>
<td>
Fraction of Java heap to use for aggregation and cogroups during shuffles, if
<code>spark.shuffle.externalSorting</code> is enabled. At any given time, the collective size of
all in-memory maps used for shuffles is bounded by this limit, beyond which the contents will
begin to spill to disk. If spills are often, consider increasing this value at the expense of
<code>spark.storage.memoryFraction</code>.
</td>
</tr>
<tr>
<td>spark.mesos.coarse</td>
<td>false</td>
@ -376,6 +387,15 @@ Apart from these, the following properties are also available, and may be useful
If set to "true", consolidates intermediate files created during a shuffle. Creating fewer files can improve filesystem performance for shuffles with large numbers of reduce tasks. It is recommended to set this to "true" when using ext4 or xfs filesystems. On ext3, this option might degrade performance on machines with many (>8) cores due to filesystem limitations.
</td>
</tr>
<tr>
<td>spark.shuffle.externalSorting</td>
<td>true</td>
<td>
If set to "true", spills in-memory maps used for shuffles to disk when a memory threshold is reached. This
threshold is specified by <code>spark.shuffle.memoryFraction</code>. Enable this especially for memory-intensive
applications.
</td>
</tr>
<tr>
<td>spark.speculation</td>
<td>false</td>