@Override
public void merge(List<InMemoryMapOutput<K, V>> inputs) throws IOException {
if (inputs == null || inputs.size() == 0) {
return;
}
TaskAttemptID dummyMapId = inputs.get(0).getMapId();
List<Segment<K, V>> inMemorySegments = new ArrayList<Segment<K, V>>();
long mergeOutputSize = createInMemorySegments(inputs, inMemorySegments, 0);
int noInMemorySegments = inMemorySegments.size();
InMemoryMapOutput<K, V> mergedMapOutputs =
unconditionalReserve(dummyMapId, mergeOutputSize, false);
Writer<K, V> writer = new InMemoryWriter<K, V>(mergedMapOutputs.getArrayStream());
LOG.info(
"Initiating Memory-to-Memory merge with "
+ noInMemorySegments
+ " segments of total-size: "
+ mergeOutputSize);
RawKeyValueIterator rIter =
Merger.merge(
jobConf,
rfs,
(Class<K>) jobConf.getMapOutputKeyClass(),
(Class<V>) jobConf.getMapOutputValueClass(),
inMemorySegments,
inMemorySegments.size(),
new Path(reduceId.toString()),
(RawComparator<K>) jobConf.getOutputKeyComparator(),
reporter,
null,
null,
null);
Merger.writeFile(rIter, writer, reporter, jobConf);
writer.close();
LOG.info(
reduceId
+ " Memory-to-Memory merge of the "
+ noInMemorySegments
+ " files in-memory complete.");
// Note the output of the merge
closeInMemoryMergedFile(mergedMapOutputs);
}
private void mergeParts() throws IOException {
// get the approximate size of the final output/index files
long finalOutFileSize = 0;
long finalIndexFileSize = 0;
Path[] filename = new Path[numSpills];
Path[] indexFileName = new Path[numSpills];
FileSystem localFs = FileSystem.getLocal(job);
for (int i = 0; i < numSpills; i++) {
filename[i] = mapOutputFile.getSpillFile(getTaskID(), i);
indexFileName[i] = mapOutputFile.getSpillIndexFile(getTaskID(), i);
finalOutFileSize += localFs.getFileStatus(filename[i]).getLen();
}
if (numSpills == 1) { // the spill is the final output
localFs.rename(filename[0], new Path(filename[0].getParent(), "file.out"));
localFs.rename(indexFileName[0], new Path(indexFileName[0].getParent(), "file.out.index"));
return;
}
// make correction in the length to include the sequence file header
// lengths for each partition
finalOutFileSize += partitions * APPROX_HEADER_LENGTH;
finalIndexFileSize = partitions * MAP_OUTPUT_INDEX_RECORD_LENGTH;
Path finalOutputFile = mapOutputFile.getOutputFileForWrite(getTaskID(), finalOutFileSize);
Path finalIndexFile =
mapOutputFile.getOutputIndexFileForWrite(getTaskID(), finalIndexFileSize);
// The output stream for the final single output file
FSDataOutputStream finalOut = localFs.create(finalOutputFile, true, 4096);
// The final index file output stream
FSDataOutputStream finalIndexOut = localFs.create(finalIndexFile, true, 4096);
if (numSpills == 0) {
// create dummy files
for (int i = 0; i < partitions; i++) {
long segmentStart = finalOut.getPos();
Writer<K, V> writer = new Writer<K, V>(job, finalOut, keyClass, valClass, codec);
writer.close();
writeIndexRecord(finalIndexOut, finalOut, segmentStart, writer);
}
finalOut.close();
finalIndexOut.close();
return;
}
{
for (int parts = 0; parts < partitions; parts++) {
// create the segments to be merged
List<Segment<K, V>> segmentList = new ArrayList<Segment<K, V>>(numSpills);
for (int i = 0; i < numSpills; i++) {
FSDataInputStream indexIn = localFs.open(indexFileName[i]);
indexIn.seek(parts * MAP_OUTPUT_INDEX_RECORD_LENGTH);
long segmentOffset = indexIn.readLong();
long rawSegmentLength = indexIn.readLong();
long segmentLength = indexIn.readLong();
indexIn.close();
Segment<K, V> s =
new Segment<K, V>(
job, localFs, filename[i], segmentOffset, segmentLength, codec, true);
segmentList.add(i, s);
if (LOG.isDebugEnabled()) {
LOG.debug(
"Index: ("
+ indexFileName[i]
+ ", "
+ segmentOffset
+ rawSegmentLength
+ ", "
+ segmentLength
+ ")");
}
}
// merge
@SuppressWarnings("unchecked")
RawKeyValueIterator kvIter =
Merger.merge(
job,
localFs,
keyClass,
valClass,
segmentList,
job.getInt("io.sort.factor", 100),
new Path(getTaskID().toString()),
job.getOutputKeyComparator(),
reporter);
// write merged output to disk
long segmentStart = finalOut.getPos();
Writer<K, V> writer = new Writer<K, V>(job, finalOut, keyClass, valClass, codec);
if (null == combinerClass
|| job.getCombineOnceOnly()
|| numSpills < minSpillsForCombine) {
Merger.writeFile(kvIter, writer, reporter, job);
} else {
combineCollector.setWriter(writer);
combineAndSpill(kvIter, combineInputCounter);
}
// close
writer.close();
// write index record
writeIndexRecord(finalIndexOut, finalOut, segmentStart, writer);
}
finalOut.close();
finalIndexOut.close();
// cleanup
for (int i = 0; i < numSpills; i++) {
localFs.delete(filename[i], true);
localFs.delete(indexFileName[i], true);
}
}
}
private RawKeyValueIterator finalMerge(
JobConf job,
FileSystem fs,
List<InMemoryMapOutput<K, V>> inMemoryMapOutputs,
List<CompressAwarePath> onDiskMapOutputs)
throws IOException {
LOG.info(
"finalMerge called with "
+ inMemoryMapOutputs.size()
+ " in-memory map-outputs and "
+ onDiskMapOutputs.size()
+ " on-disk map-outputs");
final float maxRedPer = job.getFloat(MRJobConfig.REDUCE_INPUT_BUFFER_PERCENT, 0f);
if (maxRedPer > 1.0 || maxRedPer < 0.0) {
throw new IOException(MRJobConfig.REDUCE_INPUT_BUFFER_PERCENT + maxRedPer);
}
int maxInMemReduce =
(int) Math.min(Runtime.getRuntime().maxMemory() * maxRedPer, Integer.MAX_VALUE);
// merge config params
Class<K> keyClass = (Class<K>) job.getMapOutputKeyClass();
Class<V> valueClass = (Class<V>) job.getMapOutputValueClass();
boolean keepInputs = job.getKeepFailedTaskFiles();
final Path tmpDir = new Path(reduceId.toString());
final RawComparator<K> comparator = (RawComparator<K>) job.getOutputKeyComparator();
// segments required to vacate memory
List<Segment<K, V>> memDiskSegments = new ArrayList<Segment<K, V>>();
long inMemToDiskBytes = 0;
boolean mergePhaseFinished = false;
if (inMemoryMapOutputs.size() > 0) {
TaskID mapId = inMemoryMapOutputs.get(0).getMapId().getTaskID();
inMemToDiskBytes =
createInMemorySegments(inMemoryMapOutputs, memDiskSegments, maxInMemReduce);
final int numMemDiskSegments = memDiskSegments.size();
if (numMemDiskSegments > 0 && ioSortFactor > onDiskMapOutputs.size()) {
// If we reach here, it implies that we have less than io.sort.factor
// disk segments and this will be incremented by 1 (result of the
// memory segments merge). Since this total would still be
// <= io.sort.factor, we will not do any more intermediate merges,
// the merge of all these disk segments would be directly fed to the
// reduce method
mergePhaseFinished = true;
// must spill to disk, but can't retain in-mem for intermediate merge
final Path outputPath =
mapOutputFile
.getInputFileForWrite(mapId, inMemToDiskBytes)
.suffix(Task.MERGED_OUTPUT_PREFIX);
final RawKeyValueIterator rIter =
Merger.merge(
job,
fs,
keyClass,
valueClass,
memDiskSegments,
numMemDiskSegments,
tmpDir,
comparator,
reporter,
spilledRecordsCounter,
null,
mergePhase);
Writer<K, V> writer =
new Writer<K, V>(job, fs, outputPath, keyClass, valueClass, codec, null);
try {
Merger.writeFile(rIter, writer, reporter, job);
writer.close();
onDiskMapOutputs.add(
new CompressAwarePath(
outputPath, writer.getRawLength(), writer.getCompressedLength()));
writer = null;
// add to list of final disk outputs.
} catch (IOException e) {
if (null != outputPath) {
try {
fs.delete(outputPath, true);
} catch (IOException ie) {
// NOTHING
}
}
throw e;
} finally {
if (null != writer) {
writer.close();
}
}
LOG.info(
"Merged "
+ numMemDiskSegments
+ " segments, "
+ inMemToDiskBytes
+ " bytes to disk to satisfy "
+ "reduce memory limit");
inMemToDiskBytes = 0;
memDiskSegments.clear();
} else if (inMemToDiskBytes != 0) {
LOG.info(
"Keeping "
+ numMemDiskSegments
+ " segments, "
+ inMemToDiskBytes
+ " bytes in memory for "
+ "intermediate, on-disk merge");
}
}
// segments on disk
List<Segment<K, V>> diskSegments = new ArrayList<Segment<K, V>>();
long onDiskBytes = inMemToDiskBytes;
long rawBytes = inMemToDiskBytes;
CompressAwarePath[] onDisk =
onDiskMapOutputs.toArray(new CompressAwarePath[onDiskMapOutputs.size()]);
for (CompressAwarePath file : onDisk) {
long fileLength = fs.getFileStatus(file).getLen();
onDiskBytes += fileLength;
rawBytes += (file.getRawDataLength() > 0) ? file.getRawDataLength() : fileLength;
LOG.debug("Disk file: " + file + " Length is " + fileLength);
diskSegments.add(
new Segment<K, V>(
job,
fs,
file,
codec,
keepInputs,
(file.toString().endsWith(Task.MERGED_OUTPUT_PREFIX)
? null
: mergedMapOutputsCounter),
file.getRawDataLength()));
}
LOG.info("Merging " + onDisk.length + " files, " + onDiskBytes + " bytes from disk");
Collections.sort(
diskSegments,
new Comparator<Segment<K, V>>() {
public int compare(Segment<K, V> o1, Segment<K, V> o2) {
if (o1.getLength() == o2.getLength()) {
return 0;
}
return o1.getLength() < o2.getLength() ? -1 : 1;
}
});
// build final list of segments from merged backed by disk + in-mem
List<Segment<K, V>> finalSegments = new ArrayList<Segment<K, V>>();
long inMemBytes = createInMemorySegments(inMemoryMapOutputs, finalSegments, 0);
LOG.info(
"Merging "
+ finalSegments.size()
+ " segments, "
+ inMemBytes
+ " bytes from memory into reduce");
if (0 != onDiskBytes) {
final int numInMemSegments = memDiskSegments.size();
diskSegments.addAll(0, memDiskSegments);
memDiskSegments.clear();
// Pass mergePhase only if there is a going to be intermediate
// merges. See comment where mergePhaseFinished is being set
Progress thisPhase = (mergePhaseFinished) ? null : mergePhase;
RawKeyValueIterator diskMerge =
Merger.merge(
job,
fs,
keyClass,
valueClass,
codec,
diskSegments,
ioSortFactor,
numInMemSegments,
tmpDir,
comparator,
reporter,
false,
spilledRecordsCounter,
null,
thisPhase);
diskSegments.clear();
if (0 == finalSegments.size()) {
return diskMerge;
}
finalSegments.add(
new Segment<K, V>(new RawKVIteratorReader(diskMerge, onDiskBytes), true, rawBytes));
}
return Merger.merge(
job,
fs,
keyClass,
valueClass,
finalSegments,
finalSegments.size(),
tmpDir,
comparator,
reporter,
spilledRecordsCounter,
null,
null);
}
@Override
public void merge(List<CompressAwarePath> inputs) throws IOException {
// sanity check
if (inputs == null || inputs.isEmpty()) {
LOG.info("No ondisk files to merge...");
return;
}
long approxOutputSize = 0;
int bytesPerSum = jobConf.getInt("io.bytes.per.checksum", 512);
LOG.info(
"OnDiskMerger: We have " + inputs.size() + " map outputs on disk. Triggering merge...");
// 1. Prepare the list of files to be merged.
for (CompressAwarePath file : inputs) {
approxOutputSize += localFS.getFileStatus(file).getLen();
}
// add the checksum length
approxOutputSize += ChecksumFileSystem.getChecksumLength(approxOutputSize, bytesPerSum);
// 2. Start the on-disk merge process
Path outputPath =
localDirAllocator
.getLocalPathForWrite(inputs.get(0).toString(), approxOutputSize, jobConf)
.suffix(Task.MERGED_OUTPUT_PREFIX);
Writer<K, V> writer =
new Writer<K, V>(
jobConf,
rfs,
outputPath,
(Class<K>) jobConf.getMapOutputKeyClass(),
(Class<V>) jobConf.getMapOutputValueClass(),
codec,
null);
RawKeyValueIterator iter = null;
CompressAwarePath compressAwarePath;
Path tmpDir = new Path(reduceId.toString());
try {
iter =
Merger.merge(
jobConf,
rfs,
(Class<K>) jobConf.getMapOutputKeyClass(),
(Class<V>) jobConf.getMapOutputValueClass(),
codec,
inputs.toArray(new Path[inputs.size()]),
true,
ioSortFactor,
tmpDir,
(RawComparator<K>) jobConf.getOutputKeyComparator(),
reporter,
spilledRecordsCounter,
null,
mergedMapOutputsCounter,
null);
Merger.writeFile(iter, writer, reporter, jobConf);
writer.close();
compressAwarePath =
new CompressAwarePath(outputPath, writer.getRawLength(), writer.getCompressedLength());
} catch (IOException e) {
localFS.delete(outputPath, true);
throw e;
}
closeOnDiskFile(compressAwarePath);
LOG.info(
reduceId
+ " Finished merging "
+ inputs.size()
+ " map output files on disk of total-size "
+ approxOutputSize
+ "."
+ " Local output file is "
+ outputPath
+ " of size "
+ localFS.getFileStatus(outputPath).getLen());
}
@Override
public void merge(List<InMemoryMapOutput<K, V>> inputs) throws IOException {
if (inputs == null || inputs.size() == 0) {
return;
}
// name this output file same as the name of the first file that is
// there in the current list of inmem files (this is guaranteed to
// be absent on the disk currently. So we don't overwrite a prev.
// created spill). Also we need to create the output file now since
// it is not guaranteed that this file will be present after merge
// is called (we delete empty files as soon as we see them
// in the merge method)
// figure out the mapId
TaskAttemptID mapId = inputs.get(0).getMapId();
TaskID mapTaskId = mapId.getTaskID();
List<Segment<K, V>> inMemorySegments = new ArrayList<Segment<K, V>>();
long mergeOutputSize = createInMemorySegments(inputs, inMemorySegments, 0);
int noInMemorySegments = inMemorySegments.size();
Path outputPath =
mapOutputFile
.getInputFileForWrite(mapTaskId, mergeOutputSize)
.suffix(Task.MERGED_OUTPUT_PREFIX);
Writer<K, V> writer =
new Writer<K, V>(
jobConf,
rfs,
outputPath,
(Class<K>) jobConf.getMapOutputKeyClass(),
(Class<V>) jobConf.getMapOutputValueClass(),
codec,
null);
RawKeyValueIterator rIter = null;
CompressAwarePath compressAwarePath;
try {
LOG.info("Initiating in-memory merge with " + noInMemorySegments + " segments...");
rIter =
Merger.merge(
jobConf,
rfs,
(Class<K>) jobConf.getMapOutputKeyClass(),
(Class<V>) jobConf.getMapOutputValueClass(),
inMemorySegments,
inMemorySegments.size(),
new Path(reduceId.toString()),
(RawComparator<K>) jobConf.getOutputKeyComparator(),
reporter,
spilledRecordsCounter,
null,
null);
if (null == combinerClass) {
Merger.writeFile(rIter, writer, reporter, jobConf);
} else {
combineCollector.setWriter(writer);
combineAndSpill(rIter, reduceCombineInputCounter);
}
writer.close();
compressAwarePath =
new CompressAwarePath(outputPath, writer.getRawLength(), writer.getCompressedLength());
LOG.info(
reduceId
+ " Merge of the "
+ noInMemorySegments
+ " files in-memory complete."
+ " Local file is "
+ outputPath
+ " of size "
+ localFS.getFileStatus(outputPath).getLen());
} catch (IOException e) {
// make sure that we delete the ondisk file that we created
// earlier when we invoked cloneFileAttributes
localFS.delete(outputPath, true);
throw e;
}
// Note the output of the merge
closeOnDiskFile(compressAwarePath);
}
RawKeyValueIterator merge(
Class<K> keyClass, Class<V> valueClass, int factor, int inMem, Path tmpDir)
throws IOException {
LOG.info("Merging " + segments.size() + " sorted segments");
// create the MergeStreams from the sorted map created in the constructor
// and dump the final output to a file
int numSegments = segments.size();
int origFactor = factor;
int passNo = 1;
do {
// get the factor for this pass of merge. We assume in-memory segments
// are the first entries in the segment list and that the pass factor
// doesn't apply to them
factor = getPassFactor(factor, passNo, numSegments - inMem);
if (1 == passNo) {
factor += inMem;
}
List<Segment<K, V>> segmentsToMerge = new ArrayList<Segment<K, V>>();
int segmentsConsidered = 0;
int numSegmentsToConsider = factor;
while (true) {
// extract the smallest 'factor' number of segments
// Call cleanup on the empty segments (no key/value data)
List<Segment<K, V>> mStream = getSegmentDescriptors(numSegmentsToConsider);
for (Segment<K, V> segment : mStream) {
// Initialize the segment at the last possible moment;
// this helps in ensuring we don't use buffers until we need them
segment.init();
long startPos = segment.getPosition();
boolean hasNext = segment.next();
long endPos = segment.getPosition();
totalBytesProcessed += endPos - startPos;
mergeProgress.set(totalBytesProcessed * progPerByte);
if (hasNext) {
segmentsToMerge.add(segment);
segmentsConsidered++;
} else {
segment.close();
numSegments--; // we ignore this segment for the merge
}
}
// if we have the desired number of segments
// or looked at all available segments, we break
if (segmentsConsidered == factor || segments.size() == 0) {
break;
}
numSegmentsToConsider = factor - segmentsConsidered;
}
// feed the streams to the priority queue
initialize(segmentsToMerge.size());
clear();
for (Segment<K, V> segment : segmentsToMerge) {
put(segment);
}
// if we have lesser number of segments remaining, then just return the
// iterator, else do another single level merge
if (numSegments <= factor) {
// calculate the length of the remaining segments. Required for
// calculating the merge progress
long totalBytes = 0;
for (int i = 0; i < segmentsToMerge.size(); i++) {
totalBytes += segmentsToMerge.get(i).getLength();
}
if (totalBytes != 0) // being paranoid
progPerByte = 1.0f / (float) totalBytes;
if (totalBytes != 0) mergeProgress.set(totalBytesProcessed * progPerByte);
else mergeProgress.set(1.0f); // Last pass and no segments left - we're done
LOG.info(
"Down to the last merge-pass, with "
+ numSegments
+ " segments left of total size: "
+ totalBytes
+ " bytes");
return this;
} else {
LOG.info(
"Merging "
+ segmentsToMerge.size()
+ " intermediate segments out of a total of "
+ (segments.size() + segmentsToMerge.size()));
// we want to spread the creation of temp files on multiple disks if
// available under the space constraints
long approxOutputSize = 0;
for (Segment<K, V> s : segmentsToMerge) {
approxOutputSize +=
s.getLength() + ChecksumFileSystem.getApproxChkSumLength(s.getLength());
}
Path tmpFilename = new Path(tmpDir, "intermediate").suffix("." + passNo);
Path outputFile =
lDirAlloc.getLocalPathForWrite(tmpFilename.toString(), approxOutputSize, conf);
Writer<K, V> writer = new Writer<K, V>(conf, fs, outputFile, keyClass, valueClass, codec);
writeFile(this, writer, reporter);
writer.close();
// we finished one single level merge; now clean up the priority
// queue
this.close();
// Add the newly create segment to the list of segments to be merged
Segment<K, V> tempSegment = new Segment<K, V>(conf, fs, outputFile, codec, false);
segments.add(tempSegment);
numSegments = segments.size();
Collections.sort(segments, segmentComparator);
passNo++;
}
// we are worried about only the first pass merge factor. So reset the
// factor to what it originally was
factor = origFactor;
} while (true);
}