/**
* @return the total key count in the files being merged
* @throws IOException
*/
private long prepareForMerge() throws IOException {
LOG.info("Merging " + inputFileNames);
LOG.info("Using block size: " + blockSize);
inputStoreFiles = new ArrayList<StoreFile>();
long maxKeyCount = 0;
for (String fileName : inputFileNames) {
Path filePath = new Path(fileName);
// Open without caching.
StoreFile sf = openStoreFile(filePath, false);
sf.createReader();
inputStoreFiles.add(sf);
StoreFile.Reader r = sf.getReader();
if (r != null) {
long keyCount = r.getFilterEntries();
maxKeyCount += keyCount;
LOG.info(
"Compacting: "
+ sf
+ "; keyCount = "
+ keyCount
+ "; Bloom Type = "
+ r.getBloomFilterType().toString()
+ "; Size = "
+ StringUtils.humanReadableInt(r.length()));
}
}
return maxKeyCount;
}
private static CompactionRequest createDummyRequest() throws Exception {
// "Files" are totally unused, it's Scanner class below that gives compactor fake KVs.
// But compaction depends on everything under the sun, so stub everything with dummies.
StoreFile sf = mock(StoreFile.class);
StoreFile.Reader r = mock(StoreFile.Reader.class);
when(r.length()).thenReturn(1L);
when(r.getBloomFilterType()).thenReturn(BloomType.NONE);
when(r.getHFileReader()).thenReturn(mock(HFile.Reader.class));
when(r.getStoreFileScanner(anyBoolean(), anyBoolean(), anyBoolean(), anyLong()))
.thenReturn(mock(StoreFileScanner.class));
when(sf.getReader()).thenReturn(r);
when(sf.createReader()).thenReturn(r);
return new CompactionRequest(Arrays.asList(sf));
}
/**
* Return an array of scanners corresponding to the given set of store files, And set the
* ScanQueryMatcher for each store file scanner for further optimization
*/
public static List<StoreFileScanner> getScannersForStoreFiles(
Collection<StoreFile> files,
boolean cacheBlocks,
boolean isCompaction,
ScanQueryMatcher matcher)
throws IOException {
List<StoreFileScanner> scanners = new ArrayList<StoreFileScanner>(files.size());
for (StoreFile file : files) {
StoreFile.Reader r = file.createReader();
StoreFileScanner scanner = r.getStoreFileScanner(cacheBlocks, isCompaction);
scanner.setScanQueryMatcher(matcher);
scanners.add(scanner);
}
return scanners;
}
@Override
public Boolean call() throws Exception {
Thread.currentThread().setName("reader " + readerId);
Random rand = new Random();
StoreFileScanner scanner = reader.getStoreFileScanner(true, pread);
while (System.currentTimeMillis() < endTime) {
byte[] row = createRandomRow(rand, firstRow, lastRow);
KeyValue kvToSeek = new KeyValue(row, family, createRandomQualifier(rand));
if (rand.nextDouble() < 0.0001) {
LOG.info("kvToSeek=" + kvToSeek);
}
boolean seekResult;
try {
seekResult = scanner.seek(kvToSeek);
} catch (IOException ex) {
throw new IOException("Seek failed for key " + kvToSeek + ", pread=" + pread, ex);
}
numSeeks.incrementAndGet();
if (!seekResult) {
error("Seek returned false for row " + Bytes.toStringBinary(row));
return false;
}
for (int i = 0; i < rand.nextInt(10) + 1; ++i) {
KeyValue kv = scanner.next();
numKV.incrementAndGet();
if (i == 0 && kv == null) {
error(
"scanner.next() returned null at the first iteration for "
+ "row "
+ Bytes.toStringBinary(row));
return false;
}
if (kv == null) break;
String keyHashStr = MD5Hash.getMD5AsHex(kv.getKey());
keysRead.add(keyHashStr);
totalBytes.addAndGet(kv.getLength());
}
}
return true;
}
@Override
public boolean shouldUseScanner(Scan scan, SortedSet<byte[]> columns, long oldestUnexpiredTS) {
return reader.passesTimerangeFilter(scan, oldestUnexpiredTS)
&& reader.passesBloomFilter(scan, columns);
}
/**
* Pretend we have done a seek but don't do it yet, if possible. The hope is that we find
* requested columns in more recent files and won't have to seek in older files. Creates a fake
* key/value with the given row/column and the highest (most recent) possible timestamp we might
* get from this file. When users of such "lazy scanner" need to know the next KV precisely (e.g.
* when this scanner is at the top of the heap), they run {@link #enforceSeek()}.
*
* <p>Note that this function does guarantee that the current KV of this scanner will be advanced
* to at least the given KV. Because of this, it does have to do a real seek in cases when the
* seek timestamp is older than the highest timestamp of the file, e.g. when we are trying to seek
* to the next row/column and use OLDEST_TIMESTAMP in the seek key.
*/
@Override
public boolean requestSeek(KeyValue kv, boolean forward, boolean useBloom) throws IOException {
if (kv.getFamilyLength() == 0) {
useBloom = false;
}
boolean haveToSeek = true;
if (useBloom) {
// check ROWCOL Bloom filter first.
if (reader.getBloomFilterType() == StoreFile.BloomType.ROWCOL) {
haveToSeek =
reader.passesGeneralBloomFilter(
kv.getBuffer(),
kv.getRowOffset(),
kv.getRowLength(),
kv.getBuffer(),
kv.getQualifierOffset(),
kv.getQualifierLength());
} else if (this.matcher != null && !matcher.hasNullColumnInQuery() && kv.isDeleteFamily()) {
// if there is no such delete family kv in the store file,
// then no need to seek.
haveToSeek =
reader.passesDeleteFamilyBloomFilter(
kv.getBuffer(), kv.getRowOffset(), kv.getRowLength());
}
}
delayedReseek = forward;
delayedSeekKV = kv;
if (haveToSeek) {
// This row/column might be in this store file (or we did not use the
// Bloom filter), so we still need to seek.
realSeekDone = false;
long maxTimestampInFile = reader.getMaxTimestamp();
long seekTimestamp = kv.getTimestamp();
if (seekTimestamp > maxTimestampInFile) {
// Create a fake key that is not greater than the real next key.
// (Lower timestamps correspond to higher KVs.)
// To understand this better, consider that we are asked to seek to
// a higher timestamp than the max timestamp in this file. We know that
// the next point when we have to consider this file again is when we
// pass the max timestamp of this file (with the same row/column).
cur = kv.createFirstOnRowColTS(maxTimestampInFile);
} else {
// This will be the case e.g. when we need to seek to the next
// row/column, and we don't know exactly what they are, so we set the
// seek key's timestamp to OLDEST_TIMESTAMP to skip the rest of this
// row/column.
enforceSeek();
}
return cur != null;
}
// Multi-column Bloom filter optimization.
// Create a fake key/value, so that this scanner only bubbles up to the top
// of the KeyValueHeap in StoreScanner after we scanned this row/column in
// all other store files. The query matcher will then just skip this fake
// key/value and the store scanner will progress to the next column. This
// is obviously not a "real real" seek, but unlike the fake KV earlier in
// this method, we want this to be propagated to ScanQueryMatcher.
cur = kv.createLastOnRowCol();
realSeekDone = true;
return true;
}
@Override
public long getSequenceID() {
return reader.getSequenceID();
}
public boolean runRandomReadWorkload() throws IOException {
if (inputFileNames.size() != 1) {
throw new IOException("Need exactly one input file for random reads: " + inputFileNames);
}
Path inputPath = new Path(inputFileNames.get(0));
// Make sure we are using caching.
StoreFile storeFile = openStoreFile(inputPath, true);
StoreFile.Reader reader = storeFile.createReader();
LOG.info("First key: " + Bytes.toStringBinary(reader.getFirstKey()));
LOG.info("Last key: " + Bytes.toStringBinary(reader.getLastKey()));
KeyValue firstKV = KeyValue.createKeyValueFromKey(reader.getFirstKey());
firstRow = firstKV.getRow();
KeyValue lastKV = KeyValue.createKeyValueFromKey(reader.getLastKey());
lastRow = lastKV.getRow();
byte[] family = firstKV.getFamily();
if (!Bytes.equals(family, lastKV.getFamily())) {
LOG.error(
"First and last key have different families: "
+ Bytes.toStringBinary(family)
+ " and "
+ Bytes.toStringBinary(lastKV.getFamily()));
return false;
}
if (Bytes.equals(firstRow, lastRow)) {
LOG.error(
"First and last row are the same, cannot run read workload: "
+ "firstRow="
+ Bytes.toStringBinary(firstRow)
+ ", "
+ "lastRow="
+ Bytes.toStringBinary(lastRow));
return false;
}
ExecutorService exec = Executors.newFixedThreadPool(numReadThreads + 1);
int numCompleted = 0;
int numFailed = 0;
try {
ExecutorCompletionService<Boolean> ecs = new ExecutorCompletionService<Boolean>(exec);
endTime = System.currentTimeMillis() + 1000 * durationSec;
boolean pread = true;
for (int i = 0; i < numReadThreads; ++i) ecs.submit(new RandomReader(i, reader, pread));
ecs.submit(new StatisticsPrinter());
Future<Boolean> result;
while (true) {
try {
result = ecs.poll(endTime + 1000 - System.currentTimeMillis(), TimeUnit.MILLISECONDS);
if (result == null) break;
try {
if (result.get()) {
++numCompleted;
} else {
++numFailed;
}
} catch (ExecutionException e) {
LOG.error("Worker thread failure", e.getCause());
++numFailed;
}
} catch (InterruptedException ex) {
LOG.error("Interrupted after " + numCompleted + " workers completed");
Thread.currentThread().interrupt();
continue;
}
}
} finally {
storeFile.closeReader(true);
exec.shutdown();
BlockCache c = cacheConf.getBlockCache();
if (c != null) {
c.shutdown();
}
}
LOG.info("Worker threads completed: " + numCompleted);
LOG.info("Worker threads failed: " + numFailed);
return true;
}
/**
* Do a minor/major compaction on an explicit set of storefiles from a Store.
*
* @param request the requested compaction that contains all necessary information to complete the
* compaction (i.e. the store, the files, etc.)
* @return Product of compaction or null if all cells expired or deleted and nothing made it
* through the compaction.
* @throws IOException
*/
StoreFile.Writer compact(CompactionRequest request, long maxId) throws IOException {
// Calculate maximum key count after compaction (for blooms)
// Also calculate earliest put timestamp if major compaction
int maxKeyCount = 0;
long earliestPutTs = HConstants.LATEST_TIMESTAMP;
long maxMVCCReadpoint = 0;
// pull out the interesting things from the CR for ease later
final Store store = request.getStore();
final boolean majorCompaction = request.isMajor();
final List<StoreFile> filesToCompact = request.getFiles();
for (StoreFile file : filesToCompact) {
StoreFile.Reader r = file.getReader();
if (r == null) {
LOG.warn("Null reader for " + file.getPath());
continue;
}
// NOTE: getFilterEntries could cause under-sized blooms if the user
// switches bloom type (e.g. from ROW to ROWCOL)
long keyCount =
(r.getBloomFilterType() == store.getFamily().getBloomFilterType())
? r.getFilterEntries()
: r.getEntries();
maxKeyCount += keyCount;
// Calculate the maximum MVCC readpoint used in any of the involved files
Map<byte[], byte[]> fileInfo = r.loadFileInfo();
byte[] tmp = fileInfo.get(HFileWriterV2.MAX_MEMSTORE_TS_KEY);
if (tmp != null) {
maxMVCCReadpoint = Math.max(maxMVCCReadpoint, Bytes.toLong(tmp));
}
// For major compactions calculate the earliest put timestamp
// of all involved storefiles. This is used to remove
// family delete marker during the compaction.
if (majorCompaction) {
tmp = fileInfo.get(StoreFile.EARLIEST_PUT_TS);
if (tmp == null) {
// There's a file with no information, must be an old one
// assume we have very old puts
earliestPutTs = HConstants.OLDEST_TIMESTAMP;
} else {
earliestPutTs = Math.min(earliestPutTs, Bytes.toLong(tmp));
}
}
if (LOG.isDebugEnabled()) {
LOG.debug(
"Compacting "
+ file
+ ", keycount="
+ keyCount
+ ", bloomtype="
+ r.getBloomFilterType().toString()
+ ", size="
+ StringUtils.humanReadableInt(r.length())
+ ", encoding="
+ r.getHFileReader().getEncodingOnDisk()
+ (majorCompaction ? ", earliestPutTs=" + earliestPutTs : ""));
}
}
// keep track of compaction progress
this.progress = new CompactionProgress(maxKeyCount);
// Get some configs
int compactionKVMax = getConf().getInt("hbase.hstore.compaction.kv.max", 10);
Compression.Algorithm compression = store.getFamily().getCompression();
// Avoid overriding compression setting for major compactions if the user
// has not specified it separately
Compression.Algorithm compactionCompression =
(store.getFamily().getCompactionCompression() != Compression.Algorithm.NONE)
? store.getFamily().getCompactionCompression()
: compression;
// For each file, obtain a scanner:
List<StoreFileScanner> scanners =
StoreFileScanner.getScannersForStoreFiles(filesToCompact, false, false, true);
// Make the instantiation lazy in case compaction produces no product; i.e.
// where all source cells are expired or deleted.
StoreFile.Writer writer = null;
// Find the smallest read point across all the Scanners.
long smallestReadPoint = store.getHRegion().getSmallestReadPoint();
MultiVersionConsistencyControl.setThreadReadPoint(smallestReadPoint);
try {
InternalScanner scanner = null;
try {
if (store.getHRegion().getCoprocessorHost() != null) {
scanner =
store
.getHRegion()
.getCoprocessorHost()
.preCompactScannerOpen(
store,
scanners,
majorCompaction ? ScanType.MAJOR_COMPACT : ScanType.MINOR_COMPACT,
earliestPutTs,
request);
}
if (scanner == null) {
Scan scan = new Scan();
scan.setMaxVersions(store.getFamily().getMaxVersions());
/* Include deletes, unless we are doing a major compaction */
scanner =
new StoreScanner(
store,
store.getScanInfo(),
scan,
scanners,
majorCompaction ? ScanType.MAJOR_COMPACT : ScanType.MINOR_COMPACT,
smallestReadPoint,
earliestPutTs);
}
if (store.getHRegion().getCoprocessorHost() != null) {
InternalScanner cpScanner =
store.getHRegion().getCoprocessorHost().preCompact(store, scanner, request);
// NULL scanner returned from coprocessor hooks means skip normal processing
if (cpScanner == null) {
return null;
}
scanner = cpScanner;
}
int bytesWritten = 0;
// since scanner.next() can return 'false' but still be delivering data,
// we have to use a do/while loop.
List<KeyValue> kvs = new ArrayList<KeyValue>();
// Limit to "hbase.hstore.compaction.kv.max" (default 10) to avoid OOME
boolean hasMore;
do {
hasMore = scanner.next(kvs, compactionKVMax);
if (writer == null && !kvs.isEmpty()) {
writer =
store.createWriterInTmp(
maxKeyCount,
compactionCompression,
true,
maxMVCCReadpoint >= smallestReadPoint);
}
if (writer != null) {
// output to writer:
for (KeyValue kv : kvs) {
if (kv.getMemstoreTS() <= smallestReadPoint) {
kv.setMemstoreTS(0);
}
writer.append(kv);
// update progress per key
++progress.currentCompactedKVs;
// check periodically to see if a system stop is requested
if (Store.closeCheckInterval > 0) {
bytesWritten += kv.getLength();
if (bytesWritten > Store.closeCheckInterval) {
bytesWritten = 0;
isInterrupted(store, writer);
}
}
}
}
kvs.clear();
} while (hasMore);
} finally {
if (scanner != null) {
scanner.close();
}
}
} finally {
if (writer != null) {
writer.appendMetadata(maxId, majorCompaction);
writer.close();
}
}
return writer;
}