@Test
public void testVerifyingIndexOutput() throws IOException {
Directory dir = newDirectory();
IndexOutput output = dir.createOutput("foo.bar", IOContext.DEFAULT);
int iters = scaledRandomIntBetween(10, 100);
for (int i = 0; i < iters; i++) {
BytesRef bytesRef = new BytesRef(TestUtil.randomRealisticUnicodeString(random(), 10, 1024));
output.writeBytes(bytesRef.bytes, bytesRef.offset, bytesRef.length);
}
CodecUtil.writeFooter(output);
output.close();
IndexInput indexInput = dir.openInput("foo.bar", IOContext.DEFAULT);
String checksum = Store.digestToString(CodecUtil.retrieveChecksum(indexInput));
indexInput.seek(0);
BytesRef ref = new BytesRef(scaledRandomIntBetween(1, 1024));
long length = indexInput.length();
IndexOutput verifyingOutput =
new Store.LuceneVerifyingIndexOutput(
new StoreFileMetaData("foo1.bar", length, checksum),
dir.createOutput("foo1.bar", IOContext.DEFAULT));
while (length > 0) {
if (random().nextInt(10) == 0) {
verifyingOutput.writeByte(indexInput.readByte());
length--;
} else {
int min = (int) Math.min(length, ref.bytes.length);
indexInput.readBytes(ref.bytes, ref.offset, min);
verifyingOutput.writeBytes(ref.bytes, ref.offset, min);
length -= min;
}
}
Store.verify(verifyingOutput);
verifyingOutput.writeByte((byte) 0x0);
try {
Store.verify(verifyingOutput);
fail("should be a corrupted index");
} catch (CorruptIndexException | IndexFormatTooOldException | IndexFormatTooNewException ex) {
// ok
}
IOUtils.close(indexInput, verifyingOutput, dir);
}
@Test
public void testVerifyingIndexInput() throws IOException {
Directory dir = newDirectory();
IndexOutput output = dir.createOutput("foo.bar", IOContext.DEFAULT);
int iters = scaledRandomIntBetween(10, 100);
for (int i = 0; i < iters; i++) {
BytesRef bytesRef = new BytesRef(TestUtil.randomRealisticUnicodeString(random(), 10, 1024));
output.writeBytes(bytesRef.bytes, bytesRef.offset, bytesRef.length);
}
CodecUtil.writeFooter(output);
output.close();
// Check file
IndexInput indexInput = dir.openInput("foo.bar", IOContext.DEFAULT);
long checksum = CodecUtil.retrieveChecksum(indexInput);
indexInput.seek(0);
IndexInput verifyingIndexInput =
new Store.VerifyingIndexInput(dir.openInput("foo.bar", IOContext.DEFAULT));
readIndexInputFullyWithRandomSeeks(verifyingIndexInput);
Store.verify(verifyingIndexInput);
assertThat(checksum, equalTo(((ChecksumIndexInput) verifyingIndexInput).getChecksum()));
IOUtils.close(indexInput, verifyingIndexInput);
// Corrupt file and check again
corruptFile(dir, "foo.bar", "foo1.bar");
verifyingIndexInput =
new Store.VerifyingIndexInput(dir.openInput("foo1.bar", IOContext.DEFAULT));
readIndexInputFullyWithRandomSeeks(verifyingIndexInput);
try {
Store.verify(verifyingIndexInput);
fail("should be a corrupted index");
} catch (CorruptIndexException | IndexFormatTooOldException | IndexFormatTooNewException ex) {
// ok
}
IOUtils.close(verifyingIndexInput);
IOUtils.close(dir);
}
/** Sole constructor. */
public BlockTreeTermsReader(PostingsReaderBase postingsReader, SegmentReadState state)
throws IOException {
boolean success = false;
IndexInput indexIn = null;
this.postingsReader = postingsReader;
this.segment = state.segmentInfo.name;
String termsName =
IndexFileNames.segmentFileName(segment, state.segmentSuffix, TERMS_EXTENSION);
try {
termsIn = state.directory.openInput(termsName, state.context);
version =
CodecUtil.checkIndexHeader(
termsIn,
TERMS_CODEC_NAME,
VERSION_START,
VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
String indexName =
IndexFileNames.segmentFileName(segment, state.segmentSuffix, TERMS_INDEX_EXTENSION);
indexIn = state.directory.openInput(indexName, state.context);
CodecUtil.checkIndexHeader(
indexIn,
TERMS_INDEX_CODEC_NAME,
version,
version,
state.segmentInfo.getId(),
state.segmentSuffix);
CodecUtil.checksumEntireFile(indexIn);
// Have PostingsReader init itself
postingsReader.init(termsIn, state);
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(termsIn);
// Read per-field details
seekDir(termsIn, dirOffset);
seekDir(indexIn, indexDirOffset);
final int numFields = termsIn.readVInt();
if (numFields < 0) {
throw new CorruptIndexException("invalid numFields: " + numFields, termsIn);
}
for (int i = 0; i < numFields; ++i) {
final int field = termsIn.readVInt();
final long numTerms = termsIn.readVLong();
if (numTerms <= 0) {
throw new CorruptIndexException("Illegal numTerms for field number: " + field, termsIn);
}
final int numBytes = termsIn.readVInt();
if (numBytes < 0) {
throw new CorruptIndexException(
"invalid rootCode for field number: " + field + ", numBytes=" + numBytes, termsIn);
}
final BytesRef rootCode = new BytesRef(new byte[numBytes]);
termsIn.readBytes(rootCode.bytes, 0, numBytes);
rootCode.length = numBytes;
final FieldInfo fieldInfo = state.fieldInfos.fieldInfo(field);
if (fieldInfo == null) {
throw new CorruptIndexException("invalid field number: " + field, termsIn);
}
final long sumTotalTermFreq =
fieldInfo.getIndexOptions() == IndexOptions.DOCS ? -1 : termsIn.readVLong();
final long sumDocFreq = termsIn.readVLong();
final int docCount = termsIn.readVInt();
final int longsSize = termsIn.readVInt();
if (longsSize < 0) {
throw new CorruptIndexException(
"invalid longsSize for field: " + fieldInfo.name + ", longsSize=" + longsSize,
termsIn);
}
BytesRef minTerm = readBytesRef(termsIn);
BytesRef maxTerm = readBytesRef(termsIn);
if (docCount < 0
|| docCount > state.segmentInfo.getDocCount()) { // #docs with field must be <= #docs
throw new CorruptIndexException(
"invalid docCount: " + docCount + " maxDoc: " + state.segmentInfo.getDocCount(),
termsIn);
}
if (sumDocFreq < docCount) { // #postings must be >= #docs with field
throw new CorruptIndexException(
"invalid sumDocFreq: " + sumDocFreq + " docCount: " + docCount, termsIn);
}
if (sumTotalTermFreq != -1
&& sumTotalTermFreq < sumDocFreq) { // #positions must be >= #postings
throw new CorruptIndexException(
"invalid sumTotalTermFreq: " + sumTotalTermFreq + " sumDocFreq: " + sumDocFreq,
termsIn);
}
final long indexStartFP = indexIn.readVLong();
FieldReader previous =
fields.put(
fieldInfo.name,
new FieldReader(
this,
fieldInfo,
numTerms,
rootCode,
sumTotalTermFreq,
sumDocFreq,
docCount,
indexStartFP,
longsSize,
indexIn,
minTerm,
maxTerm));
if (previous != null) {
throw new CorruptIndexException("duplicate field: " + fieldInfo.name, termsIn);
}
}
indexIn.close();
success = true;
} finally {
if (!success) {
// this.close() will close in:
IOUtils.closeWhileHandlingException(indexIn, this);
}
}
}
public void testCheckIntegrity() throws IOException {
Directory dir = newDirectory();
long luceneFileLength = 0;
try (IndexOutput output = dir.createOutput("lucene_checksum.bin", IOContext.DEFAULT)) {
int iters = scaledRandomIntBetween(10, 100);
for (int i = 0; i < iters; i++) {
BytesRef bytesRef = new BytesRef(TestUtil.randomRealisticUnicodeString(random(), 10, 1024));
output.writeBytes(bytesRef.bytes, bytesRef.offset, bytesRef.length);
luceneFileLength += bytesRef.length;
}
CodecUtil.writeFooter(output);
luceneFileLength += CodecUtil.footerLength();
}
final Adler32 adler32 = new Adler32();
long legacyFileLength = 0;
try (IndexOutput output = dir.createOutput("legacy.bin", IOContext.DEFAULT)) {
int iters = scaledRandomIntBetween(10, 100);
for (int i = 0; i < iters; i++) {
BytesRef bytesRef = new BytesRef(TestUtil.randomRealisticUnicodeString(random(), 10, 1024));
output.writeBytes(bytesRef.bytes, bytesRef.offset, bytesRef.length);
adler32.update(bytesRef.bytes, bytesRef.offset, bytesRef.length);
legacyFileLength += bytesRef.length;
}
}
final long luceneChecksum;
final long adler32LegacyChecksum = adler32.getValue();
try (IndexInput indexInput = dir.openInput("lucene_checksum.bin", IOContext.DEFAULT)) {
assertEquals(luceneFileLength, indexInput.length());
luceneChecksum = CodecUtil.retrieveChecksum(indexInput);
}
{ // positive check
StoreFileMetaData lucene =
new StoreFileMetaData(
"lucene_checksum.bin",
luceneFileLength,
Store.digestToString(luceneChecksum),
Version.LUCENE_4_8_0);
StoreFileMetaData legacy =
new StoreFileMetaData(
"legacy.bin", legacyFileLength, Store.digestToString(adler32LegacyChecksum));
assertTrue(legacy.hasLegacyChecksum());
assertFalse(lucene.hasLegacyChecksum());
assertTrue(Store.checkIntegrityNoException(lucene, dir));
assertTrue(Store.checkIntegrityNoException(legacy, dir));
}
{ // negative check - wrong checksum
StoreFileMetaData lucene =
new StoreFileMetaData(
"lucene_checksum.bin",
luceneFileLength,
Store.digestToString(luceneChecksum + 1),
Version.LUCENE_4_8_0);
StoreFileMetaData legacy =
new StoreFileMetaData(
"legacy.bin", legacyFileLength, Store.digestToString(adler32LegacyChecksum + 1));
assertTrue(legacy.hasLegacyChecksum());
assertFalse(lucene.hasLegacyChecksum());
assertFalse(Store.checkIntegrityNoException(lucene, dir));
assertFalse(Store.checkIntegrityNoException(legacy, dir));
}
{ // negative check - wrong length
StoreFileMetaData lucene =
new StoreFileMetaData(
"lucene_checksum.bin",
luceneFileLength + 1,
Store.digestToString(luceneChecksum),
Version.LUCENE_4_8_0);
StoreFileMetaData legacy =
new StoreFileMetaData(
"legacy.bin", legacyFileLength + 1, Store.digestToString(adler32LegacyChecksum));
assertTrue(legacy.hasLegacyChecksum());
assertFalse(lucene.hasLegacyChecksum());
assertFalse(Store.checkIntegrityNoException(lucene, dir));
assertFalse(Store.checkIntegrityNoException(legacy, dir));
}
{ // negative check - wrong file
StoreFileMetaData lucene =
new StoreFileMetaData(
"legacy.bin",
luceneFileLength,
Store.digestToString(luceneChecksum),
Version.LUCENE_4_8_0);
StoreFileMetaData legacy =
new StoreFileMetaData(
"lucene_checksum.bin", legacyFileLength, Store.digestToString(adler32LegacyChecksum));
assertTrue(legacy.hasLegacyChecksum());
assertFalse(lucene.hasLegacyChecksum());
assertFalse(Store.checkIntegrityNoException(lucene, dir));
assertFalse(Store.checkIntegrityNoException(legacy, dir));
}
dir.close();
}
@Test
public void testMixedChecksums() throws IOException {
final ShardId shardId = new ShardId(new Index("index"), 1);
DirectoryService directoryService = new LuceneManagedDirectoryService(random());
Store store =
new Store(
shardId,
ImmutableSettings.EMPTY,
directoryService,
randomDistributor(directoryService),
new DummyShardLock(shardId));
// this time random codec....
IndexWriter writer =
new IndexWriter(
store.directory(),
newIndexWriterConfig(random(), new MockAnalyzer(random()))
.setCodec(actualDefaultCodec()));
int docs = 1 + random().nextInt(100);
for (int i = 0; i < docs; i++) {
Document doc = new Document();
doc.add(
new TextField("id", "" + i, random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new TextField(
"body",
TestUtil.randomRealisticUnicodeString(random()),
random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new SortedDocValuesField(
"dv", new BytesRef(TestUtil.randomRealisticUnicodeString(random()))));
writer.addDocument(doc);
}
if (random().nextBoolean()) {
for (int i = 0; i < docs; i++) {
if (random().nextBoolean()) {
Document doc = new Document();
doc.add(
new TextField(
"id", "" + i, random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new TextField(
"body",
TestUtil.randomRealisticUnicodeString(random()),
random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
writer.updateDocument(new Term("id", "" + i), doc);
}
}
}
if (random().nextBoolean()) {
DirectoryReader.open(writer, random().nextBoolean()).close(); // flush
}
Store.MetadataSnapshot metadata;
// check before we committed
try {
store.getMetadata();
fail("no index present - expected exception");
} catch (IndexNotFoundException ex) {
// expected
}
assertThat(store.getMetadataOrEmpty(), is(Store.MetadataSnapshot.EMPTY)); // nothing committed
writer.commit();
writer.close();
Store.LegacyChecksums checksums = new Store.LegacyChecksums();
metadata = store.getMetadata();
assertThat(metadata.asMap().isEmpty(), is(false));
for (StoreFileMetaData meta : metadata) {
try (IndexInput input = store.directory().openInput(meta.name(), IOContext.DEFAULT)) {
if (meta.checksum() == null) {
String checksum = null;
try {
CodecUtil.retrieveChecksum(input);
fail("expected a corrupt index - posting format has not checksums");
} catch (CorruptIndexException
| IndexFormatTooOldException
| IndexFormatTooNewException ex) {
try (ChecksumIndexInput checksumIndexInput =
store.directory().openChecksumInput(meta.name(), IOContext.DEFAULT)) {
checksumIndexInput.seek(meta.length());
checksum = Store.digestToString(checksumIndexInput.getChecksum());
}
// fine - it's a postings format without checksums
checksums.add(new StoreFileMetaData(meta.name(), meta.length(), checksum, null));
}
} else {
String checksum = Store.digestToString(CodecUtil.retrieveChecksum(input));
assertThat(
"File: " + meta.name() + " has a different checksum",
meta.checksum(),
equalTo(checksum));
assertThat(meta.hasLegacyChecksum(), equalTo(false));
assertThat(meta.writtenBy(), equalTo(Version.LATEST));
}
}
}
assertConsistent(store, metadata);
checksums.write(store);
metadata = store.getMetadata();
assertThat(metadata.asMap().isEmpty(), is(false));
for (StoreFileMetaData meta : metadata) {
assertThat(
"file: " + meta.name() + " has a null checksum", meta.checksum(), not(nullValue()));
if (meta.hasLegacyChecksum()) {
try (ChecksumIndexInput checksumIndexInput =
store.directory().openChecksumInput(meta.name(), IOContext.DEFAULT)) {
checksumIndexInput.seek(meta.length());
assertThat(
meta.checksum(), equalTo(Store.digestToString(checksumIndexInput.getChecksum())));
}
} else {
try (IndexInput input = store.directory().openInput(meta.name(), IOContext.DEFAULT)) {
String checksum = Store.digestToString(CodecUtil.retrieveChecksum(input));
assertThat(
"File: " + meta.name() + " has a different checksum",
meta.checksum(),
equalTo(checksum));
assertThat(meta.hasLegacyChecksum(), equalTo(false));
assertThat(meta.writtenBy(), equalTo(Version.LATEST));
}
}
}
assertConsistent(store, metadata);
TestUtil.checkIndex(store.directory());
assertDeleteContent(store, directoryService);
IOUtils.close(store);
}
@Test
public void testNewChecksums() throws IOException {
final ShardId shardId = new ShardId(new Index("index"), 1);
DirectoryService directoryService = new LuceneManagedDirectoryService(random());
Store store =
new Store(
shardId,
ImmutableSettings.EMPTY,
directoryService,
randomDistributor(directoryService),
new DummyShardLock(shardId));
// set default codec - all segments need checksums
IndexWriter writer =
new IndexWriter(
store.directory(),
newIndexWriterConfig(random(), new MockAnalyzer(random()))
.setCodec(actualDefaultCodec()));
int docs = 1 + random().nextInt(100);
for (int i = 0; i < docs; i++) {
Document doc = new Document();
doc.add(
new TextField("id", "" + i, random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new TextField(
"body",
TestUtil.randomRealisticUnicodeString(random()),
random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new SortedDocValuesField(
"dv", new BytesRef(TestUtil.randomRealisticUnicodeString(random()))));
writer.addDocument(doc);
}
if (random().nextBoolean()) {
for (int i = 0; i < docs; i++) {
if (random().nextBoolean()) {
Document doc = new Document();
doc.add(
new TextField(
"id", "" + i, random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new TextField(
"body",
TestUtil.randomRealisticUnicodeString(random()),
random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
writer.updateDocument(new Term("id", "" + i), doc);
}
}
}
if (random().nextBoolean()) {
DirectoryReader.open(writer, random().nextBoolean()).close(); // flush
}
Store.MetadataSnapshot metadata;
// check before we committed
try {
store.getMetadata();
fail("no index present - expected exception");
} catch (IndexNotFoundException ex) {
// expected
}
assertThat(store.getMetadataOrEmpty(), is(Store.MetadataSnapshot.EMPTY)); // nothing committed
writer.commit();
writer.close();
metadata = store.getMetadata();
assertThat(metadata.asMap().isEmpty(), is(false));
for (StoreFileMetaData meta : metadata) {
try (IndexInput input = store.directory().openInput(meta.name(), IOContext.DEFAULT)) {
String checksum = Store.digestToString(CodecUtil.retrieveChecksum(input));
assertThat(
"File: " + meta.name() + " has a different checksum",
meta.checksum(),
equalTo(checksum));
assertThat(meta.hasLegacyChecksum(), equalTo(false));
assertThat(meta.writtenBy(), equalTo(Version.LATEST));
if (meta.name().endsWith(".si") || meta.name().startsWith("segments_")) {
assertThat(meta.hash().length, greaterThan(0));
}
}
}
assertConsistent(store, metadata);
TestUtil.checkIndex(store.directory());
assertDeleteContent(store, directoryService);
IOUtils.close(store);
}
private boolean doRestore() throws Exception {
Path backupPath = Paths.get(backupLocation).resolve(backupName);
SimpleDateFormat dateFormat = new SimpleDateFormat(SnapShooter.DATE_FMT, Locale.ROOT);
String restoreIndexName = "restore." + dateFormat.format(new Date());
String restoreIndexPath = core.getDataDir() + restoreIndexName;
Directory restoreIndexDir = null;
Directory indexDir = null;
try (Directory backupDir = FSDirectory.open(backupPath)) {
final Version version =
IndexFetcher.checkOldestVersion(SegmentInfos.readLatestCommit(backupDir));
restoreIndexDir =
core.getDirectoryFactory()
.get(
restoreIndexPath,
DirectoryFactory.DirContext.DEFAULT,
core.getSolrConfig().indexConfig.lockType);
// Prefer local copy.
indexDir =
core.getDirectoryFactory()
.get(
core.getIndexDir(),
DirectoryFactory.DirContext.DEFAULT,
core.getSolrConfig().indexConfig.lockType);
// Move all files from backupDir to restoreIndexDir
for (String filename : backupDir.listAll()) {
checkInterrupted();
log.info("Copying file {} to restore directory ", filename);
try (IndexInput indexInput = backupDir.openInput(filename, IOContext.READONCE)) {
Long checksum = null;
try {
checksum = CodecUtil.retrieveChecksum(indexInput);
} catch (Exception e) {
log.warn("Could not read checksum from index file: " + filename, e);
}
long length = indexInput.length();
IndexFetcher.CompareResult compareResult =
IndexFetcher.compareFile(indexDir, version, filename, length, checksum);
if (!compareResult.equal
|| (!compareResult.checkSummed
&& (filename.endsWith(".si")
|| filename.endsWith(".liv")
|| filename.startsWith("segments_")))) {
restoreIndexDir.copyFrom(backupDir, filename, filename, IOContext.READONCE);
} else {
// prefer local copy
restoreIndexDir.copyFrom(indexDir, filename, filename, IOContext.READONCE);
}
} catch (Exception e) {
throw new SolrException(
SolrException.ErrorCode.UNKNOWN, "Exception while restoring the backup index", e);
}
}
log.debug("Switching directories");
IndexFetcher.modifyIndexProps(core, restoreIndexName);
boolean success;
try {
core.getUpdateHandler().newIndexWriter(false);
openNewSearcher();
success = true;
log.info("Successfully restored to the backup index");
} catch (Exception e) {
// Rollback to the old index directory. Delete the restore index directory and mark the
// restore as failed.
log.warn("Could not switch to restored index. Rolling back to the current index");
Directory dir = null;
try {
dir =
core.getDirectoryFactory()
.get(
core.getDataDir(),
DirectoryFactory.DirContext.META_DATA,
core.getSolrConfig().indexConfig.lockType);
dir.deleteFile(IndexFetcher.INDEX_PROPERTIES);
} finally {
if (dir != null) {
core.getDirectoryFactory().release(dir);
}
}
core.getDirectoryFactory().doneWithDirectory(restoreIndexDir);
core.getDirectoryFactory().remove(restoreIndexDir);
core.getUpdateHandler().newIndexWriter(false);
openNewSearcher();
throw new SolrException(
SolrException.ErrorCode.UNKNOWN, "Exception while restoring the backup index", e);
}
if (success) {
core.getDirectoryFactory().doneWithDirectory(indexDir);
core.getDirectoryFactory().remove(indexDir);
}
return true;
} finally {
if (restoreIndexDir != null) {
core.getDirectoryFactory().release(restoreIndexDir);
}
if (indexDir != null) {
core.getDirectoryFactory().release(indexDir);
}
}
}
Lucene42DocValuesProducer(
SegmentReadState state,
String dataCodec,
String dataExtension,
String metaCodec,
String metaExtension)
throws IOException {
maxDoc = state.segmentInfo.maxDoc();
merging = false;
String metaName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, metaExtension);
// read in the entries from the metadata file.
ChecksumIndexInput in = state.directory.openChecksumInput(metaName, state.context);
boolean success = false;
ramBytesUsed = new AtomicLong(RamUsageEstimator.shallowSizeOfInstance(getClass()));
try {
version = CodecUtil.checkHeader(in, metaCodec, VERSION_START, VERSION_CURRENT);
numerics = new HashMap<>();
binaries = new HashMap<>();
fsts = new HashMap<>();
numEntries = readFields(in, state.fieldInfos);
if (version >= VERSION_CHECKSUM) {
CodecUtil.checkFooter(in);
} else {
CodecUtil.checkEOF(in);
}
success = true;
} finally {
if (success) {
IOUtils.close(in);
} else {
IOUtils.closeWhileHandlingException(in);
}
}
String dataName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, dataExtension);
this.data = state.directory.openInput(dataName, state.context);
success = false;
try {
final int version2 = CodecUtil.checkHeader(data, dataCodec, VERSION_START, VERSION_CURRENT);
if (version != version2) {
throw new CorruptIndexException(
"Format versions mismatch: meta=" + version + ", data=" + version2, data);
}
if (version >= VERSION_CHECKSUM) {
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(data);
}
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(this.data);
}
}
}
public BlockTermsReader(
TermsIndexReaderBase indexReader, PostingsReaderBase postingsReader, SegmentReadState state)
throws IOException {
this.postingsReader = postingsReader;
String filename =
IndexFileNames.segmentFileName(
state.segmentInfo.name, state.segmentSuffix, BlockTermsWriter.TERMS_EXTENSION);
in = state.directory.openInput(filename, state.context);
boolean success = false;
try {
CodecUtil.checkIndexHeader(
in,
BlockTermsWriter.CODEC_NAME,
BlockTermsWriter.VERSION_START,
BlockTermsWriter.VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
// Have PostingsReader init itself
postingsReader.init(in, state);
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(in);
// Read per-field details
seekDir(in);
final int numFields = in.readVInt();
if (numFields < 0) {
throw new CorruptIndexException("invalid number of fields: " + numFields, in);
}
for (int i = 0; i < numFields; i++) {
final int field = in.readVInt();
final long numTerms = in.readVLong();
assert numTerms >= 0;
final long termsStartPointer = in.readVLong();
final FieldInfo fieldInfo = state.fieldInfos.fieldInfo(field);
final long sumTotalTermFreq =
fieldInfo.getIndexOptions() == IndexOptions.DOCS ? -1 : in.readVLong();
final long sumDocFreq = in.readVLong();
final int docCount = in.readVInt();
final int longsSize = in.readVInt();
if (docCount < 0
|| docCount > state.segmentInfo.maxDoc()) { // #docs with field must be <= #docs
throw new CorruptIndexException(
"invalid docCount: " + docCount + " maxDoc: " + state.segmentInfo.maxDoc(), in);
}
if (sumDocFreq < docCount) { // #postings must be >= #docs with field
throw new CorruptIndexException(
"invalid sumDocFreq: " + sumDocFreq + " docCount: " + docCount, in);
}
if (sumTotalTermFreq != -1
&& sumTotalTermFreq < sumDocFreq) { // #positions must be >= #postings
throw new CorruptIndexException(
"invalid sumTotalTermFreq: " + sumTotalTermFreq + " sumDocFreq: " + sumDocFreq, in);
}
FieldReader previous =
fields.put(
fieldInfo.name,
new FieldReader(
fieldInfo,
numTerms,
termsStartPointer,
sumTotalTermFreq,
sumDocFreq,
docCount,
longsSize));
if (previous != null) {
throw new CorruptIndexException("duplicate fields: " + fieldInfo.name, in);
}
}
success = true;
} finally {
if (!success) {
in.close();
}
}
this.indexReader = indexReader;
}
Lucene70NormsProducer(
SegmentReadState state,
String dataCodec,
String dataExtension,
String metaCodec,
String metaExtension)
throws IOException {
maxDoc = state.segmentInfo.maxDoc();
String metaName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, metaExtension);
int version = -1;
// read in the entries from the metadata file.
try (ChecksumIndexInput in = state.directory.openChecksumInput(metaName, state.context)) {
Throwable priorE = null;
try {
version =
CodecUtil.checkIndexHeader(
in,
metaCodec,
VERSION_START,
VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
readFields(in, state.fieldInfos);
} catch (Throwable exception) {
priorE = exception;
} finally {
CodecUtil.checkFooter(in, priorE);
}
}
String dataName =
IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, dataExtension);
data = state.directory.openInput(dataName, state.context);
boolean success = false;
try {
final int version2 =
CodecUtil.checkIndexHeader(
data,
dataCodec,
VERSION_START,
VERSION_CURRENT,
state.segmentInfo.getId(),
state.segmentSuffix);
if (version != version2) {
throw new CorruptIndexException(
"Format versions mismatch: meta=" + version + ",data=" + version2, data);
}
// NOTE: data file is too costly to verify checksum against all the bytes on open,
// but for now we at least verify proper structure of the checksum footer: which looks
// for FOOTER_MAGIC + algorithmID. This is cheap and can detect some forms of corruption
// such as file truncation.
CodecUtil.retrieveChecksum(data);
success = true;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(this.data);
}
}
}