///<summary>
/// Added safety check on last batch of additions for < 3.0 based Lucene implementations
/// <see cref="https://issues.apache.org/jira/browse/LUCENE-3255"/>
///</summary>
///<param name="dir">return true if OK false if a potentially corrupted segments file was removed</param>
public static bool CheckLastCommitOk(FSDirectory dir)
{
string latestSeg = SegmentInfos.GetCurrentSegmentFileName(dir);
IndexInput input = dir.OpenInput(latestSeg);
int format = input.ReadInt();
input.Close();
if (format != SegmentInfos.FORMAT_DIAGNOSTICS)
{
dir.DeleteFile(latestSeg);
return(false);
}
return(true);
}
private IndexSegmentsInfo GetCurrentSegmentsInfo()
{
var segmentInfos = new SegmentInfos();
var result = new IndexSegmentsInfo();
try
{
segmentInfos.Read(directory);
result.Generation = segmentInfos.Generation;
result.SegmentsFileName = segmentInfos.GetCurrentSegmentFileName();
result.ReferencedFiles = segmentInfos.Files(directory, false);
}
catch (CorruptIndexException ex)
{
logIndexing.WarnException(string.Format("Could not read segment information for an index '{0}'", indexId), ex);
result.IsIndexCorrupted = true;
}
return(result);
}
/// <summary>Returns a <see cref="Status" /> instance detailing
/// the state of the index.
///
/// </summary>
/// <param name="onlySegments">list of specific segment names to check
///
/// <p/>As this method checks every byte in the specified
/// segments, on a large index it can take quite a long
/// time to run.
///
/// <p/><b>WARNING</b>: make sure
/// you only call this when the index is not opened by any
/// writer.
/// </param>
public virtual Status CheckIndex_Renamed_Method(List<string> onlySegments)
{
System.Globalization.NumberFormatInfo nf = System.Globalization.CultureInfo.CurrentCulture.NumberFormat;
SegmentInfos sis = new SegmentInfos();
Status result = new Status();
result.dir = dir;
try
{
sis.Read(dir);
}
catch (System.Exception t)
{
Msg("ERROR: could not read any segments file in directory");
result.missingSegments = true;
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
return result;
}
int numSegments = sis.Count;
var segmentsFileName = sis.GetCurrentSegmentFileName();
IndexInput input = null;
try
{
input = dir.OpenInput(segmentsFileName);
}
catch (System.Exception t)
{
Msg("ERROR: could not open segments file in directory");
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
result.cantOpenSegments = true;
return result;
}
int format = 0;
try
{
format = input.ReadInt();
}
catch (System.Exception t)
{
Msg("ERROR: could not read segment file version in directory");
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
result.missingSegmentVersion = true;
return result;
}
finally
{
if (input != null)
input.Close();
}
System.String sFormat = "";
bool skip = false;
if (format == SegmentInfos.FORMAT)
sFormat = "FORMAT [Lucene Pre-2.1]";
if (format == SegmentInfos.FORMAT_LOCKLESS)
sFormat = "FORMAT_LOCKLESS [Lucene 2.1]";
else if (format == SegmentInfos.FORMAT_SINGLE_NORM_FILE)
sFormat = "FORMAT_SINGLE_NORM_FILE [Lucene 2.2]";
else if (format == SegmentInfos.FORMAT_SHARED_DOC_STORE)
sFormat = "FORMAT_SHARED_DOC_STORE [Lucene 2.3]";
else
{
if (format == SegmentInfos.FORMAT_CHECKSUM)
sFormat = "FORMAT_CHECKSUM [Lucene 2.4]";
else if (format == SegmentInfos.FORMAT_DEL_COUNT)
sFormat = "FORMAT_DEL_COUNT [Lucene 2.4]";
else if (format == SegmentInfos.FORMAT_HAS_PROX)
sFormat = "FORMAT_HAS_PROX [Lucene 2.4]";
else if (format == SegmentInfos.FORMAT_USER_DATA)
sFormat = "FORMAT_USER_DATA [Lucene 2.9]";
else if (format == SegmentInfos.FORMAT_DIAGNOSTICS)
sFormat = "FORMAT_DIAGNOSTICS [Lucene 2.9]";
else if (format < SegmentInfos.CURRENT_FORMAT)
{
sFormat = "int=" + format + " [newer version of Lucene than this tool]";
skip = true;
}
else
{
sFormat = format + " [Lucene 1.3 or prior]";
}
}
result.segmentsFileName = segmentsFileName;
result.numSegments = numSegments;
result.segmentFormat = sFormat;
result.userData = sis.UserData;
System.String userDataString;
if (sis.UserData.Count > 0)
{
userDataString = " userData=" + CollectionsHelper.CollectionToString(sis.UserData);
}
else
{
userDataString = "";
}
Msg("Segments file=" + segmentsFileName + " numSegments=" + numSegments + " version=" + sFormat + userDataString);
if (onlySegments != null)
{
result.partial = true;
if (infoStream != null)
infoStream.Write("\nChecking only these segments:");
foreach(string s in onlySegments)
{
if (infoStream != null)
{
infoStream.Write(" " + s);
}
}
result.segmentsChecked.AddRange(onlySegments);
Msg(":");
}
if (skip)
{
Msg("\nERROR: this index appears to be created by a newer version of Lucene than this tool was compiled on; please re-compile this tool on the matching version of Lucene; exiting");
result.toolOutOfDate = true;
return result;
}
result.newSegments = (SegmentInfos) sis.Clone();
result.newSegments.Clear();
for (int i = 0; i < numSegments; i++)
{
SegmentInfo info = sis.Info(i);
if (onlySegments != null && !onlySegments.Contains(info.name))
continue;
var segInfoStat = new Status.SegmentInfoStatus();
result.segmentInfos.Add(segInfoStat);
Msg(" " + (1 + i) + " of " + numSegments + ": name=" + info.name + " docCount=" + info.docCount);
segInfoStat.name = info.name;
segInfoStat.docCount = info.docCount;
int toLoseDocCount = info.docCount;
SegmentReader reader = null;
try
{
Msg(" compound=" + info.GetUseCompoundFile());
segInfoStat.compound = info.GetUseCompoundFile();
Msg(" hasProx=" + info.HasProx);
segInfoStat.hasProx = info.HasProx;
Msg(" numFiles=" + info.Files().Count);
segInfoStat.numFiles = info.Files().Count;
Msg(System.String.Format(nf, " size (MB)={0:f}", new System.Object[] { (info.SizeInBytes() / (1024.0 * 1024.0)) }));
segInfoStat.sizeMB = info.SizeInBytes() / (1024.0 * 1024.0);
IDictionary<string, string> diagnostics = info.Diagnostics;
segInfoStat.diagnostics = diagnostics;
if (diagnostics.Count > 0)
{
Msg(" diagnostics = " + CollectionsHelper.CollectionToString(diagnostics));
}
int docStoreOffset = info.DocStoreOffset;
if (docStoreOffset != - 1)
{
Msg(" docStoreOffset=" + docStoreOffset);
segInfoStat.docStoreOffset = docStoreOffset;
Msg(" docStoreSegment=" + info.DocStoreSegment);
segInfoStat.docStoreSegment = info.DocStoreSegment;
Msg(" docStoreIsCompoundFile=" + info.DocStoreIsCompoundFile);
segInfoStat.docStoreCompoundFile = info.DocStoreIsCompoundFile;
}
System.String delFileName = info.GetDelFileName();
if (delFileName == null)
{
Msg(" no deletions");
segInfoStat.hasDeletions = false;
}
else
{
Msg(" has deletions [delFileName=" + delFileName + "]");
segInfoStat.hasDeletions = true;
segInfoStat.deletionsFileName = delFileName;
}
if (infoStream != null)
infoStream.Write(" test: open reader.........");
reader = SegmentReader.Get(true, info, IndexReader.DEFAULT_TERMS_INDEX_DIVISOR);
segInfoStat.openReaderPassed = true;
int numDocs = reader.NumDocs();
toLoseDocCount = numDocs;
if (reader.HasDeletions)
{
if (reader.deletedDocs.Count() != info.GetDelCount())
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs deletedDocs.count()=" + reader.deletedDocs.Count());
}
if (reader.deletedDocs.Count() > reader.MaxDoc)
{
throw new System.SystemException("too many deleted docs: MaxDoc=" + reader.MaxDoc + " vs deletedDocs.count()=" + reader.deletedDocs.Count());
}
if (info.docCount - numDocs != info.GetDelCount())
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs reader=" + (info.docCount - numDocs));
}
segInfoStat.numDeleted = info.docCount - numDocs;
Msg("OK [" + (segInfoStat.numDeleted) + " deleted docs]");
}
else
{
if (info.GetDelCount() != 0)
{
throw new System.SystemException("delete count mismatch: info=" + info.GetDelCount() + " vs reader=" + (info.docCount - numDocs));
}
Msg("OK");
}
if (reader.MaxDoc != info.docCount)
throw new System.SystemException("SegmentReader.MaxDoc " + reader.MaxDoc + " != SegmentInfos.docCount " + info.docCount);
// Test getFieldNames()
if (infoStream != null)
{
infoStream.Write(" test: fields..............");
}
ICollection<string> fieldNames = reader.GetFieldNames(IndexReader.FieldOption.ALL);
Msg("OK [" + fieldNames.Count + " fields]");
segInfoStat.numFields = fieldNames.Count;
// Test Field Norms
segInfoStat.fieldNormStatus = TestFieldNorms(fieldNames, reader);
// Test the Term Index
segInfoStat.termIndexStatus = TestTermIndex(info, reader);
// Test Stored Fields
segInfoStat.storedFieldStatus = TestStoredFields(info, reader, nf);
// Test Term Vectors
segInfoStat.termVectorStatus = TestTermVectors(info, reader, nf);
// Rethrow the first exception we encountered
// This will cause stats for failed segments to be incremented properly
if (segInfoStat.fieldNormStatus.error != null)
{
throw new SystemException("Field Norm test failed");
}
else if (segInfoStat.termIndexStatus.error != null)
{
throw new SystemException("Term Index test failed");
}
else if (segInfoStat.storedFieldStatus.error != null)
{
throw new SystemException("Stored Field test failed");
}
else if (segInfoStat.termVectorStatus.error != null)
{
throw new System.SystemException("Term Vector test failed");
}
Msg("");
}
catch (System.Exception t)
{
Msg("FAILED");
const string comment = "fixIndex() would remove reference to this segment";
Msg(" WARNING: " + comment + "; full exception:");
if (infoStream != null)
infoStream.WriteLine(t.StackTrace);
Msg("");
result.totLoseDocCount += toLoseDocCount;
result.numBadSegments++;
continue;
}
finally
{
if (reader != null)
reader.Close();
}
// Keeper
result.newSegments.Add((SegmentInfo)info.Clone());
}
if (0 == result.numBadSegments)
{
result.clean = true;
Msg("No problems were detected with this index.\n");
}
else
Msg("WARNING: " + result.numBadSegments + " broken segments (containing " + result.totLoseDocCount + " documents) detected");
return result;
}
internal ReaderCommit(SegmentInfos infos, Directory dir)
{
segmentsFileName = infos.GetCurrentSegmentFileName();
this.dir = dir;
userData = infos.UserData;
files = infos.Files(dir, true);
version = infos.Version;
generation = infos.Generation;
isOptimized = infos.Count == 1 && !infos.Info(0).HasDeletions();
}