| public static GenerationFromSegmentsFileName ( string fileName ) : long | ||
| fileName | string | |
| return | long | |
/** @see IndexReader#listCommits */
public new static ICollection <IndexCommitPoint> ListCommits(Directory dir)
{
string[] files = dir.List();
if (files == null)
{
throw new System.IO.IOException("cannot read directory " + dir + ": list() returned null");
}
ICollection <IndexCommitPoint> commits = new List <IndexCommitPoint>();
SegmentInfos latest = new SegmentInfos();
latest.Read(dir);
long currentGen = latest.GetGeneration();
commits.Add(new ReaderCommit(latest, dir));
for (int i = 0; i < files.Length; i++)
{
String fileName = files[i];
if (fileName.StartsWith(IndexFileNames.SEGMENTS) &&
!fileName.Equals(IndexFileNames.SEGMENTS_GEN) &&
SegmentInfos.GenerationFromSegmentsFileName(fileName) < currentGen)
{
SegmentInfos sis = new SegmentInfos();
try
{
// IOException allowed to throw there, in case
// segments_N is corrupt
sis.Read(dir, fileName);
}
catch (System.Exception)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
sis = null;
}
if (sis != null)
{
commits.Add(new ReaderCommit(sis, dir));
}
}
}
return(commits);
}
private void VerifyCommitOrder <T>(IList <T> commits)
where T : IndexCommit
{
if (commits.Count == 0)
{
return;
}
IndexCommit firstCommit = commits[0];
long last = SegmentInfos.GenerationFromSegmentsFileName(firstCommit.SegmentsFileName);
Assert.AreEqual(last, firstCommit.Generation);
for (int i = 1; i < commits.Count; i++)
{
IndexCommit commit = commits[i];
long now = SegmentInfos.GenerationFromSegmentsFileName(commit.SegmentsFileName);
Assert.IsTrue(now > last, "SegmentInfos commits are out-of-order");
Assert.AreEqual(now, commit.Generation);
last = now;
}
}
private void VerifyCommitOrder(System.Collections.IList commits)
{
IndexCommit firstCommit = ((IndexCommit)commits[0]);
long last = SegmentInfos.GenerationFromSegmentsFileName(firstCommit.GetSegmentsFileName());
Assert.AreEqual(last, firstCommit.GetGeneration());
long lastVersion = firstCommit.GetVersion();
long lastTimestamp = firstCommit.GetTimestamp();
for (int i = 1; i < commits.Count; i++)
{
IndexCommit commit = ((IndexCommit)commits[i]);
long now = SegmentInfos.GenerationFromSegmentsFileName(commit.GetSegmentsFileName());
long nowVersion = commit.GetVersion();
long nowTimestamp = commit.GetTimestamp();
Assert.IsTrue(now > last, "SegmentInfos commits are out-of-order");
Assert.IsTrue(nowVersion > lastVersion, "SegmentInfos versions are out-of-order");
Assert.IsTrue(nowTimestamp >= lastTimestamp, "SegmentInfos timestamps are out-of-order: now=" + nowTimestamp + " vs last=" + lastTimestamp);
Assert.AreEqual(now, commit.GetGeneration());
last = now;
lastVersion = nowVersion;
lastTimestamp = nowTimestamp;
}
}
private void VerifyCommitOrder <T>(IList <T> commits) where T : IndexCommit
{
IndexCommit firstCommit = commits[0];
long last = SegmentInfos.GenerationFromSegmentsFileName(firstCommit.SegmentsFileName);
Assert.AreEqual(last, firstCommit.Generation);
long lastVersion = firstCommit.Version;
long lastTimestamp = firstCommit.Timestamp;
for (int i = 1; i < commits.Count; i++)
{
IndexCommit commit = commits[i];
long now = SegmentInfos.GenerationFromSegmentsFileName(commit.SegmentsFileName);
long nowVersion = commit.Version;
long nowTimestamp = commit.Timestamp;
Assert.IsTrue(now > last, "SegmentInfos commits are out-of-order");
Assert.IsTrue(nowVersion > lastVersion, "SegmentInfos versions are out-of-order");
Assert.IsTrue(nowTimestamp >= lastTimestamp, "SegmentInfos timestamps are out-of-order: now=" + nowTimestamp + " vs last=" + lastTimestamp);
Assert.AreEqual(now, commit.Generation);
last = now;
lastVersion = nowVersion;
lastTimestamp = nowTimestamp;
}
}
/// <summary>
/// Initialize the deleter: find all previous commits in
/// the <see cref="Directory"/>, incref the files they reference, call
/// the policy to let it delete commits. this will remove
/// any files not referenced by any of the commits. </summary>
/// <exception cref="IOException"> if there is a low-level IO error </exception>
public IndexFileDeleter(Directory directory, IndexDeletionPolicy policy, SegmentInfos segmentInfos, InfoStream infoStream, IndexWriter writer, bool initialIndexExists)
{
this.infoStream = infoStream;
this.writer = writer;
string currentSegmentsFile = segmentInfos.GetSegmentsFileName();
if (infoStream.IsEnabled("IFD"))
{
infoStream.Message("IFD", "init: current segments file is \"" + currentSegmentsFile + "\"; deletionPolicy=" + policy);
}
this.policy = policy;
this.directory = directory;
// First pass: walk the files and initialize our ref
// counts:
long currentGen = segmentInfos.Generation;
CommitPoint currentCommitPoint = null;
string[] files = null;
try
{
files = directory.ListAll();
}
#pragma warning disable 168
catch (DirectoryNotFoundException e)
#pragma warning restore 168
{
// it means the directory is empty, so ignore it.
files = new string[0];
}
if (currentSegmentsFile != null)
{
Regex r = IndexFileNames.CODEC_FILE_PATTERN;
foreach (string fileName in files)
{
if (!fileName.EndsWith("write.lock", StringComparison.Ordinal) && !fileName.Equals(IndexFileNames.SEGMENTS_GEN, StringComparison.Ordinal) &&
(r.IsMatch(fileName) || fileName.StartsWith(IndexFileNames.SEGMENTS, StringComparison.Ordinal)))
{
// Add this file to refCounts with initial count 0:
GetRefCount(fileName);
if (fileName.StartsWith(IndexFileNames.SEGMENTS, StringComparison.Ordinal))
{
// this is a commit (segments or segments_N), and
// it's valid (<= the max gen). Load it, then
// incref all files it refers to:
if (infoStream.IsEnabled("IFD"))
{
infoStream.Message("IFD", "init: load commit \"" + fileName + "\"");
}
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, fileName);
}
#pragma warning disable 168
catch (FileNotFoundException e)
#pragma warning restore 168
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
if (infoStream.IsEnabled("IFD"))
{
infoStream.Message("IFD", "init: hit FileNotFoundException when loading commit \"" + fileName + "\"; skipping this commit point");
}
sis = null;
}
// LUCENENET specific - .NET (thankfully) only has one FileNotFoundException, so we don't need this
//catch (NoSuchFileException)
//{
// // LUCENE-948: on NFS (and maybe others), if
// // you have writers switching back and forth
// // between machines, it's very likely that the
// // dir listing will be stale and will claim a
// // file segments_X exists when in fact it
// // doesn't. So, we catch this and handle it
// // as if the file does not exist
// if (infoStream.IsEnabled("IFD"))
// {
// infoStream.Message("IFD", "init: hit FileNotFoundException when loading commit \"" + fileName + "\"; skipping this commit point");
// }
// sis = null;
//}
// LUCENENET specific - since NoSuchDirectoryException subclasses FileNotFoundException
// in Lucene, we need to catch it here to be on the safe side.
catch (System.IO.DirectoryNotFoundException)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
if (infoStream.IsEnabled("IFD"))
{
infoStream.Message("IFD", "init: hit FileNotFoundException when loading commit \"" + fileName + "\"; skipping this commit point");
}
sis = null;
}
catch (IOException /*e*/)
{
if (SegmentInfos.GenerationFromSegmentsFileName(fileName) <= currentGen && directory.FileLength(fileName) > 0)
{
throw; // LUCENENET: CA2200: Rethrow to preserve stack details (https://docs.microsoft.com/en-us/visualstudio/code-quality/ca2200-rethrow-to-preserve-stack-details)
}
else
{
// Most likely we are opening an index that
// has an aborted "future" commit, so suppress
// exc in this case
sis = null;
}
}
if (sis != null)
{
CommitPoint commitPoint = new CommitPoint(commitsToDelete, directory, sis);
if (sis.Generation == segmentInfos.Generation)
{
currentCommitPoint = commitPoint;
}
commits.Add(commitPoint);
IncRef(sis, true);
if (lastSegmentInfos == null || sis.Generation > lastSegmentInfos.Generation)
{
lastSegmentInfos = sis;
}
}
}
}
}
}
if (currentCommitPoint == null && currentSegmentsFile != null && initialIndexExists)
{
// We did not in fact see the segments_N file
// corresponding to the segmentInfos that was passed
// in. Yet, it must exist, because our caller holds
// the write lock. this can happen when the directory
// listing was stale (eg when index accessed via NFS
// client with stale directory listing cache). So we
// try now to explicitly open this commit point:
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, currentSegmentsFile);
}
catch (IOException e)
{
throw new CorruptIndexException("failed to locate current segments_N file \"" + currentSegmentsFile + "\"" + e.ToString(), e);
}
if (infoStream.IsEnabled("IFD"))
{
infoStream.Message("IFD", "forced open of current segments file " + segmentInfos.GetSegmentsFileName());
}
currentCommitPoint = new CommitPoint(commitsToDelete, directory, sis);
commits.Add(currentCommitPoint);
IncRef(sis, true);
}
// We keep commits list in sorted order (oldest to newest):
CollectionUtil.TimSort(commits);
// Now delete anything with ref count at 0. These are
// presumably abandoned files eg due to crash of
// IndexWriter.
foreach (KeyValuePair <string, RefCount> entry in refCounts)
{
RefCount rc = entry.Value;
string fileName = entry.Key;
if (0 == rc.count)
{
if (infoStream.IsEnabled("IFD"))
{
infoStream.Message("IFD", "init: removing unreferenced file \"" + fileName + "\"");
}
DeleteFile(fileName);
}
}
// Finally, give policy a chance to remove things on
// startup:
this.policy.OnInit(commits);
// Always protect the incoming segmentInfos since
// sometime it may not be the most recent commit
Checkpoint(segmentInfos, false);
startingCommitDeleted = currentCommitPoint == null ? false : currentCommitPoint.IsDeleted;
DeleteCommits();
}
/// <summary> Initialize the deleter: find all previous commits in
/// the Directory, incref the files they reference, call
/// the policy to let it delete commits. The incoming
/// segmentInfos must have been loaded from a commit point
/// and not yet modified. This will remove any files not
/// referenced by any of the commits.
/// </summary>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public IndexFileDeleter(Directory directory, IndexDeletionPolicy policy, SegmentInfos segmentInfos, System.IO.TextWriter infoStream, DocumentsWriter docWriter)
{
this.docWriter = docWriter;
this.infoStream = infoStream;
if (infoStream != null)
{
Message("init: current segments file is \"" + segmentInfos.GetCurrentSegmentFileName() + "\"; deletionPolicy=" + policy);
}
this.policy = policy;
this.directory = directory;
// First pass: walk the files and initialize our ref
// counts:
long currentGen = segmentInfos.GetGeneration();
IndexFileNameFilter filter = IndexFileNameFilter.GetFilter();
System.String[] files = directory.List();
if (files == null)
{
throw new System.IO.IOException("cannot read directory " + directory + ": list() returned null");
}
CommitPoint currentCommitPoint = null;
for (int i = 0; i < files.Length; i++)
{
System.String fileName = files[i];
if (filter.Accept(null, fileName) && !fileName.Equals(IndexFileNames.SEGMENTS_GEN))
{
// Add this file to refCounts with initial count 0:
GetRefCount(fileName);
if (fileName.StartsWith(IndexFileNames.SEGMENTS))
{
// This is a commit (segments or segments_N), and
// it's valid (<= the max gen). Load it, then
// incref all files it refers to:
if (SegmentInfos.GenerationFromSegmentsFileName(fileName) <= currentGen)
{
if (infoStream != null)
{
Message("init: load commit \"" + fileName + "\"");
}
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, fileName);
}
catch (System.IO.FileNotFoundException e)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
if (infoStream != null)
{
Message("init: hit FileNotFoundException when loading commit \"" + fileName + "\"; skipping this commit point");
}
sis = null;
}
if (sis != null)
{
CommitPoint commitPoint = new CommitPoint(this, sis);
if (sis.GetGeneration() == segmentInfos.GetGeneration())
{
currentCommitPoint = commitPoint;
}
commits.Add(commitPoint);
IncRef(sis, true);
}
}
}
}
}
if (currentCommitPoint == null)
{
// We did not in fact see the segments_N file
// corresponding to the segmentInfos that was passed
// in. Yet, it must exist, because our caller holds
// the write lock. This can happen when the directory
// listing was stale (eg when index accessed via NFS
// client with stale directory listing cache). So we
// try now to explicitly open this commit point:
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, segmentInfos.GetCurrentSegmentFileName());
}
catch (System.IO.IOException e)
{
throw new CorruptIndexException("failed to locate current segments_N file");
}
if (infoStream != null)
{
Message("forced open of current segments file " + segmentInfos.GetCurrentSegmentFileName());
}
currentCommitPoint = new CommitPoint(this, sis);
commits.Add(currentCommitPoint);
IncRef(sis, true);
}
// We keep commits list in sorted order (oldest to newest):
commits.Sort();
// Now delete anything with ref count at 0. These are
// presumably abandoned files eg due to crash of
// IndexWriter.
System.Collections.IEnumerator it = refCounts.Keys.GetEnumerator();
while (it.MoveNext())
{
System.String fileName = (System.String)it.Current;
RefCount rc = (RefCount)refCounts[fileName];
if (0 == rc.count)
{
if (infoStream != null)
{
Message("init: removing unreferenced file \"" + fileName + "\"");
}
DeleteFile(fileName);
}
}
// Finally, give policy a chance to remove things on
// startup:
policy.OnInit(commits);
// It's OK for the onInit to remove the current commit
// point; we just have to checkpoint our in-memory
// SegmentInfos to protect those files that it uses:
if (currentCommitPoint.deleted)
{
Checkpoint(segmentInfos, false);
}
DeleteCommits();
}
/// <summary>
/// Returns all commit points that exist in the <see cref="Store.Directory"/>.
/// Normally, because the default is
/// <see cref="KeepOnlyLastCommitDeletionPolicy"/>, there would be only
/// one commit point. But if you're using a custom
/// <see cref="IndexDeletionPolicy"/> then there could be many commits.
/// Once you have a given commit, you can open a reader on
/// it by calling <see cref="DirectoryReader.Open(IndexCommit)"/>
/// There must be at least one commit in
/// the <see cref="Store.Directory"/>, else this method throws
/// <see cref="IndexNotFoundException"/>. Note that if a commit is in
/// progress while this method is running, that commit
/// may or may not be returned.
/// </summary>
/// <returns> a sorted list of <see cref="Index.IndexCommit"/>s, from oldest
/// to latest. </returns>
public static IList <IndexCommit> ListCommits(Directory dir)
{
string[] files = dir.ListAll();
List <IndexCommit> commits = new List <IndexCommit>();
SegmentInfos latest = new SegmentInfos();
latest.Read(dir);
long currentGen = latest.Generation;
commits.Add(new StandardDirectoryReader.ReaderCommit(latest, dir));
for (int i = 0; i < files.Length; i++)
{
string fileName = files[i];
if (fileName.StartsWith(IndexFileNames.SEGMENTS, StringComparison.Ordinal) && !fileName.Equals(IndexFileNames.SEGMENTS_GEN, StringComparison.Ordinal) && SegmentInfos.GenerationFromSegmentsFileName(fileName) < currentGen)
{
SegmentInfos sis = new SegmentInfos();
try
{
// IOException allowed to throw there, in case
// segments_N is corrupt
sis.Read(dir, fileName);
}
catch (FileNotFoundException)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
sis = null;
}
// LUCENENET specific - .NET (thankfully) only has one FileNotFoundException, so we don't need this
//catch (NoSuchFileException)
//{
// sis = null;
//}
// LUCENENET specific - since NoSuchDirectoryException subclasses FileNotFoundException
// in Lucene, we need to catch it here to be on the safe side.
catch (System.IO.DirectoryNotFoundException)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
sis = null;
}
if (sis != null)
{
commits.Add(new StandardDirectoryReader.ReaderCommit(sis, dir));
}
}
}
// Ensure that the commit points are sorted in ascending order.
commits.Sort();
return(commits);
}
/// <summary> Initialize the deleter: find all previous commits in
/// the Directory, incref the files they reference, call
/// the policy to let it delete commits. This will remove
/// any files not referenced by any of the commits.
/// </summary>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public IndexFileDeleter(Directory directory, IndexDeletionPolicy policy, SegmentInfos segmentInfos, System.IO.StreamWriter infoStream, DocumentsWriter docWriter, HashSet <string> synced)
{
this.docWriter = docWriter;
this.infoStream = infoStream;
this.synced = synced;
if (infoStream != null)
{
Message("init: current segments file is \"" + segmentInfos.GetCurrentSegmentFileName() + "\"; deletionPolicy=" + policy);
}
this.policy = policy;
this.directory = directory;
// First pass: walk the files and initialize our ref
// counts:
long currentGen = segmentInfos.Generation;
IndexFileNameFilter filter = IndexFileNameFilter.Filter;
System.String[] files = directory.ListAll();
CommitPoint currentCommitPoint = null;
for (int i = 0; i < files.Length; i++)
{
System.String fileName = files[i];
if (filter.Accept(null, fileName) && !fileName.Equals(IndexFileNames.SEGMENTS_GEN))
{
// Add this file to refCounts with initial count 0:
GetRefCount(fileName);
if (fileName.StartsWith(IndexFileNames.SEGMENTS))
{
// This is a commit (segments or segments_N), and
// it's valid (<= the max gen). Load it, then
// incref all files it refers to:
if (infoStream != null)
{
Message("init: load commit \"" + fileName + "\"");
}
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, fileName);
}
catch (System.IO.FileNotFoundException)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
if (infoStream != null)
{
Message("init: hit FileNotFoundException when loading commit \"" + fileName + "\"; skipping this commit point");
}
sis = null;
}
catch (System.IO.IOException)
{
if (SegmentInfos.GenerationFromSegmentsFileName(fileName) <= currentGen)
{
throw;
}
else
{
// Most likely we are opening an index that
// has an aborted "future" commit, so suppress
// exc in this case
sis = null;
}
}
if (sis != null)
{
CommitPoint commitPoint = new CommitPoint(this, commitsToDelete, directory, sis);
if (sis.Generation == segmentInfos.Generation)
{
currentCommitPoint = commitPoint;
}
commits.Add(commitPoint);
IncRef(sis, true);
if (lastSegmentInfos == null || sis.Generation > lastSegmentInfos.Generation)
{
lastSegmentInfos = sis;
}
}
}
}
}
if (currentCommitPoint == null)
{
// We did not in fact see the segments_N file
// corresponding to the segmentInfos that was passed
// in. Yet, it must exist, because our caller holds
// the write lock. This can happen when the directory
// listing was stale (eg when index accessed via NFS
// client with stale directory listing cache). So we
// try now to explicitly open this commit point:
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, segmentInfos.GetCurrentSegmentFileName());
}
catch (System.IO.IOException)
{
throw new CorruptIndexException("failed to locate current segments_N file");
}
if (infoStream != null)
{
Message("forced open of current segments file " + segmentInfos.GetCurrentSegmentFileName());
}
currentCommitPoint = new CommitPoint(this, commitsToDelete, directory, sis);
commits.Add(currentCommitPoint);
IncRef(sis, true);
}
// We keep commits list in sorted order (oldest to newest):
commits.Sort();
// Now delete anything with ref count at 0. These are
// presumably abandoned files eg due to crash of
// IndexWriter.
foreach (KeyValuePair <string, RefCount> entry in refCounts)
{
string fileName = entry.Key;
RefCount rc = refCounts[fileName];
if (0 == rc.count)
{
if (infoStream != null)
{
Message("init: removing unreferenced file \"" + fileName + "\"");
}
DeleteFile(fileName);
}
}
// Finally, give policy a chance to remove things on
// startup:
policy.OnInit(commits);
// Always protect the incoming segmentInfos since
// sometime it may not be the most recent commit
Checkpoint(segmentInfos, false);
startingCommitDeleted = currentCommitPoint.IsDeleted;
DeleteCommits();
}
