/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.cassandra.io.sstable;
import java.io.*;
import java.nio.ByteBuffer;
import java.nio.charset.Charset;
import java.util.*;
import java.util.concurrent.CopyOnWriteArraySet;
import com.google.common.base.Predicates;
import com.google.common.collect.Collections2;
import com.google.common.collect.Sets;
import com.google.common.io.Files;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.apache.cassandra.db.BufferDecoratedKey;
import org.apache.cassandra.db.DecoratedKey;
import org.apache.cassandra.db.RowIndexEntry;
import org.apache.cassandra.dht.IPartitioner;
import org.apache.cassandra.io.FSWriteError;
import org.apache.cassandra.io.util.DiskOptimizationStrategy;
import org.apache.cassandra.io.util.FileUtils;
import org.apache.cassandra.io.util.RandomAccessReader;
import org.apache.cassandra.schema.TableMetadata;
import org.apache.cassandra.schema.TableMetadataRef;
import org.apache.cassandra.utils.ByteBufferUtil;
import org.apache.cassandra.utils.Pair;
import org.apache.cassandra.utils.memory.HeapAllocator;
/**
* This class is built on top of the SequenceFile. It stores
* data on disk in sorted fashion. However the sorting is upto
* the application. This class expects keys to be handed to it
* in sorted order.
*
* A separate index file is maintained as well, containing the
* SSTable keys and the offset into the SSTable at which they are found.
* Every 1/indexInterval key is read into memory when the SSTable is opened.
*
* Finally, a bloom filter file is also kept for the keys in each SSTable.
*/
public abstract class SSTable
{
static final Logger logger = LoggerFactory.getLogger(SSTable.class);
public static final int TOMBSTONE_HISTOGRAM_BIN_SIZE = 100;
public static final int TOMBSTONE_HISTOGRAM_SPOOL_SIZE = 100000;
public static final int TOMBSTONE_HISTOGRAM_TTL_ROUND_SECONDS = Integer.valueOf(System.getProperty("cassandra.streaminghistogram.roundseconds", "60"));
public final Descriptor descriptor;
protected final Set<Component> components;
public final boolean compression;
public DecoratedKey first;
public DecoratedKey last;
protected final DiskOptimizationStrategy optimizationStrategy;
protected final TableMetadataRef metadata;
protected SSTable(Descriptor descriptor, Set<Component> components, TableMetadataRef metadata, DiskOptimizationStrategy optimizationStrategy)
{
// In almost all cases, metadata shouldn't be null, but allowing null allows to create a mostly functional SSTable without
// full schema definition. SSTableLoader use that ability
assert descriptor != null;
assert components != null;
this.descriptor = descriptor;
Set<Component> dataComponents = new HashSet<>(components);
this.compression = dataComponents.contains(Component.COMPRESSION_INFO);
this.components = new CopyOnWriteArraySet<>(dataComponents);
this.metadata = metadata;
this.optimizationStrategy = Objects.requireNonNull(optimizationStrategy);
}
/**
* We use a ReferenceQueue to manage deleting files that have been compacted
* and for which no more SSTable references exist. But this is not guaranteed
* to run for each such file because of the semantics of the JVM gc. So,
* we write a marker to `compactedFilename` when a file is compacted;
* if such a marker exists on startup, the file should be removed.
*
* This method will also remove SSTables that are marked as temporary.
*
* @return true if the file was deleted
*/
public static boolean delete(Descriptor desc, Set<Component> components)
{
// remove the DATA component first if it exists
if (components.contains(Component.DATA))
FileUtils.deleteWithConfirm(desc.filenameFor(Component.DATA));
for (Component component : components)
{
if (component.equals(Component.DATA) || component.equals(Component.SUMMARY))
continue;
FileUtils.deleteWithConfirm(desc.filenameFor(component));
}
if (components.contains(Component.SUMMARY))
FileUtils.delete(desc.filenameFor(Component.SUMMARY));
logger.trace("Deleted {}", desc);
return true;
}
public TableMetadata metadata()
{
return metadata.get();
}
public IPartitioner getPartitioner()
{
return metadata().partitioner;
}
public DecoratedKey decorateKey(ByteBuffer key)
{
return getPartitioner().decorateKey(key);
}
/**
* If the given @param key occupies only part of a larger buffer, allocate a new buffer that is only
* as large as necessary.
*/
public static DecoratedKey getMinimalKey(DecoratedKey key)
{
return key.getKey().position() > 0 || key.getKey().hasRemaining() || !key.getKey().hasArray()
? new BufferDecoratedKey(key.getToken(), HeapAllocator.instance.clone(key.getKey()))
: key;
}
public String getFilename()
{
return descriptor.filenameFor(Component.DATA);
}
public String getIndexFilename()
{
return descriptor.filenameFor(Component.PRIMARY_INDEX);
}
public String getColumnFamilyName()
{
return descriptor.cfname;
}
public String getKeyspaceName()
{
return descriptor.ksname;
}
public List<String> getAllFilePaths()
{
List<String> ret = new ArrayList<>();
for (Component component : components)
ret.add(descriptor.filenameFor(component));
return ret;
}
/**
* Parse a sstable filename into both a {@link Descriptor} and {@code Component} object.
*
* @param file the filename to parse.
* @return a pair of the {@code Descriptor} and {@code Component} corresponding to {@code file} if it corresponds to
* a valid and supported sstable filename, {@code null} otherwise. Note that components of an unknown type will be
* returned as CUSTOM ones.
*/
public static Pair<Descriptor, Component> tryComponentFromFilename(File file)
{
try
{
return Descriptor.fromFilenameWithComponent(file);
}
catch (Throwable e)
{
return null;
}
}
/**
* Parse a sstable filename into a {@link Descriptor} object.
* <p>
* Note that this method ignores the component part of the filename; if this is not what you want, use
* {@link #tryComponentFromFilename} instead.
*
* @param file the filename to parse.
* @return the {@code Descriptor} corresponding to {@code file} if it corresponds to a valid and supported sstable
* filename, {@code null} otherwise.
*/
public static Descriptor tryDescriptorFromFilename(File file)
{
try
{
return Descriptor.fromFilename(file);
}
catch (Throwable e)
{
return null;
}
}
/**
* Discovers existing components for the descriptor. Slow: only intended for use outside the critical path.
*/
public static Set<Component> componentsFor(final Descriptor desc)
{
try
{
try
{
return readTOC(desc);
}
catch (FileNotFoundException e)
{
Set<Component> components = discoverComponentsFor(desc);
if (components.isEmpty())
return components; // sstable doesn't exist yet
if (!components.contains(Component.TOC))
components.add(Component.TOC);
appendTOC(desc, components);
return components;
}
}
catch (IOException e)
{
throw new IOError(e);
}
}
public static Set<Component> discoverComponentsFor(Descriptor desc)
{
Set<Component.Type> knownTypes = Sets.difference(Component.TYPES, Collections.singleton(Component.Type.CUSTOM));
Set<Component> components = Sets.newHashSetWithExpectedSize(knownTypes.size());
for (Component.Type componentType : knownTypes)
{
Component component = new Component(componentType);
if (new File(desc.filenameFor(component)).exists())
components.add(component);
}
return components;
}
/** @return An estimate of the number of keys contained in the given index file. */
protected long estimateRowsFromIndex(RandomAccessReader ifile) throws IOException
{
// collect sizes for the first 10000 keys, or first 10 megabytes of data
final int SAMPLES_CAP = 10000, BYTES_CAP = (int)Math.min(10000000, ifile.length());
int keys = 0;
while (ifile.getFilePointer() < BYTES_CAP && keys < SAMPLES_CAP)
{
ByteBufferUtil.skipShortLength(ifile);
RowIndexEntry.Serializer.skip(ifile, descriptor.version);
keys++;
}
assert keys > 0 && ifile.getFilePointer() > 0 && ifile.length() > 0 : "Unexpected empty index file: " + ifile;
long estimatedRows = ifile.length() / (ifile.getFilePointer() / keys);
ifile.seek(0);
return estimatedRows;
}
public long bytesOnDisk()
{
long bytes = 0;
for (Component component : components)
{
bytes += new File(descriptor.filenameFor(component)).length();
}
return bytes;
}
@Override
public String toString()
{
return getClass().getSimpleName() + "(" +
"path='" + getFilename() + '\'' +
')';
}
/**
* Reads the list of components from the TOC component.
* @return set of components found in the TOC
*/
protected static Set<Component> readTOC(Descriptor descriptor) throws IOException
{
File tocFile = new File(descriptor.filenameFor(Component.TOC));
List<String> componentNames = Files.readLines(tocFile, Charset.defaultCharset());
Set<Component> components = Sets.newHashSetWithExpectedSize(componentNames.size());
for (String componentName : componentNames)
{
Component component = new Component(Component.Type.fromRepresentation(componentName), componentName);
if (!new File(descriptor.filenameFor(component)).exists())
logger.error("Missing component: {}", descriptor.filenameFor(component));
else
components.add(component);
}
return components;
}
/**
* Appends new component names to the TOC component.
*/
protected static void appendTOC(Descriptor descriptor, Collection<Component> components)
{
File tocFile = new File(descriptor.filenameFor(Component.TOC));
try (PrintWriter w = new PrintWriter(new FileWriter(tocFile, true)))
{
for (Component component : components)
w.println(component.name);
}
catch (IOException e)
{
throw new FSWriteError(e, tocFile);
}
}
/**
* Registers new custom components. Used by custom compaction strategies.
* Adding a component for the second time is a no-op.
* Don't remove this - this method is a part of the public API, intended for use by custom compaction strategies.
* @param newComponents collection of components to be added
*/
public synchronized void addComponents(Collection<Component> newComponents)
{
Collection<Component> componentsToAdd = Collections2.filter(newComponents, Predicates.not(Predicates.in(components)));
appendTOC(descriptor, componentsToAdd);
components.addAll(componentsToAdd);
}
}