Package org.apache.cassandra.io.sstable.format.big

Class RowIndexEntry

java.lang.Object

org.apache.cassandra.io.sstable.AbstractRowIndexEntry

org.apache.cassandra.io.sstable.format.big.RowIndexEntry

All Implemented Interfaces:

IMeasurableMemory

public class RowIndexEntry
extends AbstractRowIndexEntry

Binary format of RowIndexEntry is defined as follows: (long) position (64 bit long, vint encoded) (int) serialized size of data that follows (32 bit int, vint encoded) -- following for indexed entries only (so serialized size > 0) (int) int/Uint DeletionTime.localDeletionTime (uint if version >= "oa" (5.0)) (long) DeletionTime.markedForDeletionAt (int) number of IndexInfo objects (32 bit int, vint encoded) (*) serialized IndexInfo objects, see below (*) offsets of serialized IndexInfo objects, since version "ma" (3.0) Each IndexInfo object's offset is relative to the first IndexInfo object.

See IndexInfo for a description of the serialized format.

For each partition, the layout of the index file looks like this:

1. partition key - prefixed with short length
2. serialized RowIndexEntry objects

Generally, we distinguish between index entries that have index samples (list of IndexInfo objects) and those who don't. For each portion of data for a single partition in the data file, an index sample is created. The size of that portion is defined by Config.column_index_size.

Index entries with less than 2 index samples, will just store the position in the data file.

Note: legacy sstables for index entries are those sstable formats that do not have an offsets table to index samples (IndexInfo objects). These are those sstables created on Cassandra versions earlier than 3.0.

For index entries with index samples we store the index samples (IndexInfo objects). The bigger the partition, the more index samples are created. Since a huge amount of index samples will "pollute" the heap and cause huge GC pressure, Cassandra 3.6 (CASSANDRA-11206) distinguishes between index entries with an "acceptable" amount of index samples per partition and those with an "enormous" amount of index samples. The barrier is controlled by the configuration parameter Config.column_index_cache_size. Index entries with a total serialized size of index samples up to column_index_cache_size will be held in an array. Index entries exceeding that value will always be accessed from disk.

This results in these classes:

• RowIndexEntry just stores the offset in the data file.
• RowIndexEntry.IndexedEntry is for index entries with index samples and used for both current and legacy sstables, which do not exceed Config.column_index_cache_size.
• RowIndexEntry.ShallowIndexedEntry is for index entries with index samples that exceed Config.column_index_cache_size for sstables with an offset table to the index samples.

Since access to index samples on disk (obviously) requires some file reader, that functionality is encapsulated in implementations of RowIndexEntry.IndexInfoRetriever. There is an implementation to access index samples of legacy sstables (without the offsets table), an implementation of access sstables with an offsets table.

Until now (Cassandra 3.x), we still support reading from legacy sstables - i.e. sstables created by Cassandra < 3.0 (see BigFormat.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	RowIndexEntry.IndexInfoRetriever	Base class to access IndexInfo objects.
static interface	RowIndexEntry.IndexSerializer
static class	RowIndexEntry.RowIndexEntryReadSizeTooLargeException
static final class	RowIndexEntry.Serializer

Field Summary

Fields inherited from class org.apache.cassandra.io.sstable.AbstractRowIndexEntry

position

Constructor Summary

Constructors

Constructor	Description
RowIndexEntry(long position)

Method Summary

Modifier and Type	Method	Description
int	blockCount()
static RowIndexEntry	create(long dataFilePosition, long indexFilePosition, DeletionTime deletionTime, long headerLength, int columnIndexCount, int indexedPartSize, List<IndexInfo> indexSamples, int[] offsets, ISerializer<IndexInfo> idxInfoSerializer, Version version)
DeletionTime	deletionTime()
BigFormat	getSSTableFormat()
boolean	indexOnHeap()
boolean	isIndexed()
RowIndexEntry.IndexInfoRetriever	openWithIndex(FileHandle indexFile)
void	serialize(DataOutputPlus out, ByteBuffer indexInfo)
void	serializeForCache(DataOutputPlus out)	Serialize this entry for key cache
long	unsharedHeapSize()

Methods inherited from class org.apache.cassandra.io.sstable.AbstractRowIndexEntry

getPosition

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

RowIndexEntry

public RowIndexEntry(long position)

Method Details

isIndexed

public boolean isIndexed()

Overrides:

isIndexed in class AbstractRowIndexEntry

Returns:

true if this index entry contains the row-level tombstone and column summary. Otherwise, caller should fetch these from the row header.

indexOnHeap

public boolean indexOnHeap()

deletionTime

public DeletionTime deletionTime()

Overrides:

deletionTime in class AbstractRowIndexEntry

blockCount

public int blockCount()

Overrides:

blockCount in class AbstractRowIndexEntry

getSSTableFormat

public BigFormat getSSTableFormat()

Specified by:

getSSTableFormat in class AbstractRowIndexEntry

unsharedHeapSize

public long unsharedHeapSize()

Returns:

the amount of on-heap memory retained by the object that might be reclaimed if the object were reclaimed, i.e. it should try to exclude globally cached data where possible, or counting portions of arrays that are referenced by the object but used by other objects only (e.g. slabbed byte-buffers), etc.

create

public static RowIndexEntry create(long dataFilePosition,
 long indexFilePosition,
 DeletionTime deletionTime,
 long headerLength,
 int columnIndexCount,
 int indexedPartSize,
 List<IndexInfo> indexSamples,
 int[] offsets,
 ISerializer<IndexInfo> idxInfoSerializer,
 Version version)

Parameters:

dataFilePosition - position of the partition in the SSTableFormat.Components.DATA file

indexFilePosition - position in the BigFormat.Components.Types.PRIMARY_INDEX of the RowIndexEntry

deletionTime - deletion time of RowIndexEntry

headerLength - deletion time of RowIndexEntry

columnIndexCount - number of IndexInfo entries in the RowIndexEntry

indexedPartSize - serialized size of all serialized IndexInfo objects and their offsets

indexSamples - list with IndexInfo offsets (if total serialized size is less than Config.column_index_cache_size

offsets - offsets of IndexInfo offsets

idxInfoSerializer - the IndexInfo serializer

openWithIndex

public RowIndexEntry.IndexInfoRetriever openWithIndex(FileHandle indexFile)

serialize

public void serialize(DataOutputPlus out,
 ByteBuffer indexInfo)
               throws IOException

Throws:

IOException

serializeForCache

public void serializeForCache(DataOutputPlus out)
                       throws IOException

Description copied from class: AbstractRowIndexEntry

Serialize this entry for key cache

Specified by:

serializeForCache in class AbstractRowIndexEntry

Parameters:

out - the output stream for serialized entry

Throws:

IOException