Package org.elasticsearch.transport
Class ReuseBuffersLZ4BlockOutputStream
org.elasticsearch.transport.ReuseBuffersLZ4BlockOutputStream
- All Implemented Interfaces:
,
,
public class ReuseBuffersLZ4BlockOutputStream
extends
This file is forked from https://github.com/lz4/lz4-java. In particular it forks the following file
net.jpountz.lz4.LZ4BlockOutputStream.
It modifies the original lz4-java code to allow the reuse of local thread local byte arrays. This prevents
the need to allocate two new byte arrays everytime a new stream is created. For the Elasticsearch use case,
a single thread should fully compress the stream in one go to avoid memory corruption.
Streaming LZ4 (not compatible with the LZ4 Frame format).
This class compresses data into fixed-size blocks of compressed data.
This class uses its own format and is not compatible with the LZ4 Frame format.
For interoperability with other LZ4 tools, use
LZ4FrameOutputStream
,
which is compatible with the LZ4 Frame format. This class remains for backward compatibility.- See Also:
LZ4BlockInputStream
,LZ4FrameOutputStream
-
Field Summary
Fields inherited from class java.io.
-
Constructor Summary
ConstructorDescriptionout)
Creates a new instance which compresses into blocks of 64 KB.out, int blockSize)
Creates a new instance which compresses with the standard LZ4 compression algorithm.out, int blockSize, net.jpountz.lz4.LZ4Compressor compressor)
Creates a new instance which checks stream integrity usingStreamingXXHash32
and doesn't sync flush.out, int blockSize, net.jpountz.lz4.LZ4Compressor compressor, checksum, boolean syncFlush)
Creates a new with configurable block size. -
Method Summary
Methods inherited from class java.io.
Methods inherited from class java.lang.
, , , , , , , , ,
-
Constructor Details
-
ReuseBuffersLZ4BlockOutputStream
public ReuseBuffersLZ4BlockOutputStream( out, int blockSize, net.jpountz.lz4.LZ4Compressor compressor, checksum, boolean syncFlush)Creates a new with configurable block size. Large blocks require more memory at compression and decompression time but should improve the compression ratio.- Parameters:
out
- the to feedblockSize
- the maximum number of bytes to try to compress at once, must be >= 64 and <= 32 Mcompressor
- theLZ4Compressor
instance to use to compress datachecksum
- the instance to use to check data for integrity.syncFlush
- true if pending data should also be flushed onflush()
-
ReuseBuffersLZ4BlockOutputStream
public ReuseBuffersLZ4BlockOutputStream( out, int blockSize, net.jpountz.lz4.LZ4Compressor compressor)Creates a new instance which checks stream integrity usingStreamingXXHash32
and doesn't sync flush.- Parameters:
out
- the to feedblockSize
- the maximum number of bytes to try to compress at once, must be >= 64 and <= 32 Mcompressor
- theLZ4Compressor
instance to use to compress data- See Also:
ReuseBuffersLZ4BlockOutputStream(OutputStream, int, LZ4Compressor, Checksum, boolean)
,StreamingXXHash32.asChecksum()
-
ReuseBuffersLZ4BlockOutputStream
public ReuseBuffersLZ4BlockOutputStream( out, int blockSize)Creates a new instance which compresses with the standard LZ4 compression algorithm.- Parameters:
out
- the to feedblockSize
- the maximum number of bytes to try to compress at once, must be >= 64 and <= 32 M- See Also:
ReuseBuffersLZ4BlockOutputStream(OutputStream, int, LZ4Compressor)
,LZ4Factory.fastCompressor()
-
ReuseBuffersLZ4BlockOutputStream
public ReuseBuffersLZ4BlockOutputStream( out)Creates a new instance which compresses into blocks of 64 KB.- Parameters:
out
- the to feed- See Also:
ReuseBuffersLZ4BlockOutputStream(OutputStream, int)
-
-
Method Details
-
write
public void write(int b) throws- Overrides:
in class
- Throws:
-
write
public void write(byte[] b, int off, int len) throws- Overrides:
in class
- Throws:
-
write
public void write(byte[] b) throws- Overrides:
in class
- Throws:
-
close
public void close() throws- Specified by:
in interface
- Specified by:
in interface
- Overrides:
in class
- Throws:
-
flush
public void flush() throwsFlushes this compressed . If the stream has been created withsyncFlush=true
, pending data will be compressed and appended to the underlying before calling on the underlying stream. Otherwise, this method just flushes the underlying stream, so pending data might not be available for reading untilfinish()
orclose()
is called.- Specified by:
in interface
- Overrides:
in class
- Throws:
-
finish
public void finish() throwsSame asclose()
except that it doesn't close the underlying stream. This can be useful if you want to keep on using the underlying stream.- Throws:
- if an I/O error occurs.
-
toString
public toString()- Overrides:
in class
-