如何tar文件的公用压缩太大，导致内存不足崩溃？

Question

在下面的代码中，如果我给（Apache）Commons压缩一个几GB大小的文件，它将崩溃，因为它会占用我所有的内存。

FileOutputStream fileOutputStream = new FileOutputStream("output.tar");
BufferedOutputStream bufferedOutputStream = new BufferedOutputStream(fileOutputStream);
GzipCompressorOutputStream gzipOutputStream = new GzipCompressorOutputStream(bufferedOutputStream);
TarArchiveOutputStream tarArchiveOutputStream = new TarArchiveOutputStream(gzipOutputStream)) {

tarArchiveOutputStream.setBigNumberMode(TarArchiveOutputStream.BIGNUMBER_POSIX);
tarArchiveOutputStream.setLongFileMode(TarArchiveOutputStream.LONGFILE_GNU);

File currentFile = new File("Huge_MultiGB_File.txt");
String relativeFilePath = currentFile.getPath();
TarArchiveEntry tarEntry = new TarArchiveEntry(currentFile, relativeFilePath);
tarEntry.setSize(currentFile.length());
tarArchiveOutputStream.putArchiveEntry(tarEntry);
tarArchiveOutputStream.write(IOUtils.toByteArray(new FileInputStream(currentFile)));
tarArchiveOutputStream.closeArchiveEntry();

Answer 1

您必须写入文件的一小部分并在循环中写入输出，而不是首先使用IOUTILS将整个文件读入内存

它或多或少是这样做的：

FileInputStream source=new FileInputStream(....somefile);
tarArchiveOutputStream; prepared to w writing

byte[] buff = new byte[1024*10]; //10kb buff
int numBytesRead = -1; //number of bytes read


while(( numBytesRead = source.read(buff)) > 0 ) {
    // while source has bytes, read from source and write
    // the same number of bytes to the tar outputstream
    tarArchiveOutputStream.write(buff, 0, numBytesRead);
   }
}