org.apache.druid.java.util.common.io

Class NativeIO

java.lang.Object

org.apache.druid.java.util.common.io.NativeIO

public class NativeIO
extends Object

Native I/O operations in order to minimize cache impact.

Method Summary

Modifier and Type	Method and Description
static void	chunkedCopy(InputStream src, File dest) Copy from an input stream to a file minimizing cache impact on the destination..
static int	getfd(FileDescriptor descriptor) Get system file descriptor (int) from FileDescriptor object.
static void	trySkipCache(int fd, long offset, long len) Remove pages from the file system page cache when they wont be accessed again

Methods inherited from class java.lang.Object

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

Method Detail

getfd

public static int getfd(FileDescriptor descriptor)

Get system file descriptor (int) from FileDescriptor object.

Parameters:

descriptor - - FileDescriptor object to get fd from

Returns:

file descriptor, -1 or error

trySkipCache

public static void trySkipCache(int fd,
                                long offset,
                                long len)

Remove pages from the file system page cache when they wont be accessed again

Parameters:

fd - The file descriptor of the source file.

offset - The offset within the file.

len - The length to be flushed.

chunkedCopy

public static void chunkedCopy(InputStream src,
                               File dest)
                        throws IOException

Copy from an input stream to a file minimizing cache impact on the destination.. This happens chunk by chunk so only at most chunk size will be present in the OS page cache. Posix (Linux, BSD) only. The implementation in this method is based on a post by Linus Torvalds here: http://lkml.iu.edu/hypermail/linux/kernel/1005.2/01845.html

Parameters:

src - Source InputStream where to copy from

dest - Destination file to copy to

Throws:

IOException