Python 中的 file.flush() 与 os.fsync()

在Python 官方文档https://docs.python.org/2/library/stdtypes.html?highlight=file%20flush#file.flush 关于file.flush() 的说明中写道：“ flush() does not necessarily write the file’s data to disk. Use flush() followed by os.fsync() to ensure this behavior.”flush方法并不强制把数据写入到磁盘文件中，那么flush方法为什么会这样呢？
其实这里有两种层面的buffering：
1. Internal buffers
2. Operating system buffers
Internal buffers 创建在 runtime/library/language 中避免每次写操作都调用系统call，依次来加速运行。当需要把数据写入文件时，先写入internal buffer, 当internal buffer 溢满时，再调用系统call写入到文件中。
所以为了确保数据写入到了文件中，file.flush() 之后要调用os.fsync()。

Example

#!/usr/bin/python

import os, sys

# Open a file
fd = os.open( "foo.txt", os.O_RDWR|os.O_CREAT )

# Write one string
os.write(fd, "This is test")

# Now you can use fsync() method.
# Infact here you would not be able to see its effect.
os.fsync(fd)

# Now read this file from the beginning
os.lseek(fd, 0, 0)
str = os.read(fd, 100)
print "Read String is : ", str

# Close opened file
os.close( fd )

print "Closed the file successfully!!"

Reference:
http://stackoverflow.com/questions/7127075/
http://www.tutorialspoint.com/python/os_fsync.htm