punch.py

#!/usr/bin/python

import ctypes
import ctypes.util

c_off_t = ctypes.c_int64

def make_fallocate():
    libc_name = ctypes.util.find_library('c')
    libc = ctypes.CDLL(libc_name)

    _fallocate = libc.fallocate
    _fallocate.restype = ctypes.c_int
    _fallocate.argtypes = [ctypes.c_int, ctypes.c_int, c_off_t, c_off_t]

    del libc
    del libc_name

    def fallocate(fd, mode, offset, len_):
        res = _fallocate(fd.fileno(), mode, offset, len_)
        if res != 0:
            raise IOError(res, 'fallocate')

    return fallocate

fallocate = make_fallocate()
del make_fallocate

FALLOC_FL_KEEP_SIZE = 0x01
FALLOC_FL_PUNCH_HOLE = 0x02

def sizeof_fmt(num, suffix='B'):
    for unit in ['','Ki','Mi','Gi','Ti','Pi','Ei','Zi']:
        if abs(num) < 1024.0:
            return "%3.0f %s%s" % (num, unit, suffix)
        num /= 1024.0
    return "%.1f%s%s" % (num, 'Yi', suffix)
    
def punch(filename, verbose, dry_run, print_summary):
    blocksize = 4096
    saved_space = 0
    extents = 0
    if verbose:
        print "processing", filename
    with open(filename, 'r+') as f:
        offset = 0
        length = 0
        last_block_was_hole = False
        while True:
            buf = f.read(blocksize)
            if not buf:
                break
            if buf == "\x00" * blocksize:
                if verbose:
                    print "punching hole at offset", offset, "length", len(buf)
                if not dry_run:
                    fallocate(f, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE,
                          offset, len(buf))
                saved_space += len(buf)
                if not last_block_was_hole:
                    extents += 1
                last_block_was_hole = True
            else:
                last_block_was_hole = False
            offset = offset + blocksize
        if print_summary:
            print "Saved space: %s" % sizeof_fmt(saved_space)
            print "Extents created: %s" % extents


if __name__ == '__main__':
    import sys
    import argparse
    parser = argparse.ArgumentParser(
        description = "Punch out the empty areas in a file, making it sparse")
    parser.add_argument('file', metavar='FILE',
                        help='file(s) to modify in-place', nargs='+')
    parser.add_argument('-s', '--summary', action="store_true", default=False,
                       help="print summary of extents created and space saved")
    parser.add_argument('-n', '--dry-run', action="store_true", default=False,
                        help="do not actually modify the file")
    parser.add_argument('-v', '--verbose', action="store_true", default=False,
                        help='be verbose')
    args = parser.parse_args()
    for filename in args.file:
        punch(filename, args.verbose, args.dry_run, args.summary)