1FDATASYNC(2) Linux Programmer's Manual FDATASYNC(2)
2
3
4
6 fdatasync - synchronize a file's in-core data with that on disk
7
9 #include <unistd.h>
10
11 int fdatasync(int fd);
12
14 fdatasync() flushes all data buffers of a file to disk (before the sys‐
15 tem call returns). It resembles fsync() but is not required to update
16 the metadata such as access time.
17
18 Applications that access databases or log files often write a tiny data
19 fragment (e.g., one line in a log file) and then call fsync() immedi‐
20 ately in order to ensure that the written data is physically stored on
21 the harddisk. Unfortunately, fsync() will always initiate two write
22 operations: one for the newly written data and another one in order to
23 update the modification time stored in the inode. If the modification
24 time is not a part of the transaction concept fdatasync() can be used
25 to avoid unnecessary inode disk write operations.
26
28 On success, zero is returned. On error, -1 is returned, and errno is
29 set appropriately.
30
32 EBADF fd is not a valid file descriptor open for writing.
33
34 EIO An error occurred during synchronization.
35
36 EROFS, EINVAL
37 fd is bound to a special file which does not support synchro‐
38 nization.
39
41 Currently (Linux 2.2) fdatasync() is equivalent to fsync().
42
44 On POSIX systems on which fdatasync() is available, _POSIX_SYNCHRO‐
45 NIZED_IO is defined in <unistd.h> to a value greater than 0. (See also
46 sysconf(3).)
47
49 POSIX.1-2001.
50
52 fsync(2), sync_file_range(2)
53 B.O. Gallmeister, POSIX.4, O'Reilly, pp. 220-223 and 343.
54
55
56
57Linux 1.3.86 1996-04-13 FDATASYNC(2)