FSYNC(2) Linux Programmer's Manual FSYNC(2)
fsync, fdatasync - synchronize a file's in-core state with storage device
int fsync(int fd);
int fdatasync(int fd);
Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
fsync(): _BSD_SOURCE || _XOPEN_SOURCE
|| /* since glibc 2.8: */ _POSIX_C_SOURCE >= 200112L
fdatasync(): _POSIX_C_SOURCE >= 199309L || _XOPEN_SOURCE >= 500
fsync() transfers ("flushes") all modified in-core data of (i.e., modified buffer cache pages for) the file referred to by the file
descriptor fd to the disk device (or other permanent storage device) where that file resides. The call blocks until the device reports
that the transfer has completed. It also flushes metadata information associated with the file (see stat(2)).
Calling fsync() does not necessarily ensure that the entry in the directory containing the file has also reached disk. For that an
explicit fsync() on a file descriptor for the directory is also needed.
fdatasync() is similar to fsync(), but does not flush modified metadata unless that metadata is needed in order to allow a subsequent data
retrieval to be correctly handled. For example, changes to st_atime or st_mtime (respectively, time of last access and time of last modi-
fication; see stat(2)) do not require flushing because they are not necessary for a subsequent data read to be handled correctly. On the
other hand, a change to the file size (st_size, as made by say ftruncate(2)), would require a metadata flush.
The aim of fdatasync() is to reduce disk activity for applications that do not require all metadata to be synchronized with the disk.
On success, these system calls return zero. On error, -1 is returned, and errno is set appropriately.
EBADF fd is not a valid file descriptor open for writing.
EIO An error occurred during synchronization.
fd is bound to a special file which does not support synchronization.
On POSIX systems on which fdatasync() is available, _POSIX_SYNCHRONIZED_IO is defined in <unistd.h> to a value greater than 0. (See also
Applications that access databases or log files often write a tiny data fragment (e.g., one line in a log file) and then call fsync() imme-
diately in order to ensure that the written data is physically stored on the harddisk. Unfortunately, fsync() will always initiate two
write operations: one for the newly written data and another one in order to update the modification time stored in the inode. If the mod-
ification time is not a part of the transaction concept fdatasync() can be used to avoid unnecessary inode disk write operations.
If the underlying hard disk has write caching enabled, then the data may not really be on permanent storage when fsync() / fdatasync()
When an ext2 file system is mounted with the sync option, directory entries are also implicitly synced by fsync().
On kernels before 2.4, fsync() on big files can be inefficient. An alternative might be to use the O_SYNC flag to open(2).
In Linux 2.2 and earlier, fdatasync() is equivalent to fsync(), and so has no performance advantage.
bdflush(2), open(2), sync(2), sync_file_range(2), hdparm(8), mount(8), sync(8), update(8)
This page is part of release 3.25 of the Linux man-pages project. A description of the project, and information about reporting bugs, can
be found at http://www.kernel.org/doc/man-pages/.
Linux 2008-11-07 FSYNC(2)