1FALLOCATE(1) User Commands FALLOCATE(1)
2
6 fallocate - preallocate or deallocate space to a file
7
9 fallocate [-c|-p|-z] [-o offset] -l length [-n] filename
10 fallocate -d [-o offset] [-l length] filename
12 fallocate -x [-o offset] -l length filename
14
16 fallocate is used to manipulate the allocated disk space for a file,
17 either to deallocate or preallocate it. For filesystems which support
18 the fallocate system call, preallocation is done quickly by allocating
19 blocks and marking them as uninitialized, requiring no IO to the data
20 blocks. This is much faster than creating a file by filling it with
21 zeroes.
22 The exit status returned by fallocate is 0 on success and 1 on failure.
24
26 The length and offset arguments may be followed by the multiplicative
27 suffixes KiB (=1024), MiB (=1024*1024), and so on for GiB, TiB, PiB,
28 EiB, ZiB, and YiB (the "iB" is optional, e.g., "K" has the same meaning
29 as "KiB") or the suffixes KB (=1000), MB (=1000*1000), and so on for
30 GB, TB, PB, EB, ZB, and YB.
31 The options --collapse-range, --dig-holes, --punch-hole, and
33 --zero-range are mutually exclusive.
34 -c, --collapse-range
36 Removes a byte range from a file, without leaving a hole. The
37 byte range to be collapsed starts at offset and continues for
38 length bytes. At the completion of the operation, the contents
39 of the file starting at the location offset+length will be
40 appended at the location offset, and the file will be length
41 bytes smaller. The option --keep-size may not be specified for
42 the collapse-range operation.
43 Available since Linux 3.15 for ext4 (only for extent-based
45 files) and XFS.
46 A filesystem may place limitations on the granularity of the
48 operation, in order to ensure efficient implementation. Typi‐
49 cally, offset and len must be a multiple of the filesystem logi‐
50 cal block size, which varies according to the filesystem type
51 and configuration. If a filesystem has such a requirement, the
52 operation will fail with the error EINVAL if this requirement is
53 violated.
54 -d, --dig-holes
56 Detect and dig holes. This makes the file sparse in-place,
57 without using extra disk space. The minimum size of the hole
58 depends on filesystem I/O block size (usually 4096 bytes).
59 Also, when using this option, --keep-size is implied. If no
60 range is specified by --offset and --length, then the entire
61 file is analyzed for holes.
62 You can think of this option as doing a "cp --sparse" and then
64 renaming the destination file to the original, without the need
65 for extra disk space.
66 See --punch-hole for a list of supported filesystems.
68 -i, --insert-range
70 Insert a hole of length bytes from offset, shifting existing
71 data.
72 -l, --length length
74 Specifies the length of the range, in bytes.
75 -n, --keep-size
77 Do not modify the apparent length of the file. This may effec‐
78 tively allocate blocks past EOF, which can be removed with a
79 truncate.
80 -o, --offset offset
82 Specifies the beginning offset of the range, in bytes.
83 -p, --punch-hole
85 Deallocates space (i.e., creates a hole) in the byte range
86 starting at offset and continuing for length bytes. Within the
87 specified range, partial filesystem blocks are zeroed, and whole
88 filesystem blocks are removed from the file. After a successful
89 call, subsequent reads from this range will return zeroes. This
90 option may not be specified at the same time as the --zero-range
91 option. Also, when using this option, --keep-size is implied.
92 Supported for XFS (since Linux 2.6.38), ext4 (since Linux 3.0),
94 Btrfs (since Linux 3.7), tmpfs (since Linux 3.5) and gfs2 (since
95 Linux 4.16).
96 -v, --verbose
98 Enable verbose mode.
99 -x, --posix
101 Enable POSIX operation mode. In that mode allocation operation
102 always completes, but it may take longer time when fast alloca‐
103 tion is not supported by the underlying filesystem.
104 -z, --zero-range
106 Zeroes space in the byte range starting at offset and continuing
107 for length bytes. Within the specified range, blocks are preal‐
108 located for the regions that span the holes in the file. After
109 a successful call, subsequent reads from this range will return
110 zeroes.
111 Zeroing is done within the filesystem preferably by converting
113 the range into unwritten extents. This approach means that the
114 specified range will not be physically zeroed out on the device
115 (except for partial blocks at the either end of the range), and
116 I/O is (otherwise) required only to update metadata.
117 Option --keep-size can be specified to prevent file length modi‐
119 fication.
120 Available since Linux 3.14 for ext4 (only for extent-based
122 files) and XFS.
123 -V, --version
125 Display version information and exit.
126 -h, --help
128 Display help text and exit.
129
131 Eric Sandeen ⟨sandeen@redhat.com⟩
132 Karel Zak ⟨kzak@redhat.com⟩
133
135 truncate(1), fallocate(2), posix_fallocate(3)
136
138 The fallocate command is part of the util-linux package and is avail‐
139 able from Linux Kernel Archive ⟨https://www.kernel.org/pub/linux/utils
140 /util-linux/⟩.
141util-linux April 2014 FALLOCATE(1)