TRUNCATE(1)

NAME

truncate — trim or punch holes in files

SYNOPSIS

truncate -s size [-co] file…

truncate -r ref [-c] file…

DESCRIPTION

Sets the size of files according to a size expression or to match ref. If the new size is smaller, data beyond it is lost; otherwise, it reads zero (and, if supported, may be a hole).

Unless -c is specified, files are created, otherwise nonexistent files are ignored.

With -s, size is in the case-insensitive format:

[+-<>/%][base][KMGTPEZY][B (with at least one of {base, KMGTPEZY, B})]

Where base is an optionally-floating-point number of bytes, defaulting to 1, which is then optionally multiplied by the relevant unit. B sets the unit multiplier to 1000 (from 1024). size is equal to

base \cdot {unit}^{mult}

, if any, or base.

If -o is specified, size is multiplied by each file's filesystem's block size (st_blksize stat(2) field).

By default, files are simply assigned size, but if any of +-<>/% is specified at the start:

+	fs + sz	size is added to the file's size
-	fs - sz	size is subtracted from the file's size
<	min(fs, sz)	smaller of size and the file's size is used
>	max(fs, sz)	larger of size and the file's size is used
/	fs - (fs % sz)	largest multiple of size not larger than the file's size (rounded down)
%	min(fs + (sz - (fs % sz)), fs)
		smallest multiple of size not smaller than the file's size (rounded up)

The result is then curried with max(0).

With -r, files are truncated to the size of ref.

File sizes are determined as their final seekable position. This is the same as the stated size for regular files, but it also means, that ref can trivially be a file or a block device, but not a FIFO, and directories will produce meaningless, potentially huge, results, Regular files need only be stat(2)table, but others must be able to be read.

OPTIONS

-c, --no-create: Don't create files and ignore nonexistent ones. The default is to create files with a=rw - umask if they don't already exist.
-s, --size=size: Set files' sizes to size, see above.
-o, --io-blocks: Multiply size by the file's I/O block size.
-r, --reference=ref: Set files' sizes to equal ref's.

EXIT STATUS

1 if a file's size couldn't be determined when needed, some files couldn't be opened (with the exception of ENOENT with -c) or truncated.

EXAMPLES

Create a multiple-of-16GB file at least as large as the swap area:

$ truncate -r /dev/zvol/babzoot/swap swap
truncate: /dev/zvol/babzoot/swap: Permission denied
# truncate -r /dev/zvol/babzoot/swap swap
# ls -lb swap
-rw-r--r-- 1 root staff 4294967296 Jul  3 22:07 swap
# truncate -s %16G swap
# ls -lb swap
-rw-r--r-- 1 root staff 17179869184 Jul  3 22:07 swap

Create files 1100 bytes longer than twice the block size:

$ truncate -os 2 ~/long /pub/er
$ truncate -s +1.1kB ~/long /pub/er
$ ls -lb ~/long /pub/er
-rw-r--r-- 1 cicada users   9292 Jul  3 22:01 /home/cicada/long
-rw-r--r-- 1 cicada users 263244 Jul  3 22:01 /pub/er

/pub is a zfs(4) dataset with recordsize=128K, whereas /home is a ext4(5) filesystem created with mke2fs -b 4096.

HISTORY

Оriginates from the GNU system, also available in FreeBSD 4.2.

The GNU system disallows size with B but without a multiplier, as well as lower-case B, and only supports integer bases.

FreeBSD doesn't support /%, B, nor -o.