1HARDLINK(1) User Commands HARDLINK(1)
2
6 hardlink - link multiple copies of a file
7
9 hardlink [options] [directory|file]...
10
12 hardlink is a tool that replaces copies of a file with either hardlinks
13 or copy-on-write clones, thus saving space.
14 hardlink first creates a binary tree of file sizes and then compares
16 the content of files that have the same size. There are two basic
17 content comparison methods. The memcmp method directly reads data
18 blocks from files and compares them. The other method is based on
19 checksums (like SHA256); in this case for each data block a checksum is
20 calculated by the Linux kernel crypto API, and this checksum is stored
21 in userspace and used for file comparisons.
22 For each file also an "intro" buffer (32 bytes) is cached. This buffer
24 is used independently from the comparison method and requested
25 cache-size and io-size. The "intro" buffer dramatically reduces
26 operations with data content as files are very often different from the
27 beginning.
28
30 -h, --help
31 Display help text and exit.
32 -V, --version
34 Print version and exit.
35 -v, --verbose
37 Verbose output, explain to the user what is being done. If
38 specified once, every hardlinked file is displayed. If specified
39 twice, it also shows every comparison.
40 -q, --quiet
42 Quiet mode, don’t print anything.
43 -n, --dry-run
45 Do not act, just print what would happen.
46 -y, --method name
48 Set the file content comparison method. The currently supported
49 methods are sha256, sha1, crc32c and memcmp. The default is sha256,
50 or memcmp if Linux Crypto API is not available. The methods based
51 on checksums are implemented in zero-copy way, in this case file
52 contents are not copied to the userspace and all calculation is
53 done in kernel.
54 --reflink[=when]
56 Create copy-on-write clones (aka reflinks) rather than hardlinks.
57 The reflinked files share only on-disk data, but the file mode and
58 owner can be different. It’s recommended to use it with
59 --ignore-owner and --ignore-mode options. This option implies
60 --skip-reflinks to ignore already cloned files.
61 The optional argument when can be never, always, or auto. If the
63 when argument is omitted, it defaults to auto, in this case,
64 hardlink checks filesystem type and uses reflinks on BTRFS and XFS
65 only, and fallback to hardlinks when creating reflink is
66 impossible. The argument always disables filesystem type detection
67 and fallback to hardlinks, in this case, only reflinks are allowed.
68 --skip-reflinks
70 Ignore already cloned files. This option may be used without
71 --reflink when creating classic hardlinks.
72 -f, --respect-name
74 Only try to link files with the same (base)name. It’s strongly
75 recommended to use long options rather than -f which is interpreted
76 in a different way by other hardlink implementations.
77 -p, --ignore-mode
79 Link and compare files even if their mode is different. Results may
80 be slightly unpredictable.
81 -o, --ignore-owner
83 Link and compare files even if their owner information (user and
84 group) differs. Results may be unpredictable.
85 -t, --ignore-time
87 Link and compare files even if their time of modification is
88 different. This is usually a good choice.
89 -c --content
91 Consider only file content, not attributes, when determining
92 whether two files are equal. Same as -pot.
93 -X, --respect-xattrs
95 Only try to link files with the same extended attributes.
96 -m, --maximize
98 Among equal files, keep the file with the highest link count.
99 -M, --minimize
101 Among equal files, keep the file with the lowest link count.
102 -O, --keep-oldest
104 Among equal files, keep the oldest file (least recent modification
105 time). By default, the newest file is kept. If --maximize or
106 --minimize is specified, the link count has a higher precedence
107 than the time of modification.
108 -x, --exclude regex
110 A regular expression which excludes files from being compared and
111 linked.
112 -i, --include regex
114 A regular expression to include files. If the option --exclude has
115 been given, this option re-includes files which would otherwise be
116 excluded. If the option is used without --exclude, only files
117 matched by the pattern are included.
118 -s, --minimum-size size
120 The minimum size to consider. By default this is 1, so empty files
121 will not be linked. The size argument may be followed by the
122 multiplicative suffixes KiB (=1024), MiB (=1024*1024), and so on
123 for GiB, TiB, PiB, EiB, ZiB and YiB (the "iB" is optional, e.g.,
124 "K" has the same meaning as "KiB").
125 -S, --maximum-size size
127 The maximum size to consider. By default this is 0 and 0 has the
128 special meaning of unlimited. The size argument may be followed by
129 the multiplicative suffixes KiB (=1024), MiB (=1024*1024), and so
130 on for GiB, TiB, PiB, EiB, ZiB and YiB (the "iB" is optional, e.g.,
131 "K" has the same meaning as "KiB").
132 -b, --io-size size
134 The size of the read(2) or sendfile(2) buffer used when comparing
135 file contents. The size argument may be followed by the
136 multiplicative suffixes KiB, MiB, etc. The "iB" is optional, e.g.,
137 "K" has the same meaning as "KiB". The default is 8KiB for memcmp
138 method and 1MiB for the other methods. The only memcmp method uses
139 process memory for the buffer, other methods use zero-copy way and
140 I/O operation is done in the kernel. The size may be altered on the
141 fly to fit a number of cached content checksums.
142 -r, --cache-size size
144 The size of the cache for content checksums. All non-memcmp methods
145 calculate checksum for each file content block (see --io-size),
146 these checksums are cached for the next comparison. The size is
147 important for large files or a large sets of files of the same
148 size. The default is 10MiB.
149
151 hardlink takes one or more directories which will be searched for files
152 to be linked.
153
155 The original hardlink implementation uses the option -f to force
156 hardlinks creation between filesystem. This very rarely usable feature
157 is no more supported by the current hardlink.
158 hardlink assumes that the trees it operates on do not change during
160 operation. If a tree does change, the result is undefined and
161 potentially dangerous. For example, if a regular file is replaced by a
162 device, hardlink may start reading from the device. If a component of a
163 path is replaced by a symbolic link or file permissions change,
164 security may be compromised. Do not run hardlink on a changing tree or
165 on a tree controlled by another user.
166
168 There are multiple hardlink implementations. The very first
169 implementation is from Jakub Jelinek for Fedora distribution, this
170 implementation has been used in util-linux between versions v2.34 to
171 v2.36. The current implementations is based on Debian version from
172 Julian Andres Klode.
173
175 For bug reports, use the issue tracker at
176 https://github.com/util-linux/util-linux/issues.
177
179 The hardlink command is part of the util-linux package which can be
180 downloaded from Linux Kernel Archive
181 <https://www.kernel.org/pub/linux/utils/util-linux/>.
182util-linux 2.38.1 2022-08-04 HARDLINK(1)