rsync Man page

Resume Wikipedia de Rsync

rsync remote synchronization, (en français : « synchronisation distante ») est un logiciel libre de synchronisation de fichiers, distribué sous licence GNU GPL. La synchronisation est unidirectionnelle, c’est-à-dire qu’elle copie les fichiers de la source en direction de la destination. rsync est donc utilisé pour réaliser des Sauvegarde incrémentielle ou décrémentielle ou pour diffuser le contenu d’un répertoire de référence.
“rsync” fonctionne sur une large gamme de systèmes d’exploitation (Microsoft Windows, GNU/Linux, Mac OS X, Unix), permettant ainsi de synchroniser des fichiers de différents systèmes d’exploitation.

Resume Wikipedia de Rsync

rsync remote synchronization, (en français : « synchronisation distante ») est un logiciel libre de synchronisation de fichiers, distribué sous licence GNU GPL. La synchronisation est unidirectionnelle, c’est-à-dire qu’elle copie les fichiers de la source en direction de la destination. rsync est donc utilisé pour réaliser des Sauvegarde incrémentielle ou décrémentielle ou pour diffuser le contenu d’un répertoire de référence.
“rsync” fonctionne sur une large gamme de systèmes d’exploitation (Microsoft Windows, GNU/Linux, Mac OS X, Unix), permettant ainsi de synchroniser des fichiers de différents systèmes d’exploitation.

rsync rsync


rsync – a fast, versatile, remote (and local) file-copying tool


Local: rsync [OPTION…] SRC… [DEST]

Access via remote shell:
Pull: rsync [OPTION…] [USER@]HOST:SRC… [DEST] Push: rsync [OPTION…] SRC… [USER@]HOST:DEST

Access via rsync daemon:
Pull: rsync [OPTION…] [USER@]HOST::SRC… [DEST] rsync [OPTION…] rsync://[USER@]HOST[:PORT]/SRC… [DEST] Push: rsync [OPTION…] SRC… [USER@]HOST::DEST
rsync [OPTION…] SRC… rsync://[USER@]HOST[:PORT]/DEST

Usages with just one SRC arg and no DEST arg will list the source files
instead of copying.


Rsync is a fast and extraordinarily versatile file copying tool. It
can copy locally, to/from another host over any remote shell, or
to/from a remote rsync daemon. It offers a large number of options
that control every aspect of its behavior and permit very flexible
specification of the set of files to be copied. It is famous for its
delta-transfer algorithm, which reduces the amount of data sent over
the network by sending only the differences between the source files
and the existing files in the destination. Rsync is widely used for
backups and mirroring and as an improved copy command for everyday use.

Rsync finds files that need to be transferred using a “quick check”
algorithm (by default) that looks for files that have changed in size
or in last-modified time. Any changes in the other preserved
attributes (as requested by options) are made on the destination file
directly when the quick check indicates that the file’s data does not
need to be updated.

Some of the additional features of rsync are:

o support for copying links, devices, owners, groups, and permis‐

o exclude and exclude-from options similar to GNU tar

o a CVS exclude mode for ignoring the same files that CVS would

o can use any transparent remote shell, including ssh or rsh

o does not require super-user privileges

o pipelining of file transfers to minimize latency costs

o support for anonymous or authenticated rsync daemons (ideal for

Rsync copies files either to or from a remote host, or locally on the
current host (it does not support copying files between two remote

There are two different ways for rsync to contact a remote system:
using a remote-shell program as the transport (such as ssh or rsh) or
contacting an rsync daemon directly via TCP. The remote-shell trans‐
port is used whenever the source or destination path contains a single
colon (:) separator after a host specification. Contacting an rsync
daemon directly happens when the source or destination path contains a
double colon (::) separator after a host specification, OR when an
rsync:// URL is specified (see also the “USING RSYNC-DAEMON FEATURES
VIA A REMOTE-SHELL CONNECTION” section for an exception to this latter

As a special case, if a single source arg is specified without a desti‐
nation, the files are listed in an output format similar to “ls -l”.

As expected, if neither the source or destination path specify a remote
host, the copy occurs locally (see also the –list-only option).

Rsync refers to the local side as the “client” and the remote side as
the “server”. Don’t confuse “server” with an rsync daemon — a daemon
is always a server, but a server can be either a daemon or a
remote-shell spawned process.

See the file README for installation instructions.

Once installed, you can use rsync to any machine that you can access
via a remote shell (as well as some that you can access using the rsync
daemon-mode protocol). For remote transfers, a modern rsync uses ssh
for its communications, but it may have been configured to use a dif‐
ferent remote shell by default, such as rsh or remsh.

You can also specify any remote shell you like, either by using the -e
command line option, or by setting the RSYNC_RSH environment variable.

Note that rsync must be installed on both the source and destination

You use rsync in the same way you use rcp. You must specify a source
and a destination, one of which may be remote.

Perhaps the best way to explain the syntax is with some examples:

rsync -t *.c foo:src/

This would transfer all files matching the pattern *.c from the current
directory to the directory src on the machine foo. If any of the files
already exist on the remote system then the rsync remote-update proto‐
col is used to update the file by sending only the differences in the
data. Note that the expansion of wildcards on the commandline (*.c)
into a list of files is handled by the shell before it runs rsync and
not by rsync itself (exactly the same as all other posix-style pro‐

rsync -avz foo:src/bar /data/tmp

This would recursively transfer all files from the directory src/bar on
the machine foo into the /data/tmp/bar directory on the local machine.
The files are transferred in “archive” mode, which ensures that sym‐
bolic links, devices, attributes, permissions, ownerships, etc. are
preserved in the transfer. Additionally, compression will be used to
reduce the size of data portions of the transfer.

rsync -avz foo:src/bar/ /data/tmp

A trailing slash on the source changes this behavior to avoid creating
an additional directory level at the destination. You can think of a
trailing / on a source as meaning “copy the contents of this directory”
as opposed to “copy the directory by name”, but in both cases the
attributes of the containing directory are transferred to the contain‐
ing directory on the destination. In other words, each of the follow‐
ing commands copies the files in the same way, including their setting
of the attributes of /dest/foo:

rsync -av /src/foo /dest
rsync -av /src/foo/ /dest/foo

Note also that host and module references don’t require a trailing
slash to copy the contents of the default directory. For example, both
of these copy the remote directory’s contents into “/dest”:

rsync -av host: /dest
rsync -av host::module /dest

You can also use rsync in local-only mode, where both the source and
destination don’t have a ’:’ in the name. In this case it behaves like
an improved copy command.

Finally, you can list all the (listable) modules available from a par‐
ticular rsync daemon by leaving off the module name:


See the following section for more details.

The syntax for requesting multiple files from a remote host is done by
specifying additional remote-host args in the same style as the first,
or with the hostname omitted. For instance, all these work:

rsync -av host:file1 :file2 host:file{3,4} /dest/
rsync -av host::modname/file{1,2} host::modname/file3 /dest/
rsync -av host::modname/file1 ::modname/file{3,4}

Older versions of rsync required using quoted spaces in the SRC, like
these examples:

rsync -av host:’dir1/file1 dir2/file2′ /dest
rsync host::’modname/dir1/file1 modname/dir2/file2′ /dest

This word-splitting still works (by default) in the latest rsync, but
is not as easy to use as the first method.

If you need to transfer a filename that contains whitespace, you can
either specify the –protect-args (-s) option, or you’ll need to escape
the whitespace in a way that the remote shell will understand. For

rsync -av host:’file\ name\ with\ spaces’ /dest

It is also possible to use rsync without a remote shell as the trans‐
port. In this case you will directly connect to a remote rsync daemon,
typically using TCP port 873. (This obviously requires the daemon to
be running on the remote system, so refer to the STARTING AN RSYNC DAE‐
MON TO ACCEPT CONNECTIONS section below for information on that.)

Using rsync in this way is the same as using it with a remote shell
except that:

o you either use a double colon :: instead of a single colon to
separate the hostname from the path, or you use an rsync:// URL.

o the first word of the “path” is actually a module name.

o the remote daemon may print a message of the day when you con‐

o if you specify no path name on the remote daemon then the list
of accessible paths on the daemon will be shown.

o if you specify no local destination then a listing of the speci‐
fied files on the remote daemon is provided.

o you must not specify the –rsh (-e) option.

An example that copies all the files in a remote module named “src”:

rsync -av host::src /dest

Some modules on the remote daemon may require authentication. If so,
you will receive a password prompt when you connect. You can avoid the
password prompt by setting the environment variable RSYNC_PASSWORD to
the password you want to use or using the –password-file option. This
may be useful when scripting rsync.

WARNING: On some systems environment variables are visible to all
users. On those systems using –password-file is recommended.

You may establish the connection via a web proxy by setting the envi‐
ronment variable RSYNC_PROXY to a hostname:port pair pointing to your
web proxy. Note that your web proxy’s configuration must support proxy
connections to port 873.

You may also establish a daemon connection using a program as a proxy
by setting the environment variable RSYNC_CONNECT_PROG to the commands
you wish to run in place of making a direct socket connection. The
string may contain the escape “%H” to represent the hostname specified
in the rsync command (so use “%%” if you need a single “%” in your
string). For example:

export RSYNC_CONNECT_PROG=’ssh proxyhost nc %H 873′
rsync -av targethost1::module/src/ /dest/
rsync -av rsync:://targethost2/module/src/ /dest/

The command specified above uses ssh to run nc (netcat) on a proxyhost,
which forwards all data to port 873 (the rsync daemon) on the targeth‐
ost (%H).

It is sometimes useful to use various features of an rsync daemon (such
as named modules) without actually allowing any new socket connections
into a system (other than what is already required to allow
remote-shell access). Rsync supports connecting to a host using a
remote shell and then spawning a single-use “daemon” server that
expects to read its config file in the home dir of the remote user.
This can be useful if you want to encrypt a daemon-style transfer’s
data, but since the daemon is started up fresh by the remote user, you
may not be able to use features such as chroot or change the uid used
by the daemon. (For another way to encrypt a daemon transfer, consider
using ssh to tunnel a local port to a remote machine and configure a
normal rsync daemon on that remote host to only allow connections from

From the user’s perspective, a daemon transfer via a remote-shell con‐
nection uses nearly the same command-line syntax as a normal rsync-dae‐
mon transfer, with the only exception being that you must explicitly
set the remote shell program on the command-line with the –rsh=COMMAND
option. (Setting the RSYNC_RSH in the environment will not turn on
this functionality.) For example:

rsync -av –rsh=ssh host::module /dest

If you need to specify a different remote-shell user, keep in mind that
the user@ prefix in front of the host is specifying the rsync-user
value (for a module that requires user-based authentication). This
means that you must give the ’-l user’ option to ssh when specifying
the remote-shell, as in this example that uses the short version of the
–rsh option:

rsync -av -e “ssh -l ssh-user” rsync-user@host::module /dest

The “ssh-user” will be used at the ssh level; the “rsync-user” will be
used to log-in to the “module”.

In order to connect to an rsync daemon, the remote system needs to have
a daemon already running (or it needs to have configured something like
inetd to spawn an rsync daemon for incoming connections on a particular
port). For full information on how to start a daemon that will han‐
dling incoming socket connections, see the rsyncd.conf(5) man page —
that is the config file for the daemon, and it contains the full
details for how to run the daemon (including stand-alone and inetd con‐

If you’re using one of the remote-shell transports for the transfer,
there is no need to manually start an rsync daemon.

Rsync always sorts the specified filenames into its internal transfer
list. This handles the merging together of the contents of identically
named directories, makes it easy to remove duplicate filenames, and may
confuse someone when the files are transferred in a different order
than what was given on the command-line.

If you need a particular file to be transferred prior to another,
either separate the files into different rsync calls, or consider using
–delay-updates (which doesn’t affect the sorted transfer order, but
does make the final file-updating phase happen much more rapidly).

Here are some examples of how I use rsync.

To backup my wife’s home directory, which consists of large MS Word
files and mail folders, I use a cron job that runs

rsync -Cavz . arvidsjaur:backup

each night over a PPP connection to a duplicate directory on my machine

To synchronize my samba source trees I use the following Makefile tar‐

rsync -avuzb –exclude ‘*~’ samba:samba/ .
rsync -Cavuzb . samba:samba/
sync: get put

this allows me to sync with a CVS directory at the other end of the
connection. I then do CVS operations on the remote machine, which saves
a lot of time as the remote CVS protocol isn’t very efficient.

I mirror a directory between my “old” and “new” ftp sites with the com‐

rsync -az -e ssh –delete ~ftp/pub/samba nimbus:”~ftp/pub/tridge”

This is launched from cron every few hours.

Here is a short summary of the options available in rsync. Please refer
to the detailed description below for a complete description.

-v, –verbose increase verbosity
–info=FLAGS fine-grained informational verbosity
–debug=FLAGS fine-grained debug verbosity
–msgs2stderr special output handling for debugging
-q, –quiet suppress non-error messages
–no-motd suppress daemon-mode MOTD (see caveat)
-c, –checksum skip based on checksum, not mod-time & size
-a, –archive archive mode; equals -rlptgoD (no -H,-A,-X)
–no-OPTION turn off an implied OPTION (e.g. –no-D)
-r, –recursive recurse into directories
-R, –relative use relative path names
–no-implied-dirs don’t send implied dirs with –relative
-b, –backup make backups (see –suffix & –backup-dir)
–backup-dir=DIR make backups into hierarchy based in DIR
–suffix=SUFFIX backup suffix (default ~ w/o –backup-dir)
-u, –update skip files that are newer on the receiver
–inplace update destination files in-place
–append append data onto shorter files
–append-verify –append w/old data in file checksum
-d, –dirs transfer directories without recursing
-l, –links copy symlinks as symlinks
-L, –copy-links transform symlink into referent file/dir
–copy-unsafe-links only “unsafe” symlinks are transformed
–safe-links ignore symlinks that point outside the tree
–munge-links munge symlinks to make them safer
-k, –copy-dirlinks transform symlink to dir into referent dir
-K, –keep-dirlinks treat symlinked dir on receiver as dir
-H, –hard-links preserve hard links
-p, –perms preserve permissions
-E, –executability preserve executability
–chmod=CHMOD affect file and/or directory permissions
-A, –acls preserve ACLs (implies -p)
-X, –xattrs preserve extended attributes
-o, –owner preserve owner (super-user only)
-g, –group preserve group
–devices preserve device files (super-user only)
–specials preserve special files
-D same as –devices –specials
-t, –times preserve modification times
-O, –omit-dir-times omit directories from –times
-J, –omit-link-times omit symlinks from –times
–super receiver attempts super-user activities
–fake-super store/recover privileged attrs using xattrs
-S, –sparse handle sparse files efficiently
–preallocate allocate dest files before writing
-n, –dry-run perform a trial run with no changes made
-W, –whole-file copy files whole (w/o delta-xfer algorithm)
-x, –one-file-system don’t cross filesystem boundaries
-B, –block-size=SIZE force a fixed checksum block-size
-e, –rsh=COMMAND specify the remote shell to use
–rsync-path=PROGRAM specify the rsync to run on remote machine
–existing skip creating new files on receiver
–ignore-existing skip updating files that exist on receiver
–remove-source-files sender removes synchronized files (non-dir)
–del an alias for –delete-during
–delete delete extraneous files from dest dirs
–delete-before receiver deletes before xfer, not during
–delete-during receiver deletes during the transfer
–delete-delay find deletions during, delete after
–delete-after receiver deletes after transfer, not during
–delete-excluded also delete excluded files from dest dirs
–ignore-missing-args ignore missing source args without error
–delete-missing-args delete missing source args from destination
–ignore-errors delete even if there are I/O errors
–force force deletion of dirs even if not empty
–max-delete=NUM don’t delete more than NUM files
–max-size=SIZE don’t transfer any file larger than SIZE
–min-size=SIZE don’t transfer any file smaller than SIZE
–partial keep partially transferred files
–partial-dir=DIR put a partially transferred file into DIR
–delay-updates put all updated files into place at end
-m, –prune-empty-dirs prune empty directory chains from file-list
–numeric-ids don’t map uid/gid values by user/group name
–usermap=STRING custom username mapping
–groupmap=STRING custom groupname mapping
–chown=USER:GROUP simple username/groupname mapping
–timeout=SECONDS set I/O timeout in seconds
–contimeout=SECONDS set daemon connection timeout in seconds
-I, –ignore-times don’t skip files that match size and time
–size-only skip files that match in size
–modify-window=NUM compare mod-times with reduced accuracy
-T, –temp-dir=DIR create temporary files in directory DIR
-y, –fuzzy find similar file for basis if no dest file
–compare-dest=DIR also compare received files relative to DIR
–copy-dest=DIR … and include copies of unchanged files
–link-dest=DIR hardlink to files in DIR when unchanged
-z, –compress compress file data during the transfer
–compress-level=NUM explicitly set compression level
–skip-compress=LIST skip compressing files with suffix in LIST
-C, –cvs-exclude auto-ignore files in the same way CVS does
-f, –filter=RULE add a file-filtering RULE
-F same as –filter=’dir-merge /.rsync-filter’
repeated: –filter=’- .rsync-filter’
–exclude=PATTERN exclude files matching PATTERN
–exclude-from=FILE read exclude patterns from FILE
–include=PATTERN don’t exclude files matching PATTERN
–include-from=FILE read include patterns from FILE
–files-from=FILE read list of source-file names from FILE
-0, –from0 all *from/filter files are delimited by 0s
-s, –protect-args no space-splitting; wildcard chars only
–address=ADDRESS bind address for outgoing socket to daemon
–port=PORT specify double-colon alternate port number
–sockopts=OPTIONS specify custom TCP options
–blocking-io use blocking I/O for the remote shell
–outbuf=N|L|B set out buffering to None, Line, or Block
–stats give some file-transfer stats
-8, –8-bit-output leave high-bit chars unescaped in output
-h, –human-readable output numbers in a human-readable format
–progress show progress during transfer
-P same as –partial –progress
-i, –itemize-changes output a change-summary for all updates
-M, –remote-option=OPTION send OPTION to the remote side only
–out-format=FORMAT output updates using the specified FORMAT
–log-file=FILE log what we’re doing to the specified FILE
–log-file-format=FMT log updates using the specified FMT
–password-file=FILE read daemon-access password from FILE
–list-only list the files instead of copying them
–bwlimit=RATE limit socket I/O bandwidth
–write-batch=FILE write a batched update to FILE
–only-write-batch=FILE like –write-batch but w/o updating dest
–read-batch=FILE read a batched update from FILE
–protocol=NUM force an older protocol version to be used
–iconv=CONVERT_SPEC request charset conversion of filenames
–checksum-seed=NUM set block/file checksum seed (advanced)
-4, –ipv4 prefer IPv4
-6, –ipv6 prefer IPv6
–version print version number
(-h) –help show this help (see below for -h comment)

Rsync can also be run as a daemon, in which case the following options
are accepted:

–daemon run as an rsync daemon
–address=ADDRESS bind to the specified address
–bwlimit=RATE limit socket I/O bandwidth
–config=FILE specify alternate rsyncd.conf file
-M, –dparam=OVERRIDE override global daemon config parameter
–no-detach do not detach from the parent
–port=PORT listen on alternate port number
–log-file=FILE override the “log file” setting
–log-file-format=FMT override the “log format” setting
–sockopts=OPTIONS specify custom TCP options
-v, –verbose increase verbosity
-4, –ipv4 prefer IPv4
-6, –ipv6 prefer IPv6
-h, –help show this help (if used after –daemon)


Rsync accepts both long (double-dash + word) and short (single-dash +
letter) options. The full list of the available options are described
below. If an option can be specified in more than one way, the choices
are comma-separated. Some options only have a long variant, not a
short. If the option takes a parameter, the parameter is only listed
after the long variant, even though it must also be specified for the
short. When specifying a parameter, you can either use the form
–option=param or replace the ’=’ with whitespace. The parameter may
need to be quoted in some manner for it to survive the shell’s com‐
mand-line parsing. Keep in mind that a leading tilde (~) in a filename
is substituted by your shell, so –option=~/foo will not change the
tilde into your home directory (remove the ’=’ for that).

–help Print a short help page describing the options available in
rsync and exit. For backward-compatibility with older versions
of rsync, the help will also be output if you use the -h option
without any other args.

print the rsync version number and exit.

-v, –verbose
This option increases the amount of information you are given
during the transfer. By default, rsync works silently. A single
-v will give you information about what files are being trans‐
ferred and a brief summary at the end. Two -v options will give
you information on what files are being skipped and slightly
more information at the end. More than two -v options should
only be used if you are debugging rsync.

In a modern rsync, the -v option is equivalent to the setting of
groups of –info and –debug options. You can choose to use
these newer options in addition to, or in place of using –ver‐
bose, as any fine-grained settings override the implied settings
of -v. Both –info and –debug have a way to ask for help that
tells you exactly what flags are set for each increase in ver‐

However, do keep in mind that a daemon’s “max verbosity” setting
will limit how high of a level the various individual flags can
be set on the daemon side. For instance, if the max is 2, then
any info and/or debug flag that is set to a higher value than
what would be set by -vv will be downgraded to the -vv level in
the daemon’s logging.

This option lets you have fine-grained control over the informa‐
tion output you want to see. An individual flag name may be
followed by a level number, with 0 meaning to silence that out‐
put, 1 being the default output level, and higher numbers
increasing the output of that flag (for those that support
higher levels). Use –info=help to see all the available flag
names, what they output, and what flag names are added for each
increase in the verbose level. Some examples:

rsync -a –info=progress2 src/ dest/
rsync -avv –info=stats2,misc1,flist0 src/ dest/

Note that –info=name’s output is affected by the –out-format
and –itemize-changes (-i) options. See those options for more
information on what is output and when.

This option was added to 3.1.0, so an older rsync on the server
side might reject your attempts at fine-grained control (if one
or more flags needed to be send to the server and the server was
too old to understand them). See also the “max verbosity”
caveat above when dealing with a daemon.

This option lets you have fine-grained control over the debug
output you want to see. An individual flag name may be followed
by a level number, with 0 meaning to silence that output, 1
being the default output level, and higher numbers increasing
the output of that flag (for those that support higher levels).
Use –debug=help to see all the available flag names, what they
output, and what flag names are added for each increase in the
verbose level. Some examples:

rsync -avvv –debug=none src/ dest/
rsync -avA –del –debug=del2,acl src/ dest/

Note that some debug messages will only be output when
–msgs2stderr is specified, especially those pertaining to I/O
and buffer debugging.

This option was added to 3.1.0, so an older rsync on the server
side might reject your attempts at fine-grained control (if one
or more flags needed to be send to the server and the server was
too old to understand them). See also the “max verbosity”
caveat above when dealing with a daemon.

This option changes rsync to send all its output directly to
stderr rather than to send messages to the client side via the
protocol (which normally outputs info messages via stdout).
This is mainly intended for debugging in order to avoid changing
the data sent via the protocol, since the extra protocol data
can change what is being tested. Keep in mind that a daemon
connection does not have a stderr channel to send messages back
to the client side, so if you are doing any daemon-transfer
debugging using this option, you should start up a daemon using
–no-detach so that you can see the stderr output on the daemon

This option has the side-effect of making stderr output get
line-buffered so that the merging of the output of 3 programs
happens in a more readable manner.

-q, –quiet
This option decreases the amount of information you are given
during the transfer, notably suppressing information messages
from the remote server. This option is useful when invoking
rsync from cron.

This option affects the information that is output by the client
at the start of a daemon transfer. This suppresses the mes‐
sage-of-the-day (MOTD) text, but it also affects the list of
modules that the daemon sends in response to the “rsync host::”
request (due to a limitation in the rsync protocol), so omit
this option if you want to request the list of modules from the

-I, –ignore-times
Normally rsync will skip any files that are already the same
size and have the same modification timestamp. This option
turns off this “quick check” behavior, causing all files to be

This modifies rsync’s “quick check” algorithm for finding files
that need to be transferred, changing it from the default of
transferring files with either a changed size or a changed
last-modified time to just looking for files that have changed
in size. This is useful when starting to use rsync after using
another mirroring system which may not preserve timestamps

When comparing two timestamps, rsync treats the timestamps as
being equal if they differ by no more than the modify-window
value. This is normally 0 (for an exact match), but you may
find it useful to set this to a larger value in some situations.
In particular, when transferring to or from an MS Windows FAT
filesystem (which represents times with a 2-second resolution),
–modify-window=1 is useful (allowing times to differ by up to 1

-c, –checksum
This changes the way rsync checks if the files have been changed
and are in need of a transfer. Without this option, rsync uses
a “quick check” that (by default) checks if each file’s size and
time of last modification match between the sender and receiver.
This option changes this to compare a 128-bit checksum for each
file that has a matching size. Generating the checksums means
that both sides will expend a lot of disk I/O reading all the
data in the files in the transfer (and this is prior to any
reading that will be done to transfer changed files), so this
can slow things down significantly.

The sending side generates its checksums while it is doing the
file-system scan that builds the list of the available files.
The receiver generates its checksums when it is scanning for
changed files, and will checksum any file that has the same size
as the corresponding sender’s file: files with either a changed
size or a changed checksum are selected for transfer.

Note that rsync always verifies that each transferred file was
correctly reconstructed on the receiving side by checking a
whole-file checksum that is generated as the file is trans‐
ferred, but that automatic after-the-transfer verification has
nothing to do with this option’s before-the-transfer “Does this
file need to be updated?” check.

For protocol 30 and beyond (first supported in 3.0.0), the
checksum used is MD5. For older protocols, the checksum used is

-a, –archive
This is equivalent to -rlptgoD. It is a quick way of saying you
want recursion and want to preserve almost everything (with -H
being a notable omission). The only exception to the above
equivalence is when –files-from is specified, in which case -r
is not implied.

Note that -a does not preserve hardlinks, because finding multi‐
ply-linked files is expensive. You must separately specify -H.

You may turn off one or more implied options by prefixing the
option name with “no-“. Not all options may be prefixed with a
“no-“: only options that are implied by other options (e.g.
–no-D, –no-perms) or have different defaults in various cir‐
cumstances (e.g. –no-whole-file, –no-blocking-io, –no-dirs).
You may specify either the short or the long option name after
the “no-” prefix (e.g. –no-R is the same as –no-relative).

For example: if you want to use -a (–archive) but don’t want -o
(–owner), instead of converting -a into -rlptgD, you could
specify -a –no-o (or -a –no-owner).

The order of the options is important: if you specify –no-r
-a, the -r option would end up being turned on, the opposite of
-a –no-r. Note also that the side-effects of the –files-from
option are NOT positional, as it affects the default state of
several options and slightly changes the meaning of -a (see the
–files-from option for more details).

-r, –recursive
This tells rsync to copy directories recursively. See also
–dirs (-d).

Beginning with rsync 3.0.0, the recursive algorithm used is now
an incremental scan that uses much less memory than before and
begins the transfer after the scanning of the first few directo‐
ries have been completed. This incremental scan only affects
our recursion algorithm, and does not change a non-recursive
transfer. It is also only possible when both ends of the trans‐
fer are at least version 3.0.0.

Some options require rsync to know the full file list, so these
options disable the incremental recursion mode. These include:
–delete-before, –delete-after, –prune-empty-dirs, and
–delay-updates. Because of this, the default delete mode when
you specify –delete is now –delete-during when both ends of
the connection are at least 3.0.0 (use –del or –delete-during
to request this improved deletion mode explicitly). See also
the –delete-delay option that is a better choice than using

Incremental recursion can be disabled using the –no-inc-recur‐
sive option or its shorter –no-i-r alias.

-R, –relative
Use relative paths. This means that the full path names speci‐
fied on the command line are sent to the server rather than just
the last parts of the filenames. This is particularly useful
when you want to send several different directories at the same
time. For example, if you used this command:

rsync -av /foo/bar/baz.c remote:/tmp/

… this would create a file named baz.c in /tmp/ on the remote
machine. If instead you used

rsync -avR /foo/bar/baz.c remote:/tmp/

then a file named /tmp/foo/bar/baz.c would be created on the
remote machine, preserving its full path. These extra path ele‐
ments are called “implied directories” (i.e. the “foo” and the
“foo/bar” directories in the above example).

Beginning with rsync 3.0.0, rsync always sends these implied
directories as real directories in the file list, even if a path
element is really a symlink on the sending side. This prevents
some really unexpected behaviors when copying the full path of a
file that you didn’t realize had a symlink in its path. If you
want to duplicate a server-side symlink, include both the sym‐
link via its path, and referent directory via its real path. If
you’re dealing with an older rsync on the sending side, you may
need to use the –no-implied-dirs option.

It is also possible to limit the amount of path information that
is sent as implied directories for each path you specify. With
a modern rsync on the sending side (beginning with 2.6.7), you
can insert a dot and a slash into the source path, like this:

rsync -avR /foo/./bar/baz.c remote:/tmp/

That would create /tmp/bar/baz.c on the remote machine. (Note
that the dot must be followed by a slash, so “/foo/.” would not
be abbreviated.) For older rsync versions, you would need to
use a chdir to limit the source path. For example, when pushing

(cd /foo; rsync -avR bar/baz.c remote:/tmp/)

(Note that the parens put the two commands into a sub-shell, so
that the “cd” command doesn’t remain in effect for future com‐
mands.) If you’re pulling files from an older rsync, use this
idiom (but only for a non-daemon transfer):

rsync -avR –rsync-path=”cd /foo; rsync” \
remote:bar/baz.c /tmp/

This option affects the default behavior of the –relative
option. When it is specified, the attributes of the implied
directories from the source names are not included in the trans‐
fer. This means that the corresponding path elements on the
destination system are left unchanged if they exist, and any
missing implied directories are created with default attributes.
This even allows these implied path elements to have big differ‐
ences, such as being a symlink to a directory on the receiving

For instance, if a command-line arg or a files-from entry told
rsync to transfer the file “path/foo/file”, the directories
“path” and “path/foo” are implied when –relative is used. If
“path/foo” is a symlink to “bar” on the destination system, the
receiving rsync would ordinarily delete “path/foo”, recreate it
as a directory, and receive the file into the new directory.
With –no-implied-dirs, the receiving rsync updates
“path/foo/file” using the existing path elements, which means
that the file ends up being created in “path/bar”. Another way
to accomplish this link preservation is to use the
–keep-dirlinks option (which will also affect symlinks to
directories in the rest of the transfer).

When pulling files from an rsync older than 3.0.0, you may need
to use this option if the sending side has a symlink in the path
you request and you wish the implied directories to be trans‐
ferred as normal directories.

-b, –backup
With this option, preexisting destination files are renamed as
each file is transferred or deleted. You can control where the
backup file goes and what (if any) suffix gets appended using
the –backup-dir and –suffix options.

Note that if you don’t specify –backup-dir, (1) the
–omit-dir-times option will be implied, and (2) if –delete is
also in effect (without –delete-excluded), rsync will add a
“protect” filter-rule for the backup suffix to the end of all
your existing excludes (e.g. -f “P *~”). This will prevent pre‐
viously backed-up files from being deleted. Note that if you
are supplying your own filter rules, you may need to manually
insert your own exclude/protect rule somewhere higher up in the
list so that it has a high enough priority to be effective
(e.g., if your rules specify a trailing inclusion/exclusion of
’*’, the auto-added rule would never be reached).

In combination with the –backup option, this tells rsync to
store all backups in the specified directory on the receiving
side. This can be used for incremental backups. You can addi‐
tionally specify a backup suffix using the –suffix option (oth‐
erwise the files backed up in the specified directory will keep
their original filenames).

Note that if you specify a relative path, the backup directory
will be relative to the destination directory, so you probably
want to specify either an absolute path or a path that starts
with “../”. If an rsync daemon is the receiver, the backup dir
cannot go outside the module’s path hierarchy, so take extra
care not to delete it or copy into it.

This option allows you to override the default backup suffix
used with the –backup (-b) option. The default suffix is a ~ if
no –backup-dir was specified, otherwise it is an empty string.

-u, –update
This forces rsync to skip any files which exist on the destina‐
tion and have a modified time that is newer than the source
file. (If an existing destination file has a modification time
equal to the source file’s, it will be updated if the sizes are

Note that this does not affect the copying of dirs, symlinks, or
other special files. Also, a difference of file format between
the sender and receiver is always considered to be important
enough for an update, no matter what date is on the objects. In
other words, if the source has a directory where the destination
has a file, the transfer would occur regardless of the time‐

This option is a transfer rule, not an exclude, so it doesn’t
affect the data that goes into the file-lists, and thus it
doesn’t affect deletions. It just limits the files that the
receiver requests to be transferred.

This option changes how rsync transfers a file when its data
needs to be updated: instead of the default method of creating a
new copy of the file and moving it into place when it is com‐
plete, rsync instead writes the updated data directly to the
destination file.

This has several effects:

o Hard links are not broken. This means the new data will
be visible through other hard links to the destination
file. Moreover, attempts to copy differing source files
onto a multiply-linked destination file will result in a
“tug of war” with the destination data changing back and

o In-use binaries cannot be updated (either the OS will
prevent this from happening, or binaries that attempt to
swap-in their data will misbehave or crash).

o The file’s data will be in an inconsistent state during
the transfer and will be left that way if the transfer is
interrupted or if an update fails.

o A file that rsync cannot write to cannot be updated.
While a super user can update any file, a normal user
needs to be granted write permission for the open of the
file for writing to be successful.

o The efficiency of rsync’s delta-transfer algorithm may be
reduced if some data in the destination file is overwrit‐
ten before it can be copied to a position later in the
file. This does not apply if you use –backup, since
rsync is smart enough to use the backup file as the basis
file for the transfer.

WARNING: you should not use this option to update files that are
being accessed by others, so be careful when choosing to use
this for a copy.

This option is useful for transferring large files with
block-based changes or appended data, and also on systems that
are disk bound, not network bound. It can also help keep a
copy-on-write filesystem snapshot from diverging the entire con‐
tents of a file that only has minor changes.

The option implies –partial (since an interrupted transfer does
not delete the file), but conflicts with –partial-dir and
–delay-updates. Prior to rsync 2.6.4 –inplace was also incom‐
patible with –compare-dest and –link-dest.

This causes rsync to update a file by appending data onto the
end of the file, which presumes that the data that already
exists on the receiving side is identical with the start of the
file on the sending side. If a file needs to be transferred and
its size on the receiver is the same or longer than the size on
the sender, the file is skipped. This does not interfere with
the updating of a file’s non-content attributes (e.g. permis‐
sions, ownership, etc.) when the file does not need to be trans‐
ferred, nor does it affect the updating of any non-regular
files. Implies –inplace, but does not conflict with –sparse
(since it is always extending a file’s length).

This works just like the –append option, but the existing data
on the receiving side is included in the full-file checksum ver‐
ification step, which will cause a file to be resent if the
final verification step fails (rsync uses a normal, non-append‐
ing –inplace transfer for the resend).

Note: prior to rsync 3.0.0, the –append option worked like
–append-verify, so if you are interacting with an older rsync
(or the transfer is using a protocol prior to 30), specifying
either append option will initiate an –append-verify transfer.

-d, –dirs
Tell the sending side to include any directories that are
encountered. Unlike –recursive, a directory’s contents are not
copied unless the directory name specified is “.” or ends with a
trailing slash (e.g. “.”, “dir/.”, “dir/”, etc.). Without this
option or the –recursive option, rsync will skip all directo‐
ries it encounters (and output a message to that effect for each
one). If you specify both –dirs and –recursive, –recursive
takes precedence.

The –dirs option is implied by the –files-from option or the
–list-only option (including an implied –list-only usage) if
–recursive wasn’t specified (so that directories are seen in
the listing). Specify –no-dirs (or –no-d) if you want to turn
this off.

There is also a backward-compatibility helper option, –old-dirs
(or –old-d) that tells rsync to use a hack of “-r
–exclude=’/*/*’” to get an older rsync to list a single direc‐
tory without recursing.

-l, –links
When symlinks are encountered, recreate the symlink on the des‐

-L, –copy-links
When symlinks are encountered, the item that they point to (the
referent) is copied, rather than the symlink. In older versions
of rsync, this option also had the side-effect of telling the
receiving side to follow symlinks, such as symlinks to directo‐
ries. In a modern rsync such as this one, you’ll need to spec‐
ify –keep-dirlinks (-K) to get this extra behavior. The only
exception is when sending files to an rsync that is too old to
understand -K — in that case, the -L option will still have the
side-effect of -K on that older receiving rsync.

This tells rsync to copy the referent of symbolic links that
point outside the copied tree. Absolute symlinks are also
treated like ordinary files, and so are any symlinks in the
source path itself when –relative is used. This option has no
additional effect if –copy-links was also specified.

This tells rsync to ignore any symbolic links which point out‐
side the copied tree. All absolute symlinks are also ignored.
Using this option in conjunction with –relative may give unex‐
pected results.

This option tells rsync to (1) modify all symlinks on the
receiving side in a way that makes them unusable but recoverable
(see below), or (2) to unmunge symlinks on the sending side that
had been stored in a munged state. This is useful if you don’t
quite trust the source of the data to not try to slip in a sym‐
link to a unexpected place.

The way rsync disables the use of symlinks is to prefix each one
with the string “/rsyncd-munged/”. This prevents the links from
being used as long as that directory does not exist. When this
option is enabled, rsync will refuse to run if that path is a
directory or a symlink to a directory.

The option only affects the client side of the transfer, so if
you need it to affect the server, specify it via
–remote-option. (Note that in a local transfer, the client
side is the sender.)

This option has no affect on a daemon, since the daemon config‐
ures whether it wants munged symlinks via its “munge symlinks”
parameter. See also the “munge-symlinks” perl script in the
support directory of the source code.

-k, –copy-dirlinks
This option causes the sending side to treat a symlink to a
directory as though it were a real directory. This is useful if
you don’t want symlinks to non-directories to be affected, as
they would be using –copy-links.

Without this option, if the sending side has replaced a direc‐
tory with a symlink to a directory, the receiving side will
delete anything that is in the way of the new symlink, including
a directory hierarchy (as long as –force or –delete is in

See also –keep-dirlinks for an analogous option for the receiv‐
ing side.

–copy-dirlinks applies to all symlinks to directories in the
source. If you want to follow only a few specified symlinks, a
trick you can use is to pass them as additional source args with
a trailing slash, using –relative to make the paths match up
right. For example:

rsync -r –relative src/./ src/./follow-me/ dest/

This works because rsync calls lstat on the source arg as
given, and the trailing slash makes lstat follow the symlink,
giving rise to a directory in the file-list which overrides the
symlink found during the scan of “src/./”.

-K, –keep-dirlinks
This option causes the receiving side to treat a symlink to a
directory as though it were a real directory, but only if it
matches a real directory from the sender. Without this option,
the receiver’s symlink would be deleted and replaced with a real

For example, suppose you transfer a directory “foo” that con‐
tains a file “file”, but “foo” is a symlink to directory “bar”
on the receiver. Without –keep-dirlinks, the receiver deletes
symlink “foo”, recreates it as a directory, and receives the
file into the new directory. With –keep-dirlinks, the receiver
keeps the symlink and “file” ends up in “bar”.

One note of caution: if you use –keep-dirlinks, you must trust
all the symlinks in the copy! If it is possible for an
untrusted user to create their own symlink to any directory, the
user could then (on a subsequent copy) replace the symlink with
a real directory and affect the content of whatever directory
the symlink references. For backup copies, you are better off
using something like a bind mount instead of a symlink to modify
your receiving hierarchy.

See also –copy-dirlinks for an analogous option for the sending

-H, –hard-links
This tells rsync to look for hard-linked files in the source and
link together the corresponding files on the destination. With‐
out this option, hard-linked files in the source are treated as
though they were separate files.

This option does NOT necessarily ensure that the pattern of hard
links on the destination exactly matches that on the source.
Cases in which the destination may end up with extra hard links
include the following:

o If the destination contains extraneous hard-links (more
linking than what is present in the source file list),
the copying algorithm will not break them explicitly.
However, if one or more of the paths have content differ‐
ences, the normal file-update process will break those
extra links (unless you are using the –inplace option).

o If you specify a –link-dest directory that contains hard
links, the linking of the destination files against the
–link-dest files can cause some paths in the destination
to become linked together due to the –link-dest associa‐

Note that rsync can only detect hard links between files that
are inside the transfer set. If rsync updates a file that has
extra hard-link connections to files outside the transfer, that
linkage will be broken. If you are tempted to use the –inplace
option to avoid this breakage, be very careful that you know how
your files are being updated so that you are certain that no
unintended changes happen due to lingering hard links (and see
the –inplace option for more caveats).

If incremental recursion is active (see –recursive), rsync may
transfer a missing hard-linked file before it finds that another
link for that contents exists elsewhere in the hierarchy. This
does not affect the accuracy of the transfer (i.e. which files
are hard-linked together), just its efficiency (i.e. copying the
data for a new, early copy of a hard-linked file that could have
been found later in the transfer in another member of the
hard-linked set of files). One way to avoid this inefficiency
is to disable incremental recursion using the –no-inc-recursive

-p, –perms
This option causes the receiving rsync to set the destination
permissions to be the same as the source permissions. (See also
the –chmod option for a way to modify what rsync considers to
be the source permissions.)

When this option is off, permissions are set as follows:

o Existing files (including updated files) retain their
existing permissions, though the –executability option
might change just the execute permission for the file.

o New files get their “normal” permission bits set to the
source file’s permissions masked with the receiving
directory’s default permissions (either the receiving
process’s umask, or the permissions specified via the
destination directory’s default ACL), and their special
permission bits disabled except in the case where a new
directory inherits a setgid bit from its parent direc‐

Thus, when –perms and –executability are both disabled,
rsync’s behavior is the same as that of other file-copy utili‐
ties, such as cp and tar.

In summary: to give destination files (both old and new) the
source permissions, use –perms. To give new files the destina‐
tion-default permissions (while leaving existing files
unchanged), make sure that the –perms option is off and use
–chmod=ugo=rwX (which ensures that all non-masked bits get
enabled). If you’d care to make this latter behavior easier to
type, you could define a popt alias for it, such as putting this
line in the file ~/.popt (the following defines the -Z option,
and includes –no-g to use the default group of the destination

rsync alias -Z –no-p –no-g –chmod=ugo=rwX

You could then use this new option in a command such as this

rsync -avZ src/ dest/

(Caveat: make sure that -a does not follow -Z, or it will
re-enable the two “–no-*” options mentioned above.)

The preservation of the destination’s setgid bit on newly-cre‐
ated directories when –perms is off was added in rsync 2.6.7.
Older rsync versions erroneously preserved the three special
permission bits for newly-created files when –perms was off,
while overriding the destination’s setgid bit setting on a
newly-created directory. Default ACL observance was added to
the ACL patch for rsync 2.6.7, so older (or non-ACL-enabled)
rsyncs use the umask even if default ACLs are present. (Keep in
mind that it is the version of the receiving rsync that affects
these behaviors.)

-E, –executability
This option causes rsync to preserve the executability (or
non-executability) of regular files when –perms is not enabled.
A regular file is considered to be executable if at least one
’x’ is turned on in its permissions. When an existing destina‐
tion file’s executability differs from that of the corresponding
source file, rsync modifies the destination file’s permissions
as follows:

o To make a file non-executable, rsync turns off all its
’x’ permissions.

o To make a file executable, rsync turns on each ’x’ per‐
mission that has a corresponding ’r’ permission enabled.

If –perms is enabled, this option is ignored.

-A, –acls
This option causes rsync to update the destination ACLs to be
the same as the source ACLs. The option also implies –perms.

The source and destination systems must have compatible ACL
entries for this option to work properly. See the –fake-super
option for a way to backup and restore ACLs that are not compat‐

-X, –xattrs
This option causes rsync to update the destination extended
attributes to be the same as the source ones.

For systems that support extended-attribute namespaces, a copy
being done by a super-user copies all namespaces except sys‐
tem.*. A normal user only copies the user.* namespace. To be
able to backup and restore non-user namespaces as a normal user,
see the –fake-super option.

Note that this option does not copy rsyncs special xattr values
(e.g. those used by –fake-super) unless you repeat the option
(e.g. -XX). This “copy all xattrs” mode cannot be used with

This option tells rsync to apply one or more comma-separated
“chmod” modes to the permission of the files in the transfer.
The resulting value is treated as though it were the permissions
that the sending side supplied for the file, which means that
this option can seem to have no effect on existing files if
–perms is not enabled.

In addition to the normal parsing rules specified in the
chmod manpage, you can specify an item that should only apply
to a directory by prefixing it with a ’D’, or specify an item
that should only apply to a file by prefixing it with a ’F’.
For example, the following will ensure that all directories get
marked set-gid, that no files are other-writable, that both are
user-writable and group-writable, and that both have consistent
executability across all bits:


Using octal mode numbers is also allowed:


It is also legal to specify multiple –chmod options, as each
additional option is just appended to the list of changes to

See the –perms and –executability options for how the result‐
ing permission value can be applied to the files in the trans‐

-o, –owner
This option causes rsync to set the owner of the destination
file to be the same as the source file, but only if the receiv‐
ing rsync is being run as the super-user (see also the –super
and –fake-super options). Without this option, the owner of
new and/or transferred files are set to the invoking user on the
receiving side.

The preservation of ownership will associate matching names by
default, but may fall back to using the ID number in some cir‐
cumstances (see also the –numeric-ids option for a full discus‐

-g, –group
This option causes rsync to set the group of the destination
file to be the same as the source file. If the receiving pro‐
gram is not running as the super-user (or if –no-super was
specified), only groups that the invoking user on the receiving
side is a member of will be preserved. Without this option, the
group is set to the default group of the invoking user on the
receiving side.

The preservation of group information will associate matching
names by default, but may fall back to using the ID number in
some circumstances (see also the –numeric-ids option for a full

This option causes rsync to transfer character and block device
files to the remote system to recreate these devices. This
option has no effect if the receiving rsync is not run as the
super-user (see also the –super and –fake-super options).

This option causes rsync to transfer special files such as named
sockets and fifos.

-D The -D option is equivalent to –devices –specials.

-t, –times
This tells rsync to transfer modification times along with the
files and update them on the remote system. Note that if this
option is not used, the optimization that excludes files that
have not been modified cannot be effective; in other words, a
missing -t or -a will cause the next transfer to behave as if it
used -I, causing all files to be updated (though rsync’s
delta-transfer algorithm will make the update fairly efficient
if the files haven’t actually changed, you’re much better off
using -t).

-O, –omit-dir-times
This tells rsync to omit directories when it is preserving modi‐
fication times (see –times). If NFS is sharing the directories
on the receiving side, it is a good idea to use -O. This option
is inferred if you use –backup without –backup-dir.

This option also has the side-effect of avoiding early creation
of directories in incremental recursion copies. The default
–inc-recursive copying normally does an early-create pass of
all the sub-directories in a parent directory in order for it to
be able to then set the modify time of the parent directory
right away (without having to delay that until a bunch of recur‐
sive copying has finished). This early-create idiom is not nec‐
essary if directory modify times are not being preserved, so it
is skipped. Since early-create directories don’t have accurate
mode, mtime, or ownership, the use of this option can help when
someone wants to avoid these partially-finished directories.

-J, –omit-link-times
This tells rsync to omit symlinks when it is preserving modifi‐
cation times (see –times).

This tells the receiving side to attempt super-user activities
even if the receiving rsync wasn’t run by the super-user. These
activities include: preserving users via the –owner option,
preserving all groups (not just the current user’s groups) via
the –groups option, and copying devices via the –devices
option. This is useful for systems that allow such activities
without being the super-user, and also for ensuring that you
will get errors if the receiving side isn’t being run as the
super-user. To turn off super-user activities, the super-user
can use –no-super.

When this option is enabled, rsync simulates super-user activi‐
ties by saving/restoring the privileged attributes via special
extended attributes that are attached to each file (as needed).
This includes the file’s owner and group (if it is not the
default), the file’s device info (device & special files are
created as empty text files), and any permission bits that we
won’t allow to be set on the real file (e.g. the real file gets
u-s,g-s,o-t for safety) or that would limit the owner’s access
(since the real super-user can always access/change a file, the
files we create can always be accessed/changed by the creating
user). This option also handles ACLs (if –acls was specified)
and non-user extended attributes (if –xattrs was specified).

This is a good way to backup data without using a super-user,
and to store ACLs from incompatible systems.

The –fake-super option only affects the side where the option
is used. To affect the remote side of a remote-shell connec‐
tion, use the –remote-option (-M) option:

rsync -av -M–fake-super /src/ host:/dest/

For a local copy, this option affects both the source and the
destination. If you wish a local copy to enable this option
just for the destination files, specify -M–fake-super. If you
wish a local copy to enable this option just for the source
files, combine –fake-super with -M–super.

This option is overridden by both –super and –no-super.

See also the “fake super” setting in the daemon’s rsyncd.conf

-S, –sparse
Try to handle sparse files efficiently so they take up less
space on the destination. Conflicts with –inplace because it’s
not possible to overwrite data in a sparse fashion.

This tells the receiver to allocate each destination file to its
eventual size before writing data to the file. Rsync will only
use the real filesystem-level preallocation support provided by
Linux’s fallocate system call or Cygwin’s posix_fallocate,
not the slow glibc implementation that writes a zero byte into
each block.

Without this option, larger files may not be entirely contiguous
on the filesystem, but with this option rsync will probably copy
more slowly. If the destination is not an extent-supporting
filesystem (such as ext4, xfs, NTFS, etc.), this option may have
no positive effect at all.

-n, –dry-run
This makes rsync perform a trial run that doesn’t make any
changes (and produces mostly the same output as a real run). It
is most commonly used in combination with the -v, –verbose
and/or -i, –itemize-changes options to see what an rsync com‐
mand is going to do before one actually runs it.

The output of –itemize-changes is supposed to be exactly the
same on a dry run and a subsequent real run (barring intentional
trickery and system call failures); if it isn’t, that’s a bug.
Other output should be mostly unchanged, but may differ in some
areas. Notably, a dry run does not send the actual data for
file transfers, so –progress has no effect, the “bytes sent”,
“bytes received”, “literal data”, and “matched data” statistics
are too small, and the “speedup” value is equivalent to a run
where no file transfers were needed.

-W, –whole-file
With this option rsync’s delta-transfer algorithm is not used
and the whole file is sent as-is instead. The transfer may be
faster if this option is used when the bandwidth between the
source and destination machines is higher than the bandwidth to
disk (especially when the “disk” is actually a networked
filesystem). This is the default when both the source and des‐
tination are specified as local paths, but only if no
batch-writing option is in effect.

-x, –one-file-system
This tells rsync to avoid crossing a filesystem boundary when
recursing. This does not limit the user’s ability to specify
items to copy from multiple filesystems, just rsync’s recursion
through the hierarchy of each directory that the user specified,
and also the analogous recursion on the receiving side during
deletion. Also keep in mind that rsync treats a “bind” mount to
the same device as being on the same filesystem.

If this option is repeated, rsync omits all mount-point directo‐
ries from the copy. Otherwise, it includes an empty directory
at each mount-point it encounters (using the attributes of the
mounted directory because those of the underlying mount-point
directory are inaccessible).

If rsync has been told to collapse symlinks (via –copy-links or
–copy-unsafe-links), a symlink to a directory on another device
is treated like a mount-point. Symlinks to non-directories are
unaffected by this option.

–existing, –ignore-non-existing
This tells rsync to skip creating files (including directories)
that do not exist yet on the destination. If this option is
combined with the –ignore-existing option, no files will be
updated (which can be useful if all you want to do is delete
extraneous files).

This option is a transfer rule, not an exclude, so it doesn’t
affect the data that goes into the file-lists, and thus it
doesn’t affect deletions. It just limits the files that the
receiver requests to be transferred.

This tells rsync to skip updating files that already exist on
the destination (this does not ignore existing directories, or
nothing would get done). See also –existing.

This option is a transfer rule, not an exclude, so it doesn’t
affect the data that goes into the file-lists, and thus it
doesn’t affect deletions. It just limits the files that the
receiver requests to be transferred.

This option can be useful for those doing backups using the
–link-dest option when they need to continue a backup run that
got interrupted. Since a –link-dest run is copied into a new
directory hierarchy (when it is used properly), using –ignore
existing will ensure that the already-handled files don’t get
tweaked (which avoids a change in permissions on the hard-linked
files). This does mean that this option is only looking at the
existing files in the destination hierarchy itself.

This tells rsync to remove from the sending side the files
(meaning non-directories) that are a part of the transfer and
have been successfully duplicated on the receiving side.

Note that you should only use this option on source files that
are quiescent. If you are using this to move files that show up
in a particular directory over to another host, make sure that
the finished files get renamed into the source directory, not
directly written into it, so that rsync can’t possibly transfer
a file that is not yet fully written. If you can’t first write
the files into a different directory, you should use a naming
idiom that lets rsync avoid transferring files that are not yet
finished (e.g. name the file “” when it is written,
rename it to “foo” when it is done, and then use the option
–exclude=’*.new’ for the rsync transfer).

Starting with 3.1.0, rsync will skip the sender-side removal
(and output an error) if the file’s size or modify time has not
stayed unchanged.

This tells rsync to delete extraneous files from the receiving
side (ones that aren’t on the sending side), but only for the
directories that are being synchronized. You must have asked
rsync to send the whole directory (e.g. “dir” or “dir/”) without
using a wildcard for the directory’s contents (e.g. “dir/*”)
since the wildcard is expanded by the shell and rsync thus gets
a request to transfer individual files, not the files’ parent
directory. Files that are excluded from the transfer are also
excluded from being deleted unless you use the –delete-excluded
option or mark the rules as only matching on the sending side
(see the include/exclude modifiers in the FILTER RULES section).

Prior to rsync 2.6.7, this option would have no effect unless
–recursive was enabled. Beginning with 2.6.7, deletions will
also occur when –dirs (-d) is enabled, but only for directories
whose contents are being copied.

This option can be dangerous if used incorrectly! It is a very
good idea to first try a run using the –dry-run option (-n) to
see what files are going to be deleted.

If the sending side detects any I/O errors, then the deletion of
any files at the destination will be automatically disabled.
This is to prevent temporary filesystem failures (such as NFS
errors) on the sending side from causing a massive deletion of
files on the destination. You can override this with the
–ignore-errors option.

The –delete option may be combined with one of the
–delete-WHEN options without conflict, as well as
–delete-excluded. However, if none of the –delete-WHEN
options are specified, rsync will choose the –delete-during
algorithm when talking to rsync 3.0.0 or newer, and the
–delete-before algorithm when talking to an older rsync. See
also –delete-delay and –delete-after.

Request that the file-deletions on the receiving side be done
before the transfer starts. See –delete (which is implied) for
more details on file-deletion.

Deleting before the transfer is helpful if the filesystem is
tight for space and removing extraneous files would help to make
the transfer possible. However, it does introduce a delay
before the start of the transfer, and this delay might cause the
transfer to timeout (if –timeout was specified). It also
forces rsync to use the old, non-incremental recursion algorithm
that requires rsync to scan all the files in the transfer into
memory at once (see –recursive).

–delete-during, –del
Request that the file-deletions on the receiving side be done
incrementally as the transfer happens. The per-directory delete
scan is done right before each directory is checked for updates,
so it behaves like a more efficient –delete-before, including
doing the deletions prior to any per-directory filter files
being updated. This option was first added in rsync version
2.6.4. See –delete (which is implied) for more details on

Request that the file-deletions on the receiving side be com‐
puted during the transfer (like –delete-during), and then
removed after the transfer completes. This is useful when com‐
bined with –delay-updates and/or –fuzzy, and is more efficient
than using –delete-after (but can behave differently, since
–delete-after computes the deletions in a separate pass after
all updates are done). If the number of removed files overflows
an internal buffer, a temporary file will be created on the
receiving side to hold the names (it is removed while open, so
you shouldn’t see it during the transfer). If the creation of
the temporary file fails, rsync will try to fall back to using
–delete-after (which it cannot do if –recursive is doing an
incremental scan). See –delete (which is implied) for more
details on file-deletion.

Request that the file-deletions on the receiving side be done
after the transfer has completed. This is useful if you are
sending new per-directory merge files as a part of the transfer
and you want their exclusions to take effect for the delete
phase of the current transfer. It also forces rsync to use the
old, non-incremental recursion algorithm that requires rsync to
scan all the files in the transfer into memory at once (see
–recursive). See –delete (which is implied) for more details
on file-deletion.

In addition to deleting the files on the receiving side that are
not on the sending side, this tells rsync to also delete any
files on the receiving side that are excluded (see –exclude).
See the FILTER RULES section for a way to make individual exclu‐
sions behave this way on the receiver, and for a way to protect
files from –delete-excluded. See –delete (which is implied)
for more details on file-deletion.

When rsync is first processing the explicitly requested source
files (e.g. command-line arguments or –files-from entries), it
is normally an error if the file cannot be found. This option
suppresses that error, and does not try to transfer the file.
This does not affect subsequent vanished-file errors if a file
was initially found to be present and later is no longer there.

This option takes the behavior of (the implied) –ignore-miss‐
ing-args option a step farther: each missing arg will become a
deletion request of the corresponding destination file on the
receiving side (should it exist). If the destination file is a
non-empty directory, it will only be successfully deleted if
–force or –delete are in effect. Other than that, this option
is independent of any other type of delete processing.

The missing source files are represented by special file-list
entries which display as a “*missing” entry in the –list-only

Tells –delete to go ahead and delete files even when there are
I/O errors.

This option tells rsync to delete a non-empty directory when it
is to be replaced by a non-directory. This is only relevant if
deletions are not active (see –delete for details).

Note for older rsync versions: –force used to still be required
when using –delete-after, and it used to be non-functional
unless the –recursive option was also enabled.

This tells rsync not to delete more than NUM files or directo‐
ries. If that limit is exceeded, all further deletions are
skipped through the end of the transfer. At the end, rsync out‐
puts a warning (including a count of the skipped deletions) and
exits with an error code of 25 (unless some more important error
condition also occurred).

Beginning with version 3.0.0, you may specify –max-delete=0 to
be warned about any extraneous files in the destination without
removing any of them. Older clients interpreted this as “unlim‐
ited”, so if you don’t know what version the client is, you can
use the less obvious –max-delete=-1 as a backward-compatible
way to specify that no deletions be allowed (though really old
versions didn’t warn when the limit was exceeded).

This tells rsync to avoid transferring any file that is larger
than the specified SIZE. The SIZE value can be suffixed with a
string to indicate a size multiplier, and may be a fractional
value (e.g. “–max-size=1.5m”).

This option is a transfer rule, not an exclude, so it doesn’t
affect the data that goes into the file-lists, and thus it
doesn’t affect deletions. It just limits the files that the
receiver requests to be transferred.

The suffixes are as follows: “K” (or “KiB”) is a kibibyte
(1024), “M” (or “MiB”) is a mebibyte (1024*1024), and “G” (or
“GiB”) is a gibibyte (1024*1024*1024). If you want the multi‐
plier to be 1000 instead of 1024, use “KB”, “MB”, or “GB”.
(Note: lower-case is also accepted for all values.) Finally, if
the suffix ends in either “+1” or “-1”, the value will be offset
by one byte in the indicated direction.

Examples: –max-size=1.5mb-1 is 1499999 bytes, and
–max-size=2g+1 is 2147483649 bytes.

Note that rsync versions prior to 3.1.0 did not allow

This tells rsync to avoid transferring any file that is smaller
than the specified SIZE, which can help in not transferring
small, junk files. See the –max-size option for a description
of SIZE and other information.

Note that rsync versions prior to 3.1.0 did not allow

-B, –block-size=BLOCKSIZE
This forces the block size used in rsync’s delta-transfer algo‐
rithm to a fixed value. It is normally selected based on the
size of each file being updated. See the technical report for

-e, –rsh=COMMAND
This option allows you to choose an alternative remote shell
program to use for communication between the local and remote
copies of rsync. Typically, rsync is configured to use ssh by
default, but you may prefer to use rsh on a local network.

If this option is used with [user@]host::module/path, then the
remote shell COMMAND will be used to run an rsync daemon on the
remote host, and all data will be transmitted through that
remote shell connection, rather than through a direct socket
connection to a running rsync daemon on the remote host. See
NECTION” above.

Command-line arguments are permitted in COMMAND provided that
COMMAND is presented to rsync as a single argument. You must
use spaces (not tabs or other whitespace) to separate the com‐
mand and args from each other, and you can use single- and/or
double-quotes to preserve spaces in an argument (but not back‐
slashes). Note that doubling a single-quote inside a sin‐
gle-quoted string gives you a single-quote; likewise for dou‐
ble-quotes (though you need to pay attention to which quotes
your shell is parsing and which quotes rsync is parsing). Some

-e ‘ssh -p 2234’
-e ‘ssh -o “ProxyCommand nohup ssh firewall nc -w1 %h %p”‘

(Note that ssh users can alternately customize site-specific
connect options in their .ssh/config file.)

You can also choose the remote shell program using the RSYNC_RSH
environment variable, which accepts the same range of values as

See also the –blocking-io option which is affected by this

Use this to specify what program is to be run on the remote
machine to start-up rsync. Often used when rsync is not in the
default remote-shell’s path (e.g.
–rsync-path=/usr/local/bin/rsync). Note that PROGRAM is run
with the help of a shell, so it can be any program, script, or
command sequence you’d care to run, so long as it does not cor‐
rupt the standard-in & standard-out that rsync is using to com‐

One tricky example is to set a different default directory on
the remote machine for use with the –relative option. For

rsync -avR –rsync-path=”cd /a/b && rsync” host:c/d /e/

-M, –remote-option=OPTION
This option is used for more advanced situations where you want
certain effects to be limited to one side of the transfer only.
For instance, if you want to pass –log-file=FILE and
–fake-super to the remote system, specify it like this:

rsync -av -M –log-file=foo -M–fake-super src/ dest/

If you want to have an option affect only the local side of a
transfer when it normally affects both sides, send its negation
to the remote side. Like this:

rsync -av -x -M–no-x src/ dest/

Be cautious using this, as it is possible to toggle an option
that will cause rsync to have a different idea about what data
to expect next over the socket, and that will make it fail in a
cryptic fashion.

Note that it is best to use a separate –remote-option for each
option you want to pass. This makes your useage compatible with
the –protect-args option. If that option is off, any spaces in
your remote options will be split by the remote shell unless you
take steps to protect them.

When performing a local transfer, the “local” side is the sender
and the “remote” side is the receiver.

Note some versions of the popt option-parsing library have a bug
in them that prevents you from using an adjacent arg with an
equal in it next to a short option letter (e.g.
-M–log-file=/tmp/foo. If this bug affects your version of
popt, you can use the version of popt that is included with

-C, –cvs-exclude
This is a useful shorthand for excluding a broad range of files
that you often don’t want to transfer between systems. It uses a
similar algorithm to CVS to determine if a file should be

The exclude list is initialized to exclude the following items
(these initial items are marked as perishable — see the FILTER
RULES section):

RCS SCCS CVS CVS.adm RCSLOG cvslog.* tags TAGS
.make.state .nse_depinfo *~ #* .#* ,* _$* *$ *.old *.bak
*.BAK *.orig *.rej .del-* *.a *.olb *.o *.obj *.so *.exe
*.Z *.elc *.ln core .svn/ .git/ .hg/ .bzr/

then, files listed in a $HOME/.cvsignore are added to the list
and any files listed in the CVSIGNORE environment variable (all
cvsignore names are delimited by whitespace).

Finally, any file is ignored if it is in the same directory as a
.cvsignore file and matches one of the patterns listed therein.
Unlike rsync’s filter/exclude files, these patterns are split on
whitespace. See the cvs(1) manual for more information.

If you’re combining -C with your own –filter rules, you should
note that these CVS excludes are appended at the end of your own
rules, regardless of where the -C was placed on the com‐
mand-line. This makes them a lower priority than any rules you
specified explicitly. If you want to control where these CVS
excludes get inserted into your filter rules, you should omit
the -C as a command-line option and use a combination of –fil‐
ter=:C and –filter=-C (either on your command-line or by
putting the “:C” and “-C” rules into a filter file with your
other rules). The first option turns on the per-directory scan‐
ning for the .cvsignore file. The second option does a one-time
import of the CVS excludes mentioned above.

-f, –filter=RULE
This option allows you to add rules to selectively exclude cer‐
tain files from the list of files to be transferred. This is
most useful in combination with a recursive transfer.

You may use as many –filter options on the command line as you
like to build up the list of files to exclude. If the filter
contains whitespace, be sure to quote it so that the shell gives
the rule to rsync as a single argument. The text below also
mentions that you can use an underscore to replace the space
that separates a rule from its arg.

See the FILTER RULES section for detailed information on this

-F The -F option is a shorthand for adding two –filter rules to
your command. The first time it is used is a shorthand for this

–filter=’dir-merge /.rsync-filter’

This tells rsync to look for per-directory .rsync-filter files
that have been sprinkled through the hierarchy and use their
rules to filter the files in the transfer. If -F is repeated,
it is a shorthand for this rule:

–filter=’exclude .rsync-filter’

This filters out the .rsync-filter files themselves from the

See the FILTER RULES section for detailed information on how
these options work.

This option is a simplified form of the –filter option that
defaults to an exclude rule and does not allow the full
rule-parsing syntax of normal filter rules.

See the FILTER RULES section for detailed information on this

This option is related to the –exclude option, but it specifies
a FILE that contains exclude patterns (one per line). Blank
lines in the file and lines starting with ’;’ or ’#’ are
ignored. If FILE is -, the list will be read from standard

This option is a simplified form of the –filter option that
defaults to an include rule and does not allow the full
rule-parsing syntax of normal filter rules.

See the FILTER RULES section for detailed information on this

This option is related to the –include option, but it specifies
a FILE that contains include patterns (one per line). Blank
lines in the file and lines starting with ’;’ or ’#’ are
ignored. If FILE is -, the list will be read from standard

Using this option allows you to specify the exact list of files
to transfer (as read from the specified FILE or – for standard
input). It also tweaks the default behavior of rsync to make
transferring just the specified files and directories easier:

o The –relative (-R) option is implied, which preserves
the path information that is specified for each item in
the file (use –no-relative or –no-R if you want to turn
that off).

o The –dirs (-d) option is implied, which will create
directories specified in the list on the destination
rather than noisily skipping them (use –no-dirs or
–no-d if you want to turn that off).

o The –archive (-a) option’s behavior does not imply
–recursive (-r), so specify it explicitly, if you want

o These side-effects change the default state of rsync, so
the position of the –files-from option on the com‐
mand-line has no bearing on how other options are parsed
(e.g. -a works the same before or after –files-from, as
does –no-R and all other options).

The filenames that are read from the FILE are all relative to
the source dir — any leading slashes are removed and no “..”
references are allowed to go higher than the source dir. For
example, take this command:

rsync -a –files-from=/tmp/foo /usr remote:/backup

If /tmp/foo contains the string “bin” (or even “/bin”), the
/usr/bin directory will be created as /backup/bin on the remote
host. If it contains “bin/” (note the trailing slash), the
immediate contents of the directory would also be sent (without
needing to be explicitly mentioned in the file — this began in
version 2.6.4). In both cases, if the -r option was enabled,
that dir’s entire hierarchy would also be transferred (keep in
mind that -r needs to be specified explicitly with –files-from,
since it is not implied by -a). Also note that the effect of
the (enabled by default) –relative option is to duplicate only
the path info that is read from the file — it does not force
the duplication of the source-spec path (/usr in this case).

In addition, the –files-from file can be read from the remote
host instead of the local host if you specify a “host:” in front
of the file (the host must match one end of the transfer). As a
short-cut, you can specify just a prefix of “:” to mean “use the
remote end of the transfer”. For example:

rsync -a –files-from=:/path/file-list src:/ /tmp/copy

This would copy all the files specified in the /path/file-list
file that was located on the remote “src” host.

If the –iconv and –protect-args options are specified and the
–files-from filenames are being sent from one host to another,
the filenames will be translated from the sending host’s charset
to the receiving host’s charset.

NOTE: sorting the list of files in the –files-from input helps
rsync to be more efficient, as it will avoid re-visiting the
path elements that are shared between adjacent entries. If the
input is not sorted, some path elements (implied directories)
may end up being scanned multiple times, and rsync will eventu‐
ally unduplicate them after they get turned into file-list ele‐

-0, –from0
This tells rsync that the rules/filenames it reads from a file
are terminated by a null (’\0’) character, not a NL, CR, or
CR+LF. This affects –exclude-from, –include-from,
–files-from, and any merged files specified in a –filter rule.
It does not affect –cvs-exclude (since all names read from a
.cvsignore file are split on whitespace).

-s, –protect-args
This option sends all filenames and most options to the remote
rsync without allowing the remote shell to interpret them. This
means that spaces are not split in names, and any non-wildcard
special characters are not translated (such as ~, $, ;, &,
etc.). Wildcards are expanded on the remote host by rsync
(instead of the shell doing it).

If you use this option with –iconv, the args related to the
remote side will also be translated from the local to the remote
character-set. The translation happens before wild-cards are
expanded. See also the –files-from option.

You may also control this option via the RSYNC_PROTECT_ARGS
environment variable. If this variable has a non-zero value,
this option will be enabled by default, otherwise it will be
disabled by default. Either state is overridden by a manually
specified positive or negative version of this option (note that
–no-s and –no-protect-args are the negative versions). Since
this option was first introduced in 3.0.0, you’ll need to make
sure it’s disabled if you ever need to interact with a remote
rsync that is older than that.

Rsync can also be configured (at build time) to have this option
enabled by default (with is overridden by both the environment
and the command-line). This option will eventually become a new
default setting at some as-yet-undetermined point in the future.

-T, –temp-dir=DIR
This option instructs rsync to use DIR as a scratch directory
when creating temporary copies of the files transferred on the
receiving side. The default behavior is to create each tempo‐
rary file in the same directory as the associated destination
file. Beginning with rsync 3.1.1, the temp-file names inside
the specified DIR will not be prefixed with an extra dot (though
they will still have a random suffix added).

This option is most often used when the receiving disk partition
does not have enough free space to hold a copy of the largest
file in the transfer. In this case (i.e. when the scratch
directory is on a different disk partition), rsync will not be
able to rename each received temporary file over the top of the
associated destination file, but instead must copy it into
place. Rsync does this by copying the file over the top of the
destination file, which means that the destination file will
contain truncated data during this copy. If this were not done
this way (even if the destination file were first removed, the
data locally copied to a temporary file in the destination
directory, and then renamed into place) it would be possible for
the old file to continue taking up disk space (if someone had it
open), and thus there might not be enough room to fit the new
version on the disk at the same time.

If you are using this option for reasons other than a shortage
of disk space, you may wish to combine it with the
–delay-updates option, which will ensure that all copied files
get put into subdirectories in the destination hierarchy, await‐
ing the end of the transfer. If you don’t have enough room to
duplicate all the arriving files on the destination partition,
another way to tell rsync that you aren’t overly concerned about
disk space is to use the –partial-dir option with a relative
path; because this tells rsync that it is OK to stash off a copy
of a single file in a subdir in the destination hierarchy, rsync
will use the partial-dir as a staging area to bring over the
copied file, and then rename it into place from there. (Specify‐
ing a –partial-dir with an absolute path does not have this

-y, –fuzzy
This option tells rsync that it should look for a basis file for
any destination file that is missing. The current algorithm
looks in the same directory as the destination file for either a
file that has an identical size and modified-time, or a simi‐
larly-named file. If found, rsync uses the fuzzy basis file to
try to speed up the transfer.

If the option is repeated, the fuzzy scan will also be done in
any matching alternate destination directories that are speci‐
fied via –compare-dest, –copy-dest, or –link-dest.

Note that the use of the –delete option might get rid of any
potential fuzzy-match files, so either use –delete-after or
specify some filename exclusions if you need to prevent this.

This option instructs rsync to use DIR on the destination
machine as an additional hierarchy to compare destination files
against doing transfers (if the files are missing in the desti‐
nation directory). If a file is found in DIR that is identical
to the sender’s file, the file will NOT be transferred to the
destination directory. This is useful for creating a sparse
backup of just files that have changed from an earlier backup.
This option is typically used to copy into an empty (or newly
created) directory.

Beginning in version 2.6.4, multiple –compare-dest directories
may be provided, which will cause rsync to search the list in
the order specified for an exact match. If a match is found
that differs only in attributes, a local copy is made and the
attributes updated. If a match is not found, a basis file from
one of the DIRs will be selected to try to speed up the trans‐

If DIR is a relative path, it is relative to the destination
directory. See also –copy-dest and –link-dest.

NOTE: beginning with version 3.1.0, rsync will remove a file
from a non-empty destination hierarchy if an exact match is
found in one of the compare-dest hierarchies (making the end
result more closely match a fresh copy).

This option behaves like –compare-dest, but rsync will also
copy unchanged files found in DIR to the destination directory
using a local copy. This is useful for doing transfers to a new
destination while leaving existing files intact, and then doing
a flash-cutover when all files have been successfully trans‐

Multiple –copy-dest directories may be provided, which will
cause rsync to search the list in the order specified for an
unchanged file. If a match is not found, a basis file from one
of the DIRs will be selected to try to speed up the transfer.

If DIR is a relative path, it is relative to the destination
directory. See also –compare-dest and –link-dest.

This option behaves like –copy-dest, but unchanged files are
hard linked from DIR to the destination directory. The files
must be identical in all preserved attributes (e.g. permissions,
possibly ownership) in order for the files to be linked
together. An example:

rsync -av –link-dest=$PWD/prior_dir host:src_dir/ new_dir/

If file’s aren’t linking, double-check their attributes. Also
check if some attributes are getting forced outside of rsync’s
control, such a mount option that squishes root to a single
user, or mounts a removable drive with generic ownership (such
as OS X’s “Ignore ownership on this volume” option).

Beginning in version 2.6.4, multiple –link-dest directories may
be provided, which will cause rsync to search the list in the
order specified for an exact match. If a match is found that
differs only in attributes, a local copy is made and the
attributes updated. If a match is not found, a basis file from
one of the DIRs will be selected to try to speed up the trans‐

This option works best when copying into an empty destination
hierarchy, as existing files may get their attributes tweaked,
and that can affect alternate destination files via hard-links.
Also, itemizing of changes can get a bit muddled. Note that
prior to version 3.1.0, an alternate-directory exact match would
never be found (nor linked into the destination) when a destina‐
tion file already exists.

Note that if you combine this option with –ignore-times, rsync
will not link any files together because it only links identical
files together as a substitute for transferring the file, never
as an additional check after the file is updated.

If DIR is a relative path, it is relative to the destination
directory. See also –compare-dest and –copy-dest.

Note that rsync versions prior to 2.6.1 had a bug that could
prevent –link-dest from working properly for a non-super-user
when -o was specified (or implied by -a). You can work-around
this bug by avoiding the -o option when sending to an old rsync.

-z, –compress
With this option, rsync compresses the file data as it is sent
to the destination machine, which reduces the amount of data
being transmitted — something that is useful over a slow con‐

Note that this option typically achieves better compression
ratios than can be achieved by using a compressing remote shell
or a compressing transport because it takes advantage of the
implicit information in the matching data blocks that are not
explicitly sent over the connection. This matching-data com‐
pression comes at a cost of CPU, though, and can be disabled by
repeating the -z option, but only if both sides are at least
version 3.1.1.

Note that if your version of rsync was compiled with an external
zlib (instead of the zlib that comes packaged with rsync) then
it will not support the old-style compression, only the
new-style (repeated-option) compression. In the future this
new-style compression will likely become the default.

The client rsync requests new-style compression on the server
via the –new-compress option, so if you see that option
rejected it means that the server is not new enough to support
-zz. Rsync also accepts the –old-compress option for a future
time when new-style compression becomes the default.

See the –skip-compress option for the default list of file suf‐
fixes that will not be compressed.

Explicitly set the compression level to use (see –compress)
instead of letting it default. If NUM is non-zero, the –com‐
press option is implied.

Override the list of file suffixes that will not be compressed.
The LIST should be one or more file suffixes (without the dot)
separated by slashes (/).

You may specify an empty string to indicate that no file should
be skipped.

Simple character-class matching is supported: each must consist
of a list of letters inside the square brackets (e.g. no special
classes, such as “[:alpha:]”, are supported, and ’-’ has no spe‐
cial meaning).

The characters asterisk (*) and question-mark (?) have no spe‐
cial meaning.

Here’s an example that specifies 6 suffixes to skip (since 1 of
the 5 rules matches 2 suffixes):


The default list of suffixes that will not be compressed is this
(in this version of rsync):

7z ace avi bz2 deb gpg gz iso jpeg jpg lz lzma lzo mov mp3 mp4
ogg png rar rpm rzip tbz tgz tlz txz xz z zip

This list will be replaced by your –skip-compress list in all
but one situation: a copy from a daemon rsync will add your
skipped suffixes to its list of non-compressing files (and its
list may be configured to a different default).

With this option rsync will transfer numeric group and user IDs
rather than using user and group names and mapping them at both

By default rsync will use the username and groupname to deter‐
mine what ownership to give files. The special uid 0 and the
special group 0 are never mapped via user/group names even if
the –numeric-ids option is not specified.

If a user or group has no name on the source system or it has no
match on the destination system, then the numeric ID from the
source system is used instead. See also the comments on the
“use chroot” setting in the rsyncd.conf manpage for information
on how the chroot setting affects rsync’s ability to look up the
names of the users and groups and what you can do about it.

–usermap=STRING, –groupmap=STRING
These options allow you to specify users and groups that should
be mapped to other values by the receiving side. The STRING is
one or more FROM:TO pairs of values separated by commas. Any
matching FROM value from the sender is replaced with a TO value
from the receiver. You may specify usernames or user IDs for
the FROM and TO values, and the FROM value may also be a
wild-card string, which will be matched against the sender’s
names (wild-cards do NOT match against ID numbers, though see
below for why a ’*’ matches everything). You may instead spec‐
ify a range of ID numbers via an inclusive range: LOW-HIGH. For

–usermap=0-99:nobody,wayne:admin,*:normal –groupmap=usr:1,1:usr

The first match in the list is the one that is used. You should
specify all your user mappings using a single –usermap option,
and/or all your group mappings using a single –groupmap option.

Note that the sender’s name for the 0 user and group are not
transmitted to the receiver, so you should either match these
values using a 0, or use the names in effect on the receiving
side (typically “root”). All other FROM names match those in
use on the sending side. All TO names match those in use on the
receiving side.

Any IDs that do not have a name on the sending side are treated
as having an empty name for the purpose of matching. This
allows them to be matched via a “*” or using an empty name. For

–usermap=:nobody –groupmap=*:nobody

When the –numeric-ids option is used, the sender does not send
any names, so all the IDs are treated as having an empty name.
This means that you will need to specify numeric FROM values if
you want to map these nameless IDs to different values.

For the –usermap option to have any effect, the -o (–owner)
option must be used (or implied), and the receiver will need to
be running as a super-user (see also the –fake-super option).
For the –groupmap option to have any effect, the -g (–groups)
option must be used (or implied), and the receiver will need to
have permissions to set that group.

This option forces all files to be owned by USER with group
GROUP. This is a simpler interface than using –usermap and
–groupmap directly, but it is implemented using those options
internally, so you cannot mix them. If either the USER or GROUP
is empty, no mapping for the omitted user/group will occur. If
GROUP is empty, the trailing colon may be omitted, but if USER
is empty, a leading colon must be supplied.

If you specify “–chown=foo:bar, this is exactly the same as
specifying “–usermap=*:foo –groupmap=*:bar”, only easier.

This option allows you to set a maximum I/O timeout in seconds.
If no data is transferred for the specified time then rsync will
exit. The default is 0, which means no timeout.

This option allows you to set the amount of time that rsync will
wait for its connection to an rsync daemon to succeed. If the
timeout is reached, rsync exits with an error.

By default rsync will bind to the wildcard address when connect‐
ing to an rsync daemon. The –address option allows you to
specify a specific IP address (or hostname) to bind to. See
also this option in the –daemon mode section.

This specifies an alternate TCP port number to use rather than
the default of 873. This is only needed if you are using the
double-colon (::) syntax to connect with an rsync daemon (since
the URL syntax has a way to specify the port as a part of the
URL). See also this option in the –daemon mode section.

This option can provide endless fun for people who like to tune
their systems to the utmost degree. You can set all sorts of
socket options which may make transfers faster (or slower!).
Read the man page for the setsockopt() system call for details
on some of the options you may be able to set. By default no
special socket options are set. This only affects direct socket
connections to a remote rsync daemon. This option also exists
in the –daemon mode section.

This tells rsync to use blocking I/O when launching a remote
shell transport. If the remote shell is either rsh or remsh,
rsync defaults to using blocking I/O, otherwise it defaults to
using non-blocking I/O. (Note that ssh prefers non-blocking

This sets the output buffering mode. The mode can be None (aka
Unbuffered), Line, or Block (aka Full). You may specify as lit‐
tle as a single letter for the mode, and use upper or lower

The main use of this option is to change Full buffering to Line
buffering when rsync’s output is going to a file or pipe.

-i, –itemize-changes
Requests a simple itemized list of the changes that are being
made to each file, including attribute changes. This is exactly
the same as specifying –out-format=’%i %n%L’. If you repeat
the option, unchanged files will also be output, but only if the
receiving rsync is at least version 2.6.7 (you can use -vv with
older versions of rsync, but that also turns on the output of
other verbose messages).

The “%i” escape has a cryptic output that is 11 letters long.
The general format is like the string YXcstpoguax, where Y is
replaced by the type of update being done, X is replaced by the
file-type, and the other letters represent attributes that may
be output if they are being modified.

The update types that replace the Y are as follows:

o A < means that a file is being transferred to the remote host (sent). o A > means that a file is being transferred to the local
host (received).

o A c means that a local change/creation is occurring for
the item (such as the creation of a directory or the
changing of a symlink, etc.).

o A h means that the item is a hard link to another item
(requires –hard-links).

o A . means that the item is not being updated (though it
might have attributes that are being modified).

o A * means that the rest of the itemized-output area con‐
tains a message (e.g. “deleting”).

The file-types that replace the X are: f for a file, a d for a
directory, an L for a symlink, a D for a device, and a S for a
special file (e.g. named sockets and fifos).

The other letters in the string above are the actual letters
that will be output if the associated attribute for the item is
being updated or a “.” for no change. Three exceptions to this
are: (1) a newly created item replaces each letter with a “+”,
(2) an identical item replaces the dots with spaces, and (3) an
unknown attribute replaces each letter with a “?” (this can hap‐
pen when talking to an older rsync).

The attribute that is associated with each letter is as follows:

o A c means either that a regular file has a different
checksum (requires –checksum) or that a symlink, device,
or special file has a changed value. Note that if you
are sending files to an rsync prior to 3.0.1, this change
flag will be present only for checksum-differing regular

o A s means the size of a regular file is different and
will be updated by the file transfer.

o A t means the modification time is different and is being
updated to the sender’s value (requires –times). An
alternate value of T means that the modification time
will be set to the transfer time, which happens when a
file/symlink/device is updated without –times and when a
symlink is changed and the receiver can’t set its time.
(Note: when using an rsync 3.0.0 client, you might see
the s flag combined with t instead of the proper T flag
for this time-setting failure.)

o A p means the permissions are different and are being
updated to the sender’s value (requires –perms).

o An o means the owner is different and is being updated to
the sender’s value (requires –owner and super-user priv‐

o A g means the group is different and is being updated to
the sender’s value (requires –group and the authority to
set the group).

o The u slot is reserved for future use.

o The a means that the ACL information changed.

o The x means that the extended attribute information

One other output is possible: when deleting files, the “%i”
will output the string “*deleting” for each item that is being
removed (assuming that you are talking to a recent enough rsync
that it logs deletions instead of outputting them as a verbose

This allows you to specify exactly what the rsync client outputs
to the user on a per-update basis. The format is a text string
containing embedded single-character escape sequences prefixed
with a percent (%) character. A default format of “%n%L” is
assumed if either –info=name or -v is specified (this tells you
just the name of the file and, if the item is a link, where it
points). For a full list of the possible escape characters, see
the “log format” setting in the rsyncd.conf manpage.

Specifying the –out-format option implies the –info=name
option, which will mention each file, dir, etc. that gets
updated in a significant way (a transferred file, a recreated
symlink/device, or a touched directory). In addition, if the
itemize-changes escape (%i) is included in the string (e.g. if
the –itemize-changes option was used), the logging of names
increases to mention any item that is changed in any way (as
long as the receiving side is at least 2.6.4). See the –item‐
ize-changes option for a description of the output of “%i”.

Rsync will output the out-format string prior to a file’s trans‐
fer unless one of the transfer-statistic escapes is requested,
in which case the logging is done at the end of the file’s
transfer. When this late logging is in effect and –progress is
also specified, rsync will also output the name of the file
being transferred prior to its progress information (followed,
of course, by the out-format output).

This option causes rsync to log what it is doing to a file.
This is similar to the logging that a daemon does, but can be
requested for the client side and/or the server side of a
non-daemon transfer. If specified as a client option, transfer
logging will be enabled with a default format of “%i %n%L”. See
the –log-file-format option if you wish to override this.

Here’s a example command that requests the remote side to log
what is happening:

rsync -av –remote-option=–log-file=/tmp/rlog src/ dest/

This is very useful if you need to debug why a connection is
closing unexpectedly.

This allows you to specify exactly what per-update logging is
put into the file specified by the –log-file option (which must
also be specified for this option to have any effect). If you
specify an empty string, updated files will not be mentioned in
the log file. For a list of the possible escape characters, see
the “log format” setting in the rsyncd.conf manpage.

The default FORMAT used if –log-file is specified and this
option is not is ’%i %n%L’.

This tells rsync to print a verbose set of statistics on the
file transfer, allowing you to tell how effective rsync’s
delta-transfer algorithm is for your data. This option is
equivalent to –info=stats2 if combined with 0 or 1 -v options,
or –info=stats3 if combined with 2 or more -v options.

The current statistics are as follows:

o Number of files is the count of all “files” (in the
generic sense), which includes directories, symlinks,
etc. The total count will be followed by a list of
counts by filetype (if the total is non-zero). For exam‐
ple: “(reg: 5, dir: 3, link: 2, dev: 1, special: 1)”
lists the totals for regular files, directories, sym‐
links, devices, and special files. If any of value is 0,
it is completely omitted from the list.

o Number of created files is the count of how many “files”
(generic sense) were created (as opposed to updated).
The total count will be followed by a list of counts by
filetype (if the total is non-zero).

o Number of deleted files is the count of how many “files”
(generic sense) were created (as opposed to updated).
The total count will be followed by a list of counts by
filetype (if the total is non-zero). Note that this line
is only output if deletions are in effect, and only if
protocol 31 is being used (the default for rsync 3.1.x).

o Number of regular files transferred is the count of nor‐
mal files that were updated via rsync’s delta-transfer
algorithm, which does not include dirs, symlinks, etc.
Note that rsync 3.1.0 added the word “regular” into this

o Total file size is the total sum of all file sizes in the
transfer. This does not count any size for directories
or special files, but does include the size of symlinks.

o Total transferred file size is the total sum of all files
sizes for just the transferred files.

o Literal data is how much unmatched file-update data we
had to send to the receiver for it to recreate the
updated files.

o Matched data is how much data the receiver got locally
when recreating the updated files.

o File list size is how big the file-list data was when the
sender sent it to the receiver. This is smaller than the
in-memory size for the file list due to some compressing
of duplicated data when rsync sends the list.

o File list generation time is the number of seconds that
the sender spent creating the file list. This requires a
modern rsync on the sending side for this to be present.

o File list transfer time is the number of seconds that the
sender spent sending the file list to the receiver.

o Total bytes sent is the count of all the bytes that rsync
sent from the client side to the server side.

o Total bytes received is the count of all non-message
bytes that rsync received by the client side from the
server side. “Non-message” bytes means that we don’t
count the bytes for a verbose message that the server
sent to us, which makes the stats more consistent.

-8, –8-bit-output
This tells rsync to leave all high-bit characters unescaped in
the output instead of trying to test them to see if they’re
valid in the current locale and escaping the invalid ones. All
control characters (but never tabs) are always escaped, regard‐
less of this option’s setting.

The escape idiom that started in 2.6.7 is to output a literal
backslash (\) and a hash (#), followed by exactly 3 octal dig‐
its. For example, a newline would output as “\#012”. A literal
backslash that is in a filename is not escaped unless it is fol‐
lowed by a hash and 3 digits (0-9).

-h, –human-readable
Output numbers in a more human-readable format. There are 3
possible levels: (1) output numbers with a separator between
each set of 3 digits (either a comma or a period, depending on
if the decimal point is represented by a period or a comma); (2)
output numbers in units of 1000 (with a character suffix for
larger units — see below); (3) output numbers in units of 1024.

The default is human-readable level 1. Each -h option increases
the level by one. You can take the level down to 0 (to output
numbers as pure digits) by specifing the –no-human-readable
(–no-h) option.

The unit letters that are appended in levels 2 and 3 are: K
(kilo), M (mega), G (giga), or T (tera). For example, a
1234567-byte file would output as 1.23M in level-2 (assuming
that a period is your local decimal point).

Backward compatibility note: versions of rsync prior to 3.1.0
do not support human-readable level 1, and they default to level
0. Thus, specifying one or two -h options will behave in a com‐
parable manner in old and new versions as long as you didn’t
specify a –no-h option prior to one or more -h options. See
the –list-only option for one difference.

By default, rsync will delete any partially transferred file if
the transfer is interrupted. In some circumstances it is more
desirable to keep partially transferred files. Using the –par‐
tial option tells rsync to keep the partial file which should
make a subsequent transfer of the rest of the file much faster.

A better way to keep partial files than the –partial option is
to specify a DIR that will be used to hold the partial data
(instead of writing it out to the destination file). On the
next transfer, rsync will use a file found in this dir as data
to speed up the resumption of the transfer and then delete it
after it has served its purpose.

Note that if –whole-file is specified (or implied), any par‐
tial-dir file that is found for a file that is being updated
will simply be removed (since rsync is sending files without
using rsync’s delta-transfer algorithm).

Rsync will create the DIR if it is missing (just the last dir —
not the whole path). This makes it easy to use a relative path
(such as “–partial-dir=.rsync-partial”) to have rsync create
the partial-directory in the destination file’s directory when
needed, and then remove it again when the partial file is

If the partial-dir value is not an absolute path, rsync will add
an exclude rule at the end of all your existing excludes. This
will prevent the sending of any partial-dir files that may exist
on the sending side, and will also prevent the untimely deletion
of partial-dir items on the receiving side. An example: the
above –partial-dir option would add the equivalent of “-f ‘-p
.rsync-partial/'” at the end of any other filter rules.

If you are supplying your own exclude rules, you may need to add
your own exclude/hide/protect rule for the partial-dir because
(1) the auto-added rule may be ineffective at the end of your
other rules, or (2) you may wish to override rsync’s exclude
choice. For instance, if you want to make rsync clean-up any
left-over partial-dirs that may be lying around, you should
specify –delete-after and add a “risk” filter rule, e.g. -f ‘R
.rsync-partial/’. (Avoid using –delete-before or –delete-dur‐
ing unless you don’t need rsync to use any of the left-over par‐
tial-dir data during the current run.)

IMPORTANT: the –partial-dir should not be writable by other
users or it is a security risk. E.g. AVOID “/tmp”.

You can also set the partial-dir value the RSYNC_PARTIAL_DIR
environment variable. Setting this in the environment does not
force –partial to be enabled, but rather it affects where par‐
tial files go when –partial is specified. For instance,
instead of using –partial-dir=.rsync-tmp along with –progress,
you could set RSYNC_PARTIAL_DIR=.rsync-tmp in your environment
and then just use the -P option to turn on the use of the
.rsync-tmp dir for partial transfers. The only times that the
–partial option does not look for this environment value are
(1) when –inplace was specified (since –inplace conflicts with
–partial-dir), and (2) when –delay-updates was specified (see

For the purposes of the daemon-config’s “refuse options” set‐
ting, –partial-dir does not imply –partial. This is so that a
refusal of the –partial option can be used to disallow the
overwriting of destination files with a partial transfer, while
still allowing the safer idiom provided by –partial-dir.

This option puts the temporary file from each updated file into
a holding directory until the end of the transfer, at which time
all the files are renamed into place in rapid succession. This
attempts to make the updating of the files a little more atomic.
By default the files are placed into a directory named “.~tmp~”
in each file’s destination directory, but if you’ve specified
the –partial-dir option, that directory will be used instead.
See the comments in the –partial-dir section for a discussion
of how this “.~tmp~” dir will be excluded from the transfer, and
what you can do if you want rsync to cleanup old “.~tmp~” dirs
that might be lying around. Conflicts with –inplace and

This option uses more memory on the receiving side (one bit per
file transferred) and also requires enough free disk space on
the receiving side to hold an additional copy of all the updated
files. Note also that you should not use an absolute path to
–partial-dir unless (1) there is no chance of any of the files
in the transfer having the same name (since all the updated
files will be put into a single directory if the path is abso‐
lute) and (2) there are no mount points in the hierarchy (since
the delayed updates will fail if they can’t be renamed into

See also the “atomic-rsync” perl script in the “support” subdir
for an update algorithm that is even more atomic (it uses
–link-dest and a parallel hierarchy of files).

-m, –prune-empty-dirs
This option tells the receiving rsync to get rid of empty direc‐
tories from the file-list, including nested directories that
have no non-directory children. This is useful for avoiding the
creation of a bunch of useless directories when the sending
rsync is recursively scanning a hierarchy of files using
include/exclude/filter rules.

Note that the use of transfer rules, such as the –min-size
option, does not affect what goes into the file list, and thus
does not leave directories empty, even if none of the files in a
directory match the transfer rule.

Because the file-list is actually being pruned, this option also
affects what directories get deleted when a delete is active.
However, keep in mind that excluded files and directories can
prevent existing items from being deleted due to an exclude both
hiding source files and protecting destination files. See the
perishable filter-rule option for how to avoid this.

You can prevent the pruning of certain empty directories from
the file-list by using a global “protect” filter. For instance,
this option would ensure that the directory “emptydir” was kept
in the file-list:

–filter ’protect emptydir/’

Here’s an example that copies all .pdf files in a hierarchy,
only creating the necessary destination directories to hold the
.pdf files, and ensures that any superfluous files and directo‐
ries in the destination are removed (note the hide filter of
non-directories being used instead of an exclude):

rsync -avm –del –include=’*.pdf’ -f ’hide,! */’ src/ dest

If you didn’t want to remove superfluous destination files, the
more time-honored options of “–include=’*/’ –exclude=’*'”
would work fine in place of the hide-filter (if that is more
natural to you).

This option tells rsync to print information showing the
progress of the transfer. This gives a bored user something to
watch. With a modern rsync this is the same as specifying
–info=flist2,name,progress, but any user-supplied settings for
those info flags takes precedence (e.g. “–info=flist0

While rsync is transferring a regular file, it updates a
progress line that looks like this:

782448 63% 110.64kB/s 0:00:04

In this example, the receiver has reconstructed 782448 bytes or
63% of the sender’s file, which is being reconstructed at a rate
of 110.64 kilobytes per second, and the transfer will finish in
4 seconds if the current rate is maintained until the end.

These statistics can be misleading if rsync’s delta-transfer
algorithm is in use. For example, if the sender’s file consists
of the basis file followed by additional data, the reported rate
will probably drop dramatically when the receiver gets to the
literal data, and the transfer will probably take much longer to
finish than the receiver estimated as it was finishing the
matched part of the file.

When the file transfer finishes, rsync replaces the progress
line with a summary line that looks like this:

1,238,099 100% 146.38kB/s 0:00:08 (xfr#5, to-chk=169/396)

In this example, the file was 1,238,099 bytes long in total, the
average rate of transfer for the whole file was 146.38 kilobytes
per second over the 8 seconds that it took to complete, it was
the 5th transfer of a regular file during the current rsync ses‐
sion, and there are 169 more files for the receiver to check (to
see if they are up-to-date or not) remaining out of the 396
total files in the file-list.

In an incremental recursion scan, rsync won’t know the total
number of files in the file-list until it reaches the ends of
the scan, but since it starts to transfer files during the scan,
it will display a line with the text “ir-chk” (for incremental
recursion check) instead of “to-chk” until the point that it
knows the full size of the list, at which point it will switch
to using “to-chk”. Thus, seeing “ir-chk” lets you know that the
total count of files in the file list is still going to increase
(and each time it does, the count of files left to check will
increase by the number of the files added to the list).

-P The -P option is equivalent to –partial –progress. Its pur‐
pose is to make it much easier to specify these two options for
a long transfer that may be interrupted.

There is also a –info=progress2 option that outputs statistics
based on the whole transfer, rather than individual files. Use
this flag without outputting a filename (e.g. avoid -v or spec‐
ify –info=name0 if you want to see how the transfer is doing
without scrolling the screen with a lot of names. (You don’t
need to specify the –progress option in order to use

This option allows you to provide a password for accessing an
rsync daemon via a file or via standard input if FILE is -. The
file should contain just the password on the first line (all
other lines are ignored). Rsync will exit with an error if FILE
is world readable or if a root-run rsync command finds a
non-root-owned file.

This option does not supply a password to a remote shell trans‐
port such as ssh; to learn how to do that, consult the remote
shell’s documentation. When accessing an rsync daemon using a
remote shell as the transport, this option only comes into
effect after the remote shell finishes its authentication (i.e.
if you have also specified a password in the daemon’s config

This option will cause the source files to be listed instead of
transferred. This option is inferred if there is a single
source arg and no destination specified, so its main uses are:
(1) to turn a copy command that includes a destination arg into
a file-listing command, or (2) to be able to specify more than
one source arg (note: be sure to include the destination). Cau‐
tion: keep in mind that a source arg with a wild-card is
expanded by the shell into multiple args, so it is never safe to
try to list such an arg without using this option. For example:

rsync -av –list-only foo* dest/

Starting with rsync 3.1.0, the sizes output by –list-only are
affected by the –human-readable option. By default they will
contain digit separators, but higher levels of readability will
output the sizes with unit suffixes. Note also that the column
width for the size output has increased from 11 to 14 characters
for all human-readable levels. Use –no-h if you want just dig‐
its in the sizes, and the old column width of 11 characters.

Compatibility note: when requesting a remote listing of files
from an rsync that is version 2.6.3 or older, you may encounter
an error if you ask for a non-recursive listing. This is
because a file listing implies the –dirs option w/o –recur‐
sive, and older rsyncs don’t have that option. To avoid this
problem, either specify the –no-dirs option (if you don’t need
to expand a directory’s content), or turn on recursion and
exclude the content of subdirectories: -r –exclude=’/*/*’.

This option allows you to specify the maximum transfer rate for
the data sent over the socket, specified in units per second.
The RATE value can be suffixed with a string to indicate a size
multiplier, and may be a fractional value (e.g.
“–bwlimit=1.5m”). If no suffix is specified, the value will be
assumed to be in units of 1024 bytes (as if “K” or “KiB” had
been appended). See the –max-size option for a description of
all the available suffixes. A value of zero specifies no limit.

For backward-compatibility reasons, the rate limit will be
rounded to the nearest KiB unit, so no rate smaller than 1024
bytes per second is possible.

Rsync writes data over the socket in blocks, and this option
both limits the size of the blocks that rsync writes, and tries
to keep the average transfer rate at the requested limit. Some
“burstiness” may be seen where rsync writes out a block of data
and then sleeps to bring the average rate into compliance.

Due to the internal buffering of data, the –progress option may
not be an accurate reflection on how fast the data is being
sent. This is because some files can show up as being rapidly
sent when the data is quickly buffered, while other can show up
as very slow when the flushing of the output buffer occurs.
This may be fixed in a future version.

Record a file that can later be applied to another identical
destination with –read-batch. See the “BATCH MODE” section for
details, and also the –only-write-batch option.

Works like –write-batch, except that no updates are made on the
destination system when creating the batch. This lets you
transport the changes to the destination system via some other
means and then apply the changes via –read-batch.

Note that you can feel free to write the batch directly to some
portable media: if this media fills to capacity before the end
of the transfer, you can just apply that partial transfer to the
destination and repeat the whole process to get the rest of the
changes (as long as you don’t mind a partially updated destina‐
tion system while the multi-update cycle is happening).

Also note that you only save bandwidth when pushing changes to a
remote system because this allows the batched data to be
diverted from the sender into the batch file without having to
flow over the wire to the receiver (when pulling, the sender is
remote, and thus can’t write the batch).

Apply all of the changes stored in FILE, a file previously gen‐
erated by –write-batch. If FILE is -, the batch data will be
read from standard input. See the “BATCH MODE” section for

Force an older protocol version to be used. This is useful for
creating a batch file that is compatible with an older version
of rsync. For instance, if rsync 2.6.4 is being used with the
–write-batch option, but rsync 2.6.3 is what will be used to
run the –read-batch option, you should use “–protocol=28” when
creating the batch file to force the older protocol version to
be used in the batch file (assuming you can’t upgrade the rsync
on the reading system).

Rsync can convert filenames between character sets using this
option. Using a CONVERT_SPEC of “.” tells rsync to look up the
default character-set via the locale setting. Alternately, you
can fully specify what conversion to do by giving a local and a
remote charset separated by a comma in the order
–iconv=LOCAL,REMOTE, e.g. –iconv=utf8,iso88591. This order
ensures that the option will stay the same whether you’re push‐
ing or pulling files. Finally, you can specify either
–no-iconv or a CONVERT_SPEC of “-” to turn off any conversion.
The default setting of this option is site-specific, and can
also be affected via the RSYNC_ICONV environment variable.

For a list of what charset names your local iconv library sup‐
ports, you can run “iconv –list”.

If you specify the –protect-args option (-s), rsync will trans‐
late the filenames you specify on the command-line that are
being sent to the remote host. See also the –files-from

Note that rsync does not do any conversion of names in filter
files (including include/exclude files). It is up to you to
ensure that you’re specifying matching rules that can match on
both sides of the transfer. For instance, you can specify extra
include/exclude rules if there are filename differences on the
two sides that need to be accounted for.

When you pass an –iconv option to an rsync daemon that allows
it, the daemon uses the charset specified in its “charset” con‐
figuration parameter regardless of the remote charset you actu‐
ally pass. Thus, you may feel free to specify just the local
charset for a daemon transfer (e.g. –iconv=utf8).

-4, –ipv4 or -6, –ipv6
Tells rsync to prefer IPv4/IPv6 when creating sockets. This
only affects sockets that rsync has direct control over, such as
the outgoing socket when directly contacting an rsync daemon.
See also these options in the –daemon mode section.

If rsync was complied without support for IPv6, the –ipv6
option will have no effect. The –version output will tell you
if this is the case.

Set the checksum seed to the integer NUM. This 4 byte checksum
seed is included in each block and MD4 file checksum calculation
(the more modern MD5 file checksums don’t use a seed). By
default the checksum seed is generated by the server and
defaults to the current time() . This option is used to set a
specific checksum seed, which is useful for applications that
want repeatable block checksums, or in the case where the user
wants a more random checksum seed. Setting NUM to 0 causes
rsync to use the default of time() for checksum seed.



The options allowed when starting an rsync daemon are as follows:

This tells rsync that it is to run as a daemon. The daemon you
start running may be accessed using an rsync client using the
host::module or rsync://host/module/ syntax.

If standard input is a socket then rsync will assume that it is
being run via inetd, otherwise it will detach from the current
terminal and become a background daemon. The daemon will read
the config file (rsyncd.conf) on each connect made by a client
and respond to requests accordingly. See the rsyncd.conf(5) man
page for more details.

By default rsync will bind to the wildcard address when run as a
daemon with the –daemon option. The –address option allows
you to specify a specific IP address (or hostname) to bind to.
This makes virtual hosting possible in conjunction with the
–config option. See also the “address” global option in the
rsyncd.conf manpage.

This option allows you to specify the maximum transfer rate for
the data the daemon sends over the socket. The client can still
specify a smaller –bwlimit value, but no larger value will be
allowed. See the client version of this option (above) for some
extra details.

This specifies an alternate config file than the default. This
is only relevant when –daemon is specified. The default is
/etc/rsyncd.conf unless the daemon is running over a remote
shell program and the remote user is not the super-user; in that
case the default is rsyncd.conf in the current directory (typi‐
cally $HOME).

-M, –dparam=OVERRIDE
This option can be used to set a daemon-config parameter when
starting up rsync in daemon mode. It is equivalent to adding
the parameter at the end of the global settings prior to the
first module’s definition. The parameter names can be specified
without spaces, if you so desire. For instance:

rsync –daemon -M pidfile=/path/

When running as a daemon, this option instructs rsync to not
detach itself and become a background process. This option is
required when running as a service on Cygwin, and may also be
useful when rsync is supervised by a program such as daemontools
or AIX’s System Resource Controller. –no-detach is also recom‐
mended when rsync is run under a debugger. This option has no
effect if rsync is run from inetd or sshd.

This specifies an alternate TCP port number for the daemon to
listen on rather than the default of 873. See also the “port”
global option in the rsyncd.conf manpage.

This option tells the rsync daemon to use the given log-file
name instead of using the “log file” setting in the config file.

This option tells the rsync daemon to use the given FORMAT
string instead of using the “log format” setting in the config
file. It also enables “transfer logging” unless the string is
empty, in which case transfer logging is turned off.

This overrides the socket options setting in the rsyncd.conf
file and has the same syntax.

-v, –verbose
This option increases the amount of information the daemon logs
during its startup phase. After the client connects, the dae‐
mon’s verbosity level will be controlled by the options that the
client used and the “max verbosity” setting in the module’s con‐
fig section.

-4, –ipv4 or -6, –ipv6
Tells rsync to prefer IPv4/IPv6 when creating the incoming sock‐
ets that the rsync daemon will use to listen for connections.
One of these options may be required in older versions of Linux
to work around an IPv6 bug in the kernel (if you see an “address
already in use” error when nothing else is using the port, try
specifying –ipv6 or –ipv4 when starting the daemon).

If rsync was complied without support for IPv6, the –ipv6
option will have no effect. The –version output will tell you
if this is the case.

-h, –help
When specified after –daemon, print a short help page describ‐
ing the options available for starting an rsync daemon.

The filter rules allow for flexible selection of which files to trans‐
fer (include) and which files to skip (exclude). The rules either
directly specify include/exclude patterns or they specify a way to
acquire more include/exclude patterns (e.g. to read them from a file).

As the list of files/directories to transfer is built, rsync checks
each name to be transferred against the list of include/exclude pat‐
terns in turn, and the first matching pattern is acted on: if it is an
exclude pattern, then that file is skipped; if it is an include pattern
then that filename is not skipped; if no matching pattern is found,
then the filename is not skipped.

Rsync builds an ordered list of filter rules as specified on the com‐
mand-line. Filter rules have the following syntax:


You have your choice of using either short or long RULE names, as
described below. If you use a short-named rule, the ’,’ separating the
RULE from the MODIFIERS is optional. The PATTERN or FILENAME that fol‐
lows (when present) must come after either a single space or an under‐
score (_). Here are the available rule prefixes:

exclude, – specifies an exclude pattern.
include, + specifies an include pattern.
merge, . specifies a merge-file to read for more rules.
dir-merge, : specifies a per-directory merge-file.
hide, H specifies a pattern for hiding files from the transfer.
show, S files that match the pattern are not hidden.
protect, P specifies a pattern for protecting files from dele‐
risk, R files that match the pattern are not protected.
clear, ! clears the current include/exclude list (takes no arg)

When rules are being read from a file, empty lines are ignored, as are
comment lines that start with a “#”.

Note that the –include/–exclude command-line options do not allow the
full range of rule parsing as described above — they only allow the
specification of include/exclude patterns plus a “!” token to clear the
list (and the normal comment parsing when rules are read from a file).
If a pattern does not begin with “- ” (dash, space) or “+ ” (plus,
space), then the rule will be interpreted as if “+ ” (for an include
option) or “- ” (for an exclude option) were prefixed to the string. A
–filter option, on the other hand, must always contain either a short
or long rule name at the start of the rule.

Note also that the –filter, –include, and –exclude options take one
rule/pattern each. To add multiple ones, you can repeat the options on
the command-line, use the merge-file syntax of the –filter option, or
the –include-from/–exclude-from options.

You can include and exclude files by specifying patterns using the “+”,
“-“, etc. filter rules (as introduced in the FILTER RULES section
above). The include/exclude rules each specify a pattern that is
matched against the names of the files that are going to be trans‐
ferred. These patterns can take several forms:

o if the pattern starts with a / then it is anchored to a particu‐
lar spot in the hierarchy of files, otherwise it is matched
against the end of the pathname. This is similar to a leading ^
in regular expressions. Thus “/foo” would match a name of “foo”
at either the “root of the transfer” (for a global rule) or in
the merge-file’s directory (for a per-directory rule). An
unqualified “foo” would match a name of “foo” anywhere in the
tree because the algorithm is applied recursively from the top
down; it behaves as if each path component gets a turn at being
the end of the filename. Even the unanchored “sub/foo” would
match at any point in the hierarchy where a “foo” was found
within a directory named “sub”. See the section on ANCHORING
INCLUDE/EXCLUDE PATTERNS for a full discussion of how to specify
a pattern that matches at the root of the transfer.

o if the pattern ends with a / then it will only match a direc‐
tory, not a regular file, symlink, or device.

o rsync chooses between doing a simple string match and wildcard
matching by checking if the pattern contains one of these three
wildcard characters: ’*’, ’?’, and ’[’ .

o a ’*’ matches any path component, but it stops at slashes.

o use ’**’ to match anything, including slashes.

o a ’?’ matches any character except a slash (/).

o a ’[’ introduces a character class, such as [a-z] or

o in a wildcard pattern, a backslash can be used to escape a wild‐
card character, but it is matched literally when no wildcards
are present. This means that there is an extra level of back‐
slash removal when a pattern contains wildcard characters com‐
pared to a pattern that has none. e.g. if you add a wildcard to
“foo\bar” (which matches the backslash) you would need to use
“foo\\bar*” to avoid the “\b” becoming just “b”.

o if the pattern contains a / (not counting a trailing /) or a
“**”, then it is matched against the full pathname, including
any leading directories. If the pattern doesn’t contain a / or a
“**”, then it is matched only against the final component of the
filename. (Remember that the algorithm is applied recursively
so “full filename” can actually be any portion of a path from
the starting directory on down.)

o a trailing “dir_name/***” will match both the directory (as if
“dir_name/” had been specified) and everything in the directory
(as if “dir_name/**” had been specified). This behavior was
added in version 2.6.7.

Note that, when using the –recursive (-r) option (which is implied by
-a), every subcomponent of every path is visited from the top down, so
include/exclude patterns get applied recursively to each subcomponent’s
full name (e.g. to include “/foo/bar/baz” the subcomponents “/foo” and
“/foo/bar” must not be excluded). The exclude patterns actually
short-circuit the directory traversal stage when rsync finds the files
to send. If a pattern excludes a particular parent directory, it can
render a deeper include pattern ineffectual because rsync did not
descend through that excluded section of the hierarchy. This is par‐
ticularly important when using a trailing ’*’ rule. For instance, this
won’t work:

+ /some/path/this-file-will-not-be-found
+ /file-is-included
– *

This fails because the parent directory “some” is excluded by the ’*’
rule, so rsync never visits any of the files in the “some” or
“some/path” directories. One solution is to ask for all directories in
the hierarchy to be included by using a single rule: “+ */” (put it
somewhere before the “- *” rule), and perhaps use the
–prune-empty-dirs option. Another solution is to add specific include
rules for all the parent dirs that need to be visited. For instance,
this set of rules works fine:

+ /some/
+ /some/path/
+ /some/path/this-file-is-found
+ /file-also-included
– *

Here are some examples of exclude/include matching:

o “- *.o” would exclude all names matching *.o

o “- /foo” would exclude a file (or directory) named foo in the
transfer-root directory

o “- foo/” would exclude any directory named foo

o “- /foo/*/bar” would exclude any file named bar which is at two
levels below a directory named foo in the transfer-root direc‐

o “- /foo/**/bar” would exclude any file named bar two or more
levels below a directory named foo in the transfer-root direc‐

o The combination of “+ */”, “+ *.c”, and “- *” would include all
directories and C source files but nothing else (see also the
–prune-empty-dirs option)

o The combination of “+ foo/”, “+ foo/bar.c”, and “- *” would
include only the foo directory and foo/bar.c (the foo directory
must be explicitly included or it would be excluded by the “*”)

The following modifiers are accepted after a “+” or “-“:

o A / specifies that the include/exclude rule should be matched
against the absolute pathname of the current item. For example,
“-/ /etc/passwd” would exclude the passwd file any time the
transfer was sending files from the “/etc” directory, and “-/
subdir/foo” would always exclude “foo” when it is in a dir named
“subdir”, even if “foo” is at the root of the current transfer.

o A ! specifies that the include/exclude should take effect if the
pattern fails to match. For instance, “-! */” would exclude all

o A C is used to indicate that all the global CVS-exclude rules
should be inserted as excludes in place of the “-C”. No arg
should follow.

o An s is used to indicate that the rule applies to the sending
side. When a rule affects the sending side, it prevents files
from being transferred. The default is for a rule to affect
both sides unless –delete-excluded was specified, in which case
default rules become sender-side only. See also the hide (H)
and show (S) rules, which are an alternate way to specify send‐
ing-side includes/excludes.

o An r is used to indicate that the rule applies to the receiving
side. When a rule affects the receiving side, it prevents files
from being deleted. See the s modifier for more info. See also
the protect (P) and risk (R) rules, which are an alternate way
to specify receiver-side includes/excludes.

o A p indicates that a rule is perishable, meaning that it is
ignored in directories that are being deleted. For instance,
the -C option’s default rules that exclude things like “CVS” and
“*.o” are marked as perishable, and will not prevent a directory
that was removed on the source from being deleted on the desti‐

You can merge whole files into your filter rules by specifying either a
merge (.) or a dir-merge (:) filter rule (as introduced in the FILTER
RULES section above).

There are two kinds of merged files — single-instance (’.’) and
per-directory (’:’). A single-instance merge file is read one time,
and its rules are incorporated into the filter list in the place of the
“.” rule. For per-directory merge files, rsync will scan every direc‐
tory that it traverses for the named file, merging its contents when
the file exists into the current list of inherited rules. These
per-directory rule files must be created on the sending side because it
is the sending side that is being scanned for the available files to
transfer. These rule files may also need to be transferred to the
receiving side if you want them to affect what files don’t get deleted

Some examples:

merge /etc/rsync/default.rules
. /etc/rsync/default.rules
dir-merge .per-dir-filter
dir-merge,n- .non-inherited-per-dir-excludes
:n- .non-inherited-per-dir-excludes

The following modifiers are accepted after a merge or dir-merge rule:

o A – specifies that the file should consist of only exclude pat‐
terns, with no other rule-parsing except for in-file comments.

o A + specifies that the file should consist of only include pat‐
terns, with no other rule-parsing except for in-file comments.

o A C is a way to specify that the file should be read in a
CVS-compatible manner. This turns on ’n’, ’w’, and ’-’, but
also allows the list-clearing token (!) to be specified. If no
filename is provided, “.cvsignore” is assumed.

o A e will exclude the merge-file name from the transfer; e.g.
“dir-merge,e .rules” is like “dir-merge .rules” and “- .rules”.

o An n specifies that the rules are not inherited by subdirecto‐

o A w specifies that the rules are word-split on whitespace
instead of the normal line-splitting. This also turns off com‐
ments. Note: the space that separates the prefix from the rule
is treated specially, so “- foo + bar” is parsed as two rules
(assuming that prefix-parsing wasn’t also disabled).

o You may also specify any of the modifiers for the “+” or “-”
rules (above) in order to have the rules that are read in from
the file default to having that modifier set (except for the !
modifier, which would not be useful). For instance, “merge,-/
.excl” would treat the contents of .excl as absolute-path
excludes, while “dir-merge,s .filt” and “:sC” would each make
all their per-directory rules apply only on the sending side.
If the merge rule specifies sides to affect (via the s or r mod‐
ifier or both), then the rules in the file must not specify
sides (via a modifier or a rule prefix such as hide).

Per-directory rules are inherited in all subdirectories of the direc‐
tory where the merge-file was found unless the ’n’ modifier was used.
Each subdirectory’s rules are prefixed to the inherited per-directory
rules from its parents, which gives the newest rules a higher priority
than the inherited rules. The entire set of dir-merge rules are
grouped together in the spot where the merge-file was specified, so it
is possible to override dir-merge rules via a rule that got specified
earlier in the list of global rules. When the list-clearing rule (“!”)
is read from a per-directory file, it only clears the inherited rules
for the current merge file.

Another way to prevent a single rule from a dir-merge file from being
inherited is to anchor it with a leading slash. Anchored rules in a
per-directory merge-file are relative to the merge-file’s directory, so
a pattern “/foo” would only match the file “foo” in the directory where
the dir-merge filter file was found.

Here’s an example filter file which you’d specify via –filter=”.

merge /home/user/.global-filter
– *.gz
dir-merge .rules
+ *.[ch] – *.o

This will merge the contents of the /home/user/.global-filter file at
the start of the list and also turns the “.rules” filename into a
per-directory filter file. All rules read in prior to the start of the
directory scan follow the global anchoring rules (i.e. a leading slash
matches at the root of the transfer).

If a per-directory merge-file is specified with a path that is a parent
directory of the first transfer directory, rsync will scan all the par‐
ent dirs from that starting point to the transfer directory for the
indicated per-directory file. For instance, here is a common filter
(see -F):

–filter=’: /.rsync-filter’

That rule tells rsync to scan for the file .rsync-filter in all direc‐
tories from the root down through the parent directory of the transfer
prior to the start of the normal directory scan of the file in the
directories that are sent as a part of the transfer. (Note: for an
rsync daemon, the root is always the same as the module’s “path”.)

Some examples of this pre-scanning for per-directory files:

rsync -avF /src/path/ /dest/dir
rsync -av –filter=’: ../../.rsync-filter’ /src/path/ /dest/dir
rsync -av –filter=’: .rsync-filter’ /src/path/ /dest/dir

The first two commands above will look for “.rsync-filter” in “/” and
“/src” before the normal scan begins looking for the file in
“/src/path” and its subdirectories. The last command avoids the par‐
ent-dir scan and only looks for the “.rsync-filter” files in each
directory that is a part of the transfer.

If you want to include the contents of a “.cvsignore” in your patterns,
you should use the rule “:C”, which creates a dir-merge of the .cvsig‐
nore file, but parsed in a CVS-compatible manner. You can use this to
affect where the –cvs-exclude (-C) option’s inclusion of the
per-directory .cvsignore file gets placed into your rules by putting
the “:C” wherever you like in your filter rules. Without this, rsync
would add the dir-merge rule for the .cvsignore file at the end of all
your other rules (giving it a lower priority than your command-line
rules). For example:

cat < out.dat

then look at out.dat. If everything is working correctly then out.dat
should be a zero length file. If you are getting the above error from
rsync then you will probably find that out.dat contains some text or
data. Look at the contents and try to work out what is producing it.
The most common cause is incorrectly configured shell startup scripts
(such as .cshrc or .profile) that contain output statements for
non-interactive logins.

If you are having trouble debugging filter patterns, then try specify‐
ing the -vv option. At this level of verbosity rsync will show why
each individual file is included or excluded.

0 Success

1 Syntax or usage error

2 Protocol incompatibility

3 Errors selecting input/output files, dirs

4 Requested action not supported: an attempt was made to manipu‐
late 64-bit files on a platform that cannot support them; or an
option was specified that is supported by the client and not by
the server.

5 Error starting client-server protocol

6 Daemon unable to append to log-file

10 Error in socket I/O

11 Error in file I/O

12 Error in rsync protocol data stream

13 Errors with program diagnostics

14 Error in IPC code

20 Received SIGUSR1 or SIGINT

21 Some error returned by waitpid()

22 Error allocating core memory buffers

23 Partial transfer due to error

24 Partial transfer due to vanished source files

25 The –max-delete limit stopped deletions

30 Timeout in data send/receive

35 Timeout waiting for daemon connection

The CVSIGNORE environment variable supplements any ignore pat‐
terns in .cvsignore files. See the –cvs-exclude option for more

Specify a default –iconv setting using this environment vari‐
able. (First supported in 3.0.0.)

Specify a non-zero numeric value if you want the –protect-args
option to be enabled by default, or a zero value to make sure
that it is disabled by default. (First supported in 3.1.0.)

The RSYNC_RSH environment variable allows you to override the
default shell used as the transport for rsync. Command line
options are permitted after the command name, just as in the -e

The RSYNC_PROXY environment variable allows you to redirect your
rsync client to use a web proxy when connecting to a rsync dae‐
mon. You should set RSYNC_PROXY to a hostname:port pair.

Setting RSYNC_PASSWORD to the required password allows you to
run authenticated rsync connections to an rsync daemon without
user intervention. Note that this does not supply a password to
a remote shell transport such as ssh; to learn how to do that,
consult the remote shell’s documentation.



The USER or LOGNAME environment variables are used to determine
the default username sent to an rsync daemon. If neither is
set, the username defaults to “nobody”.

HOME The HOME environment variable is used to find the user’s default
.cvsignore file.

/etc/rsyncd.conf or rsyncd.conf




times are transferred as *nix time_t values

When transferring to FAT filesystems rsync may re-sync unmodified
files. See the comments on the –modify-window option.

file permissions, devices, etc. are transferred as native numerical

see also the comments on the –delete option

Please report bugs! See the web site at

This man page is current for version 3.1.1 of rsync.



The options –server and –sender are used internally by rsync, and
should never be typed by a user under normal circumstances. Some
awareness of these options may be needed in certain scenarios, such as
when setting up a login that can only run an rsync command. For
instance, the support directory of the rsync distribution has an exam‐
ple script named rrsync (for restricted rsync) that can be used with a
restricted ssh login.

rsync is distributed under the GNU General Public License. See the
file COPYING for details.

A WEB site is available at The site includes
an FAQ-O-Matic which may cover questions unanswered by this manual

The primary ftp site for rsync is

We would be delighted to hear from you if you like this program.
Please contact the mailing-list at

This program uses the excellent zlib compression library written by
Jean-loup Gailly and Mark Adler.

Special thanks go out to: John Van Essen, Matt McCutchen, Wesley W.
Terpstra, David Dykstra, Jos Backus, Sebastian Krahmer, Martin Pool,
and our gone-but-not-forgotten compadre, J.W. Schultz.

Thanks also to Richard Brent, Brendan Mackay, Bill Waite, Stephen Roth‐
well and David Bell. I’ve probably missed some people, my apologies if
I have.


rsync was originally written by Andrew Tridgell and Paul Mackerras.
Many people have later contributed to it. It is currently maintained
by Wayne Davison.

Mailing lists for support and development are available at

22 Jun 2014 rsync