NAME
s3cmd - tool for managing Amazon S3 storage space and Amazon CloudFront
content delivery network
SYNOPSIS
s3cmd [OPTIONS] COMMAND [PARAMETERS]
DESCRIPTION
s3cmd is a command line client for copying files to/from Amazon S3
(Simple Storage Service) and performing other related tasks, for
instance creating and removing buckets, listing objects, etc.
s3cmd can do several actions specified by the following commands.
mb s3://BUCKET
Make bucket
rb s3://BUCKET
Remove bucket
ls [s3://BUCKET[/PREFIX]]
List objects or buckets
la List all object in all buckets
put FILE [FILE...] s3://BUCKET[/PREFIX]
Put file into bucket (i.e. upload to S3)
get s3://BUCKET/OBJECT LOCAL_FILE
Get file from bucket (i.e. download from S3)
del s3://BUCKET/OBJECT
Delete file from bucket
sync LOCAL_DIR s3://BUCKET[/PREFIX]
Backup a directory tree to S3
sync s3://BUCKET[/PREFIX] LOCAL_DIR
Restore a tree from S3 to local directory
cp s3://BUCKET1/OBJECT1 s3://BUCKET2[/OBJECT2], mv s3://BUCKET1/OBJECT1
s3://BUCKET2[/OBJECT2]
Make a copy of a file (cp) or move a file (mv). Destination can
be in the same bucket with a different name or in another bucket
with the same or different name. Adding --acl-public will make
the destination object publicly accessible (see below).
setacl s3://BUCKET[/OBJECT]
Modify Access control list for Bucket or Files. Use with
--acl-public or --acl-private
info s3://BUCKET[/OBJECT]
Get various information about a Bucket or Object
du [s3://BUCKET[/PREFIX]]
Disk usage - amount of data stored in S3
Commands for CloudFront management
cflist List CloudFront distribution points
cfinfo [cf://DIST_ID]
Display CloudFront distribution point parameters
cfcreate s3://BUCKET
Create CloudFront distribution point
cfdelete cf://DIST_ID
Delete CloudFront distribution point
cfmodify cf://DIST_ID
Change CloudFront distribution point parameters
OPTIONS
Some of the below specified options can have their default values set
in s3cmd config file (by default $HOME/.s3cmd). As it’s a simple text
file feel free to open it with your favorite text editor and do any
changes you like.
Config file related options.
--configure
Invoke interactive (re)configuration tool. Don’t worry, you
won’t lose your settings on subsequent runs.
-c FILE, --config=FILE
Config file name. Defaults to $HOME/.s3cfg
--dump-config
Dump current configuration after parsing config files and
command line options and exit.
Options specific for file transfer commands (sync, put and get):
-n, --dry-run
Only show what should be uploaded or downloaded but don’t
actually do it. May still perform S3 requests to get bucket
listings and other in formation though.
--delete-removed
Delete remote objects with no corresponding local file when
syncing to S3 or delete local files with no corresponding object
in S3 when syncing from S3.
--no-delete-removed
Don’t delete remote objects. Default for sync command.
-p, --preserve
Preserve filesystem attributes (mode, ownership, timestamps).
Default for sync command.
--no-preserve
Don’t store filesystem attributes with uploaded files.
--exclude GLOB
Exclude files matching GLOB (a.k.a. shell-style wildcard) from
sync. See FILE TRANSFERS section and http://s3tools.org/s3cmd-
sync for more information.
--exclude-from FILE
Same as --exclude but reads GLOBs from the given FILE instead of
expecting them on the command line.
--rexclude REGEXP
Same as --exclude but works with REGEXPs (Regular expressions).
--rexclude-from FILE
Same as --exclude-from but works with REGEXPs.
--include=GLOB, --include-from=FILE, --rinclude=REGEXP,
--rinclude-from=FILE
Filenames and paths matching GLOB or REGEXP will be included
even if previously excluded by one of --(r)exclude(-from)
patterns
--continue
Continue getting a partially downloaded file (only for get
command). This comes handy once download of a large file, say an
ISO image, from a S3 bucket fails and a partially downloaded
file is left on the disk. Unfortunately put command doesn’t
support restarting of failed upload due to Amazon S3
limitations.
--skip-existing
Skip over files that exist at the destination (only for get and
sync commands).
-m MIME/TYPE, --mime-type=MIME/TYPE
Default MIME-type to be set for objects stored.
-M, --guess-mime-type
Guess MIME‐type of files by their extension. Falls back to
default MIME‐Type as specified by --mime-type option
--add-header=NAME:VALUE
Add a given HTTP header to the upload request. Can be used
multiple times with different header names. For instance set
’Expires’ or ’Cache-Control’ headers (or both) using this
options if you like.
-P, --acl-public
Store objects with permissions allowing read for anyone. See
http://s3tools.org/s3cmd-public for details and hints for
storing publicly accessible files.
--acl-private
Store objects with default ACL allowing access for you only.
-e, --encrypt
Use GPG encryption to protect stored objects from unauthorized
access. See http://s3tools.org/s3cmd-public for details about
encryption.
--no-encrypt
Don’t encrypt files.
Options for CloudFront commands:
See http://s3tools.org/s3cmd-cloudfront for more details.
--enable
Enable given CloudFront distribution (only for cfmodify command)
--disable
Enable given CloudFront distribution (only for cfmodify command)
--cf-add-cname=CNAME
Add given CNAME to a CloudFront distribution (only for cfcreate
and cfmodify commands)
--cf-remove-cname=CNAME
Remove given CNAME from a CloudFront distribution (only for
cfmodify command)
--cf-comment=COMMENT
Set COMMENT for a given CloudFront distribution (only for
cfcreate and cfmodify commands)
Options common for all commands (where it makes sense indeed):
-r, --recursive
Recursive upload, download or removal. When used with del it can
remove all the files in a bucket.
-f, --force
Force overwrite and other dangerous operations. Can be used to
remove a non-empty buckets with s3cmd rb --force s3://bkt
--bucket-location=BUCKET_LOCATION
Specify datacentre where to create the bucket. Possible values
are US (default) or EU.
-H, --human-readable-sizes
Print sizes in human readable form.
--list-md5
Include MD5 sums in bucket listings (only for ls command).
--progress, --no-progress
Display or don’t display progress meter. When running on TTY
(e.g. console or xterm) the default is to display progress
meter. If not on TTY (e.g. output is redirected somewhere or
running from cron) the default is to not display progress meter.
--encoding=ENCODING
Override autodetected terminal and filesystem encoding
(character set).
-v, --verbose
Enable verbose output.
-d, --debug
Enable debug output.
-h, --help
Show the help message and exit
--version
Show s3cmd version and exit.
FILE TRANSFERS
One of the most powerful commands of s3cmd is s3cmd sync used for
synchronising complete directory trees to or from remote S3 storage. To
some extent s3cmd put and s3cmd get share a similar behaviour with
sync.
Basic usage common in backup scenarios is as simple as:
s3cmd sync /local/path/ s3://test-bucket/backup/
This command will find all files under /local/path directory and copy
them to corresponding paths under s3://test-bucket/backup on the remote
side. For example:
/local/path/file1.ext -> s3://bucket/backup/file1.ext
/local/path/dir123/file2.bin -> s3://bucket/backup/dir123/file2.bin
However if the local path doesn’t end with a slash the last directory’s
name is used on the remote side as well. Compare these with the
previous example:
s3cmd sync /local/path s3://test-bucket/backup/
will sync:
/local/path/file1.ext -> s3://bucket/backup/path/file1.ext
/local/path/dir123/file2.bin -> s3://bucket/backup/path/dir123/file2.bin
To retrieve the files back from S3 use inverted syntax:
s3cmd sync s3://test-bucket/backup/ /tmp/restore/
that will download files:
s3://bucket/backup/file1.ext -> /tmp/restore/file1.ext
s3://bucket/backup/dir123/file2.bin -> /tmp/restore/dir123/file2.bin
Without the trailing slash on source the behaviour is similar to what
has been demonstrated with upload:
s3cmd sync s3://test-bucket/backup /tmp/restore/
will download the files as:
s3://bucket/backup/file1.ext -> /tmp/restore/backup/file1.ext
s3://bucket/backup/dir123/file2.bin -> /tmp/restore/backup/dir123/file2.bin
All source file names, the bold ones above, are matched against exclude
rules and those that match are then re-checked against include rules to
see whether they should be excluded or kept in the source list.
For the purpose of --exclude and --include matching only the bold file
names above are used. For instance only path/file1.ext is tested
against the patterns, not /local/path/file1.ext
Both --exclude and --include work with shell-style wildcards (a.k.a.
GLOB). For a greater flexibility s3cmd provides Regular-expression
versions of the two exclude options named --rexclude and --rinclude.
The options with ...-from suffix (eg --rinclude-from) expect a filename
as an argument. Each line of such a file is treated as one pattern.
There is only one set of patterns built from all --(r)exclude(-from)
options and similarly for include variant. Any file excluded with eg
--exclude can be put back with a pattern found in --rinclude-from list.
Run s3cmd with --dry-run to verify that your rules work as expected.
Use together with --debug get detailed information about matching file
names against exclude and include rules.
For example to exclude all files with ".jpg" extension except those
beginning with a number use:
--exclude ’*.jpg’ --rinclude ’[0-9].*.jpg’
SEE ALSO
For the most up to date list of options run s3cmd --help
For more info about usage, examples and other related info visit
project homepage at
http://s3tools.org
AUTHOR
Written by Michal Ludvig <michal@logix.cz>
CONTACT, SUPPORT
Prefered way to get support is our mailing list:
s3tools-general@lists.sourceforge.net
REPORTING BUGS
Report bugs to s3tools-bugs@lists.sourceforge.net
COPYRIGHT
Copyright © 2007,2008,2009 Michal Ludvig <http://www.logix.cz/michal>
This is free software. You may redistribute copies of it under the
terms of the GNU General Public License version 2
<http://www.gnu.org/licenses/gpl.html>. There is NO WARRANTY, to the
extent permitted by law.
s3cmd(1)