elementalnet/s3s3mirror

Fork of s3s3mirror

This fork is based on https://github.com/cobbzilla/s3s3mirror/tree/master. It supports using different client connections for the source and destination buckets for example when mirroring from one S3 compatible storage to another. It also adds on-the-fly encryption, decryption and reencryption. The following encryption algorithms are supported:

• SSE-C
• CSE AES-256
• CSE AES/GCM-256
• CSE AES/GCM-256 (in strict mode)

For generating the master key a custom key derivation function based on PBKDF2 and SHA-256 is used. The salt is constant! For compatibility with existing applications the KDF, rounds and salt should probably be made configurable.

It was tested with OpenJDK 1.8 and most of the dependencies have been updated. The extended test suite runs mostly successful. There seems to be a small problem with mirroring from an SSE encrypted bucket so another SSE encrypted bucket which I couldn't figure out yet. The test suite was run against a Minio and a Google endpoint. The Google endpoint requires a special option (see tests/.s3cfg) and a patched AWS Java SDK which is included as a Git submodule.

All endpoint configuration is now done in an .s3cfg file which I have slightly extended. See tests/.s3cfg for examples.

s3s3mirror

A utility for mirroring content from one S3 bucket to another.

Designed to be lightning-fast and highly concurrent, with modest CPU and memory requirements.

An object will be copied if and only if at least one of the following holds true:

• The object does not exist in the destination bucket.
• The size or ETag of the object in the destination bucket are different from the size/ETag in the source bucket.

When copying, the source metadata and ACL lists are also copied to the destination object.

Note: https://github.com/cobbzilla/s3s3mirror/tree/2.0-stable supports copying to/from local directories.

Motivation

I started with "s3cmd sync" but found that with buckets containing many thousands of objects, it was incredibly slow to start and consumed massive amounts of memory. So I designed s3s3mirror to start copying immediately with an intelligently chosen "chunk size" and to operate in a highly-threaded, streaming fashion, so memory requirements are much lower.

Running with 100 threads, I found the gating factor to be how fast I could list items from the source bucket (!?!) Which makes me wonder if there is any way to do this faster. I'm sure there must be, but this is pretty damn fast.

AWS Credentials

• s3s3mirror will first look for credentials in your system environment. If variables named AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY are defined, then these will be used.
• Next, it checks for a ~/.s3cfg file (which you might have for using s3cmd). If present, the access key and secret key are read from there.
• IAM Roles can be used on EC2 instances by specifying the --iam flag
• If none of the above is found, it will error out and refuse to run.

System Requirements

• Java 6 or Java 7

Note: currently there are problems compiling under Java 8. If you're including s3s3mirror in a larger project that uses Java8, compile it with Java 7 first, and compile your other code with Java 8. It should be fine to run with Java 8, just some issues with the compiler.

Building

mvn package

Note that s3s3mirror now has a prebuilt jar checked in to github, so you'll only need to do this if you've been playing with the source code. The above command requires that Maven 3 is installed.

License

s3s3mirror is available under the https://www.apache.org/licenses/LICENSE-2.0.

Usage

s3s3mirror.sh [options] <source-bucket>[/src-prefix/path/...] <destination-bucket>[/dest-prefix/path/...]

Versions

The 1.x branch (currently master) has been in use by the most number of people and is the most battle tested.

The 2.x branch supports copying between S3 and any local filesystem. It has seen heavy use and performs well, but is not as widely used as the 1.x branch.

In the near future, the 1.x branch will offshoot from master, and the 2.x branch will be merged into master. There are a handful of features on the 1.x branch that have not yet been ported to 2.x. If you can live without them, I encourage you to use the 2.x branch. If you really need them, I encourage you to port them to the 2.x branch, if you have the ability.

Options

-c (--ctime) N           : Only copy objects whose Last-Modified date is younger than this many days
                           For other time units, use these suffixes: y (years), M (months), d (days), w (weeks),
                                                                     h (hours), m (minutes), s (seconds)
-i (--iam) : Atte*** to use IAM Role if invoked on an EC2 instance
-P (--profile) VAL        : Use a specific profile from your credential file (~/.aws/config)
-m (--max-connections) N  : Maximum number of connections to S3 (default 100)
-n (--dry-run)            : Do not actually do anything, but show what would be done (default false)
-r (--max-retries) N      : Maximum number of retries for S3 requests (default 5)
-p (--prefix) VAL         : Only copy objects whose keys start with this prefix
-d (--dest-prefix) VAL    : Destination prefix (replacing the one specified in --prefix, if any)
-e (--endpoint) VAL       : AWS endpoint to use (or set AWS_ENDPOINT in your environment)
-X (--delete-removed)     : Delete objects from the destination bucket if they do not exist in the source bucket
-t (--max-threads) N      : Maximum number of threads (default 100)
-v (--verbose)            : Verbose output (default false)
-z (--proxy) VAL          : host:port of proxy server to use.
                            Defaults to proxy_host and proxy_port defined in ~/.s3cfg,
                            or no proxy if these values are not found in ~/.s3cfg
-u (--upload-part-size) N : The upload size (in bytes) of each part uploaded as part of a multipart request
                            for files that are greater than the max allowed file size of 5368709120 bytes (5 GB)
                            Defaults to 4294967296 bytes (4 GB)
-C (--cross-account-copy) : Copy across AWS accounts. Only Resource-based policies are supported (as
                            specified by AWS documentation) for cross account copying
                            Default is false (copying within same account, preserving ACLs across copies)
                            If this option is active, the owner of the destination bucket will receive full control
                            
-s (--ssl)                    : Use SSL for all S3 api operations (default false)
-E (--server-side-encryption) : Enable AWS managed server-side encryption (default false)
-l (--storage-class)		  : S3 storage class "Standard" or "ReducedRedundancy" (default Standard)

Examples

Copy everything from a bucket named "source" to another bucket named "dest"

s3s3mirror.sh source dest

Copy everything from "source" to "dest", but only copy objects created or modified within the past week

s3s3mirror.sh -c 7 source dest
s3s3mirror.sh -c 7d source dest
s3s3mirror.sh -c 1w source dest
s3s3mirror.sh --ctime 1w source dest

Copy everything from "source/foo" to "dest/bar"

s3s3mirror.sh source/foo dest/bar
s3s3mirror.sh -p foo -d bar source dest

Copy everything from "source/foo" to "dest/bar" and delete anything in "dest/bar" that does not exist in "source/foo"

s3s3mirror.sh -X source/foo dest/bar
s3s3mirror.sh --delete-removed source/foo dest/bar
s3s3mirror.sh -p foo -d bar -X source dest
s3s3mirror.sh -p foo -d bar --delete-removed source dest

Copy within a single bucket -- copy everything from "source/foo" to "source/bar"

s3s3mirror.sh source/foo source/bar
s3s3mirror.sh -p foo -d bar source source

BAD IDEA: If copying within a single bucket, do not put the destination below the source

s3s3mirror.sh source/foo source/foo/subfolder
s3s3mirror.sh -p foo -d foo/subfolder source source

This might cause recursion and raise your AWS bill unnecessarily

If you've enjoyed using s3s3mirror and are looking for a warm-fuzzy feeling, *** dropping a little somethin' into my tip jar