s3-bsync

Bidirectional syncing tool to sync local filesystem directories with S3 buckets. Developed by Josh Stockin and licensed under the MIT License.

Work in progress (v0.1.0). Not in a functional or usable state. Do NOT use this unless you know what you are doing.

Behavior

After an initial sync (manually handling conflicts and uncommon files), the S3 bucket maintains precedence. Files with the same size and modify time on both hosts are ignored. A newer copy of a file always overwrites the corresponding old, regardless of changes in the old. (In other words, there is no manual conflict resolution after the first sync. Conflicting files are handled automatically as described here. This script is meant to run without input or output by default, in a cron job for example.) Untracked files, in either S3 or on the local machine, are copied to the opposite host and tracked. Tracked files that are moved or removed on either host are moved or removed on the corresponding host, with the tracking adjusted accordingly. Ultimately, after a sync, the .state.s3sync state tracking file should match the contents of the S3 bucket's and local synced directories.

Installation

Depends on python3 and aws-cli. Both can be installed with your package manager. Requires Python modules pip and setuptools if you want to install on your system path using one of the methods listed below. Python module python-gnupg optionally required if you wish to use GPG encryption options.

Install with one of the following:

• ./install.sh [<python interpreter>?] (Preferred)
• python3 -m pip install .
• python3 ./setup.py install (Not recommended)

Uninstall with one of the following:

• ./install.sh uninstall [<python interpreter>?] (Preferred)
• python3 -m pip uninstall s3-bsync

install.sh is a frontend for pip (un)install, configured by setuptools in setup.py. The script automatically performs compatibility checks on Python interpreter and other required dependencies.

Root permissions are not required. This program does not manage S3 authentication or aws-cli credentials. You must do this yourself with the aws configure command, or through some other means of IAM/S3 policy.

Usage

usage: s3-bsync [--help] [--version] [--init] [--debug] [--dryrun] [--file SYNCFILE]
                [--dump] [--purge] [--overwrite] [--dir PATH S3_DEST] [--rmdir RMPATH]

Bidirectional syncing tool to sync local filesystem directories with S3 buckets.

optional arguments:
  --help, -h, -?      Display this help message and exit.
  --version, -v       Display program and version information and exit.

program behavior:
  The program runs in sync mode by default.

  --init, -i          Run in initialize (edit) mode. This allows tracking file
                      management and directory options to be used. (default: False)
  --debug             Enables debug mode, which prints program information to stdout.
                      (default: False)
  --dryrun            Run program logic without making changes. Useful when paired with
                      debug mode to see what changes would be made. (default: False)

tracking file management:
  Configuring the tracking file.

  --file SYNCFILE     The s3sync state file used to store tracking and state
                      information. It should resolve to an absolute path. (default:
                      ['~/.state.s3sync'])
  --dump              Dump s3sync state file configuration and exit. (default: False)
  --purge             Deletes the tracking configuration file if it exists and exits.
                      Requires init mode. (default: False)
  --overwrite         Overwrite tracking file with new directory maps instead of
                      appending. Requires init mode. (default: False)

directory mapping:
  Requires initialize mode to be enabled.

  --dir PATH S3_DEST  Directory map to detail which local directory corresponds to S3
                      bucket and key prefix. Can be used multiple times to set multiple
                      directories. Local directories must be absolute. S3 destination in
                      `s3://bucket-name/prefix` format. Example: `--dir
                      /home/josh/Documents s3://joshstockin/Documents`
  --rmdir RMPATH      Remove tracked directory map by local directory identifier.
                      Running `--rmdir /home/josh/Documents` would remove the directory
                      map from the s3syncfile and stop tracking/syncing that directory.

Source files

setup.py manages installation metadata. install.sh handles installation and uninstallation using pip.

Created files and .s3syncignore

The default file used to store sync information is ~/.state.s3sync, but this location can be reconfigured. The file uses the binary s3sync file format specified later in this document. If you want to intentionally ignore untracked files, use a .s3syncignore file, in the same manner as .gitignore.

s3sync file format

The .state.s3sync file saved in home directory defines the state of tracked objects from the specified S3 buckets and key prefixes used in the last sync.

Control bytes

90 - Begin bucket block
91 - End bucket block
92 - Begin directory map
93 - End directory map
94 - Begin object block
95 - End object block
96 - ETag type MD5
97 - ETag type null-terminated string (non-MD5)
98
99
9A - Begin metadata block
9B - End metadata block
9C
9D - File signature byte
9E
9F - File signature byte

File structure

Version 1 of the s3sync file format.

Header {
    File signature - 4 bytes - 9D 9F 53 33
    File version   - 1 byte  - 01
}
Metadata block {
    Begin metadata block control byte - 9A
    Last synced time                  - 8 bytes uint
    End metadata block control byte   - 9B
}
Bucket block {
    Begin bucket block control byte - 90
    Bucket name                     - null-terminated string
    Directory map {
        Begin directory map block control byte - 92
        Path to local directory                - null-terminated string
        S3 key prefix (no `/` termination)     - null-terminated string
        Compress (gzip level)                  - 0-11 (1 byte)
        Recursive sync                         - 1 byte boolean
        GPG encryption enabled                 - 1 byte boolean
        GPG encryption email                   - null-terminated string
        End directory map block control byte   - 93
    }...
    Recorded object {
        Begin object block control byte - 94
        Key                             - null-terminated string
        Last modified time              - 8 bytes uint
        ETag type                       - 96 or 97
        ETag                            - 16 bytes or null-terminated string
        File size                       - 8 bytes uint
        End object block control byte   - 95
    }...
    End bucket block control byte - 91
}...

Copyright

This program is copyrighted by Joshua Stockin and licensed under the MIT License.

A form of the following should be present in each source file.

s3-bsync Copyright (c) 2022 Joshua Stockin
<https://joshstock.in>
<https://git.joshstock.in/s3-bsync>

This software is licensed and distributed under the terms of the MIT License.
See the MIT License in the LICENSE file of this project's root folder.

This comment block and its contents, including this disclaimer, MUST be
preserved in all copies or distributions of this software's source.

<https://joshstock.in> | josh@joshstock.in | joshuas3#9641