2 unstable releases
✓ Uses Rust 2018 edition
|new 0.2.0||Mar 14, 2019|
|0.1.0||Mar 13, 2019|
#183 in Command line utilities
4 downloads per month
A high-performance, caching Git LFS server with an AWS S3 back-end.
AWS S3 permanent storage back-end.
A configurable local disk cache to speed up downloads (and reduce your S3 bill).
Corruption-resilient local disk cache. Even if the disk is getting blasted by cosmic rays, it'll find corrupted LFS objects and purge them from the cache transparently. The client should never notice this happening.
Encryption of LFS objects in both the cache and in permanent storage.
Separation of GitHub organizations and projects. Just specify the org and project names in the URL and they are automatically created. If two projects share many LFS objects, have them use the same URL to save on storage space.
A tiny (<10MB) Docker image (jasonwhite0/rudolfs).
The back-end storage code is very modular and composable. PRs for implementing other storage back-ends are welcome. If you begin working on this, please let us know by submitting an issue.
- There is no client authentication. This is meant to be run in an internal network with clients you trust, not on the internet with malicious actors.
All LFS objects are encrypted with the xchacha20 symmetric stream cipher. You must generate a 32-byte encryption key before starting the server.
Generating a random key is easy:
openssl rand -hex 32
Keep this secret and save it in a password manager so you don't lose it. We will pass this to the server below.
- If the key ever changes, all existing LFS objects will become garbage. When the Git LFS client attempts to download them, the SHA256 verification step will fail.
- LFS objects in both the cache and in permanent storage are encrypted. However, objects are decrypted before being sent to the LFS client, so take any necessary precautions to keep your intellectual property safe.
For testing during development, it is easiest to run it with Cargo. Create
a file called
test.sh (this path is already ignored by
# Your AWS credentials. export AWS_ACCESS_KEY_ID=XXXXXXXXXXXXXXXXXXXX export AWS_SECRET_ACCESS_KEY=XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX export AWS_DEFAULT_REGION=us-west-1 # Change this to the output of `openssl rand -hex 32`. KEY=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx cargo run -- \ --cache-dir cache \ --host localhost:8080 \ --s3-bucket foobar \ --max-cache-size 10GiB \ --key $KEY
Note: Always use a different S3 bucket, cache directory, and encryption key than what you use in your production environment.
To run in a production environment, it is easiest to use
.envfile next to
docker-compose.ymlwith the configuration variables:
AWS_ACCESS_KEY_ID=XXXXXXXXXXXXXXXXXXXQ AWS_SECRET_ACCESS_KEY=XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX AWS_DEFAULT_REGION=us-west-1 LFS_ENCRYPTION_KEY=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx LFS_S3_BUCKET=my-bucket LFS_MAX_CACHE_SIZE=10GB
Use the provided
docker-compose.ymlfile to run a production environment:
docker-compose up -d
[Optional]: It is best to use nginx as a reverse proxy for this server. Use it to enable TLS. How to configure this is better covered by other tutorials on the internet.
- A bigger cache is (almost) always better. Try to use ~85% of the available disk space.
- The cache data is stored in a Docker volume named
rudolfs_data. If you want to delete it, run
docker volume rm rudolfs_data.
AWS credentials must be provided to the server so that it can make requests to
the S3 bucket specified on the command line (with
Your AWS credentials will be searched for in the following order:
- Environment variables:
- AWS credentials file. Usually located at
- IAM instance profile. Will only work if running on an EC2 instance with an instance profile/role.
The AWS region is read from the
variable. If it is malformed, it will fall back to
us-east-1. If it is not
present it will fall back on the value associated with the current profile in
~/.aws/config or the file specified by the
variable. If that is malformed or absent it will fall back to
.lfsconfig to the root of your Git repository:
[lfs] url = "http://gitlfs.example.com:8080/api/my-org/my-project"
Optionally, I also recommend changing these global settings to speed things up:
# Increase the number of worker threads git config --global lfs.concurrenttransfers 64 # Use a global LFS cache to make re-cloning faster git config --global lfs.storage ~/.cache/lfs
This was developed at Environmental Systems Research Institute (Esri) who have graciously allowed me to retain the copyright and publish it as open source software.