Open source lakeFS, Git for data lakes, on Ubuntu 24.04 LTS by cloudimg. Bring Git style branching, committing and merging to the data in your object storage. A per VM admin access key and a per VM encryption key are created on first boot (no shipped secrets), the web UI and lakectl CLI are fronted by nginx on port 80, and the versioned data plus metadata live on a dedicated Azure data disk. Apache 2.0 licensed. 24/7 expert support. Not affiliated with or endorsed by Treeverse or lakeFS.
## lakeFS 1.83 on Ubuntu 24.04 LTS by cloudimg
lakeFS is open source data version control, often described as Git for data lakes. It brings Git style branching, committing, merging and rollback to the data in your object storage, so data teams can build reproducible, isolated and safe data workflows over the same storage they already use. The cloudimg image installs the official pinned lakeFS 1.83 server and lakectl client binaries, verified against a sha256 checksum, and runs the server directly under systemd as a dedicated lakefs system user. Docker is not required. You interact with lakeFS through its web UI, the lakectl command line client, the REST API, or the S3 compatible gateway. Backed by 24/7 expert support.
No Shipped Secrets
lakeFS holds two secrets that must never be shared across machines: the admin access key and secret that authenticate API and UI access, and the encryption key that protects all stored credentials. This image ships neither. On the first boot of every VM a random encryption key is created and a headless setup generates a per instance admin access key and secret, then proves that the admin authenticates while anonymous access is rejected. The admin credentials are written to a root only file at /root/lakefs-credentials.txt.
Self Contained Storage on a Dedicated Disk
The image ships a local block adapter for the versioned data and an embedded key value store for the metadata, both on a dedicated 20 GiB Azure data disk mounted at /var/lib/lakefs. nginx on port 80 fronts the lakeFS server, which binds to loopback only, and an unauthenticated health endpoint answers load balancer probes. For production you point the block store at Azure Blob Storage or S3 and can back the metadata with PostgreSQL.
What is Included
* lakeFS 1.83, the official pinned server binary, run under systemd as an unprivileged lakefs system user
* The lakectl command line client for branching, committing and merging your data from the shell
* nginx on port 80 fronting the loopback bound web UI, REST API and S3 gateway, with an unauthenticated health probe
* A per VM admin access key and secret plus a per VM encryption key generated on first boot, no shipped secrets
* A local block store and embedded key value metadata store on a dedicated Azure data disk
* lakefs.service and nginx.service as systemd units, enabled and active
Use Cases
Reproducible data pipelines with isolated branches for experiments, safe ingestion with commit and rollback, continuous integration for data, and a fully controlled self hosted alternative to hosted data version control. Point the block store at Azure Blob or S3 to version data already in your object storage.
lakeFS is licensed under the Apache License 2.0. lakeFS is a trademark of Treeverse Ltd.; cloudimg is not affiliated with or endorsed by Treeverse or the lakeFS project. All product and company names are trademarks or registered trademarks of their respective holders and are used only to identify the software.