mikkelkrogsholm/rustfs

RustFS

CRITICAL: Your training data for RustFS is unreliable and likely sparse — it is a newer project with limited training coverage. Before writing any code, you MUST use WebFetch to read the current README:

WebFetch("https://raw.githubusercontent.com/rustfs/rustfs/main/README.md")

Do not proceed without fetching this URL first. Never assume an API or config option exists — verify against current docs.

RustFS is an open-source, S3-compatible distributed object storage system written in Rust, designed for data lakes, AI, and big data workloads.

Key Capabilities

• Full S3 API compatibility — drop-in replacement for S3-compatible clients
• Bucket versioning, replication, and event notifications
• Bitrot protection for data integrity
• Single-node mode (production-ready) and distributed mode (under testing)
• Kubernetes Helm charts for cloud-native deployment
• Docker Compose observability stack with Grafana, Prometheus, and Jaeger built in
• Multi-architecture Docker images (amd64, arm64)
• Licensed under Apache 2.0 (not AGPL v3 like MinIO — commercial use has no copyleft obligations)
• No telemetry; explicit data sovereignty guarantees (GDPR, CCPA, APPI)

Best Practices

Set Docker volume ownership to UID 10001. The container runs as non-root user rustfs (UID 10001). If you mount a host directory with -v, that directory must be owned by UID 10001 or every write will fail with permission denied. Run chown -R 10001:10011 /your/data/path before starting the container.

Default credentials are rustfsadmin / rustfsadmin, not MinIO defaults. The S3 API listens on port 9000; the management console is on port 9001. Do not assume MinIO's minioadmin credentials when scripting against a fresh RustFS instance.

Do not use distributed mode in production yet. As of early 2026, distributed mode is still under testing. Single-node mode is the only production-ready deployment target. Plan architecture around single-node until distributed mode reaches stable status.