By ByteBot
MinIO was archived on February 13, 2026. No new commits. No security patches. Known CVEs sitting unfixed in a read-only repository while AIStor — the commercial successor — quietly absorbs former community users. If you are still running MinIO in production, you are running unsupported infrastructure. RustFS 1.0 hit general availability on September 16, and it is the most credible Apache 2.0 replacement to land yet. But “production-ready” requires unpacking.
Why the Urgency Is Real
MinIO did not die overnight. The company raised $126 million, built an ecosystem that crossed one billion Docker pulls, then spent five years methodically walking it back — AGPL relicense, admin console removal, binary distribution dropped, and finally the February archival. The community got a read-only tombstone.
The security math is unforgiving. CVEs have been disclosed against the last community release. MinIO Inc. ships fixes inside AIStor; none reach the public codebase. Every day your deployment stays on archived MinIO is a day you are accumulating exposure with no upstream remedy. That is not hypothetical risk — it is the current state.
What RustFS 1.0 GA Actually Ships
After 2.5 years of development — first commit February 2024, open-source July 2025, beta April 2026, RC August 2026 — RustFS 1.0.0 is out with a substantial feature set covering the core of what MinIO’s community edition once offered:
- Full S3-compatible API: PUT, GET, DELETE, multipart upload, versioning, lifecycle management
- Erasure coding (GA, not preview)
- IAM with policies, OIDC, SSO
- Web console included — the admin interface MinIO stripped from community users
- WebDAV, Swift API, FTP/FTPS, and MCP support
- Single binary, x86-64 and ARM64 (GNU and musl), SBOM and provenance artifacts included
- Apache 2.0 license — the dealbreaker for anyone who could not live with AGPL
The project sits at 33,000+ GitHub stars, 10 million+ Docker Hub pulls, and 160+ contributors. That is not a weekend experiment — that is an ecosystem with momentum.
Three Paths to Migrate From MinIO
Migration strategy depends on your deployment. Three options exist, each with a different risk profile:
Binary swap — the lowest-friction path for same-machine moves. Since alpha.89 in March 2026, RustFS reads MinIO’s on-disk format directly. Point RustFS at the existing data directory, keep the same ports and credentials, and your buckets, objects, and IAM policies show up intact. This is currently preview-scoped with documented encryption caveats, so test in staging before production.
mc mirror — cross-server moves using MinIO Client. Standard object-by-object transfer with no on-disk format assumptions. Slower than a binary swap but more straightforward and does not depend on format compatibility.
Godwit Sync — for production environments where you need version history, Object Lock compliance, and chunk-level resume. If you are moving regulated data or cannot afford to re-transfer on failure, this is the option to evaluate.
For a new deployment, the Docker quick start gets you running on ports 9000 (S3 API) and 9001 (web console):
services:
rustfs:
image: rustfs/rustfs:latest
ports:
- "9000:9000"
- "9001:9001"
volumes:
- ./data:/data
environment:
RUSTFS_ACCESS_KEY: your-access-key
RUSTFS_SECRET_KEY: your-secret-keyDo not use rustfsadmin as either credential. Generate real keys.
The Rust Advantage — and Where It Ends
The no-GC architecture is not marketing. In January 2026, a MinIO cluster hit a 340-millisecond GC pause during a batch ingest job writing 2,000+ small objects per second. That pause caused an ML pipeline’s checkpoint-write retry logic to time out and fail. Rust does not have a garbage collector — that failure mode does not exist in RustFS by design.
On benchmarks: RustFS leads on PUT-heavy workloads and small-object write IOPS. It still trails MinIO on large sequential GET bandwidth, though that gap is closing. For write-intensive workloads — ML checkpointing, log ingestion, event storage — the performance case is real.
What Is Not Production-Ready Yet
The GA label covers the core engine. Several surrounding features are still preview:
- S3 Tables / Iceberg native storage — preview only, not for production analytics pipelines
- On-disk MinIO binary swap — works, but preview-scoped with encryption caveats
- Multi-site replication — functional, fewer battle-tested runbooks than MinIO once had
RustFS has had growing pains in public. A hardcoded RPC authentication bypass was reported and patched, and early documentation made overclaims the team subsequently corrected. These are normal project maturity signals — but they matter when planning a production migration.
If your primary requirement is maximum production confidence and you are not constrained by the MinIO timeline, SeaweedFS has more production validation across mixed workloads. If you need Apache 2.0, an active Rust codebase, no GC, and a web console not locked behind a commercial tier — RustFS 1.0 is the clearest answer available right now.
The Bottom Line
RustFS 1.0 is not “use this blindly.” It is “this is the project that earned a 1.0 by solving the right problem at the right time.” MinIO left a gap that needed a production-grade, permissively-licensed, actively-maintained replacement. RustFS is the first to credibly fill it. The binary drops to Docker in minutes, and the GA announcement covers the full feature list. If you are running MinIO today, the migration path exists. The question is no longer whether to move — it is when and how.
