MinIO archived its repo on April 25. The community fork already has the admin console back.

The GitHub banner on minio/minio now reads, simply, “This repository was archived by the owner on Apr 25, 2026. It is now read-only.” The README that follows is shorter than it used to be: a “NO LONGER MAINTAINED” header, then two links pointing readers at AIStor Free and AIStor Enterprise, the company’s renamed commercial successors. Eleven years of one of the most-deployed open-source object stores ended with a Friday push.

What MinIO actually shut

The archive isn’t a sudden rug-pull; it’s the punctuation at the end of a long sentence. The Pigsty team’s retrospective lays out the timeline:

• 2021: Apache 2.0 swapped for AGPL v3. Already a signal that the company wanted enterprise customers buying licenses, not running the free edition behind nginx.
• May 2025: the admin console got ripped out of the community edition. The web UI for user management, bucket policies, lifecycle rules, and access control became enterprise-only. Self-hosted users who didn’t want to drive mc from the command line had to buy AIStor.
• October 2025: pre-built binaries and Docker images stopped shipping. Anyone running MinIO from apt or docker pull had to start building from source.
• December 2025: status updated to “maintenance mode,” and an InfoQ post that month collected the rising chorus of users asking what alternative S3 implementations to migrate to (Ceph, SeaweedFS, Garage).
• February 2026: the README started saying “no longer maintained.”
• April 25, 2026: the repository went read-only.

What MinIO didn’t do, even at the end, was point users at a fork. The README points at AIStor Free and AIStor Enterprise. AIStor Free has feature gaps the old community edition didn’t. The migration is “buy a license” or “leave.”

The fork that already shipped

The community started moving in February. Ruohang Feng, the founder of the Pigsty PostgreSQL distribution, runs MinIO as a backing store for Pigsty’s WAL archives, so the dependency was production-critical for him personally. His fork lives at pgsty/minio and the maintenance scope is narrow on purpose:

• Admin console restored. The console submodule got reverted to the pre-removal commit. Users get the web UI back without an enterprise subscription.
• Binaries shipping again. Docker images and RPM/DEB packages are built through CI/CD on every commit. Drop-in replacements for the official artifacts that stopped in October.
• Docs forked. The fork hosts updated documentation at silo.pigsty.io and the CLI is mc-compatible, so existing scripts don’t change.
• Maintenance, not features. Feng’s commit is “supply-chain stability.” The plan is CVE patching and bug fixes, with AI coding tools used to keep the workload tractable for one maintainer.

That last point is what makes this fork different from the doomed forks that follow most company-led-OSS ruptures: it’s not trying to out-feature the original. It’s just trying to keep the freeze-in-amber version of MinIO running for the people whose ops are already on it. That’s a much more achievable scope.

Why this isn’t a Ruby Central situation

Ruby Central’s RubyGems crisis ended with a financial collapse and a working board taking over a community-essential package registry. MinIO’s archive is the other shape of OSS-to-enterprise: the company is fine, the product is fine, the customers are fine. What got shut is the free edition that was never the company’s strategic priority.

That’s also why the fork has runway. MinIO Inc. doesn’t need to fight pgsty/minio in court; AIStor’s pitch is to enterprises that need a vendor relationship, not to homelab users. The two can coexist as long as the fork doesn’t try to steal commercial customers (and Feng has zero interest in that).

What this means for you

If you run a MinIO cluster in production, the question isn’t whether to migrate today; it’s what your one-year plan looks like. Three options, in roughly increasing order of effort:

1. Switch to pgsty/minio. The drop-in path. You keep your mc scripts, your buckets, your access policies. You take a dependency on a one-person fork, which means CVEs get patched as Feng patches them and you should pin to commit hashes, not floating tags.
2. Migrate to AIStor Free or Enterprise. The vendor-supported path. Buy a license, get back the admin console plus official binaries and a support contract. This is the path MinIO Inc. wants you on, and if you’re a regulated shop that needs to point at a vendor for compliance, it’s the cleaner answer.
3. Migrate off MinIO entirely. Garage, SeaweedFS, and Ceph have all matured to the point where they’re real options for self-hosted S3. Garage in particular is a much smaller blast radius if you don’t need the full Ceph feature set. This is the highest effort but also the cleanest exit from the MinIO question altogether.

My read: the pgsty fork is the right call for the next 12 months for most teams. It buys time without forcing a migration during the same week the original repo went dark. But pin the version, set up a CVE alert on the fork, and start a slow eval of an alternative store in parallel. One-maintainer forks are stable until they aren’t.