ScaleFS

Reliable storage that resizes as your needs change.

ScaleFS is a copy-on-write file system with redundancy you control.
Add, remove and replace disks — grow or shrink a pool and choose your level of protection — all without taking your data offline.

Download ScaleFS View on GitHub

Always consistent

Copy-on-write means a write either lands whole or not at all — data is never left half-written.

No downtime

Reconfigure the pool live — capacity, failover and layout all migrate online.

Yours to move

Disk pools can travel between computers, internally or in external enclosures.

Setup only requires two choices

Choose any combination of drives.
Then choose your protection level.

No RAID levels to learn, no layouts to hand-tune. ScaleFS builds a pool that fits.

1

Pick your drives

Choose the disks to include — any sizes, any mix, internal or external.

2

Choose your protection level

Pick your level of protection — how many drives can fail without losing data. More safety or more space, your call.

3

ScaleFS builds your pool

It lays out a protected pool that's sized to your drives, ready to use immediately.

Available now

A file system for the modern age

Underneath the volume manager is a modern file system — copy-on-write, checksummed end to end, and resilient against power loss.

copy-on-write

Snapshots in an instant

Because ScaleFS is copy-on-write, a snapshot is just a named point in time — no duplicate copies, no reserved space, no downtime. Pull yesterday's version of a file out of a snapshot while the live system keeps moving.

CRC-32C

Every byte checked, every read

All data and metadata carries a checksum, verified on every read. fsck catches a single flipped bit — and names the file it damaged. Silent corruption doesn't stay silent.

atomic commits

Crash-safe by design, validated in the real world

Every commit is a barrier-ordered, all-or-nothing transaction.

POSIX

Behaves like a file system should

Hard links, symlinks, special files, extended attributes and setcap capabilities all round-trip.

zero-dependency Rust

Nothing between you and your data

A dependency-free core written in Rust — every line that touches your bytes lives in an open-source, auditable repository. Two self-locating superblock copies per drive mean a damaged head or tail can't hide your pool.

salvage

Recovery tools for extreme scenarios

If you lose more drives than your protection covers, read-only recovery tools pull back every file that can still be read and flag anything that can't. Open APIs let other software or AI agents do the same.

The volume manager

A pool that bends to your hardware.

Set up a multi-drive pool today. Here's what the volume manager does.

Multi-drive pools

One pool spans many drives, and every membership change happens while you keep working.

  • Add, replace and remove drives without downtime — when a drive fails, ScaleFS rebuilds its data onto a replacement from the copies on the surviving drives
  • Protect with n-way mirroring or Reed-Solomon parity — tolerate up to 3 drive failures; data and snapshots keep serving
  • scrub re-checks every copy of every byte, repairing any silent corruption from a verified-good copy on another drive
  • Protection is a policy — how many drives may fail — not a fixed RAID mode

Elastic by policy

Size and safety are dials you turn any time — not decisions you're stuck with.

  • Grow or shrink capacity on demand — add drives, or swap small ones for large
  • Decide how many drives may fail — trade capacity for safety, your call
  • Start on a single disk and grow to many — then shrink back down whenever you like, all with no downtime

Runs on your hardware

From a single internal drive to multi-bay external stations — with no enclosure lock-in; move your drives anytime.

  • Internal drives or external single- and multi-drive stations
  • Pools move between computers — every drive carries the pool's identity
  • Understands dual-actuator drives — positioning data safely and keeping both heads busy
  • The scalefs command-line tool works on macOS, Linux and Windows

From home to enterprise

The same file system whether it's one spare disk or a rack of them.

  • Start with a single drive at home; grow into an enterprise array
  • You choose the drives and how many can fail — ScaleFS lays out the rest
  • One open-source build for everyone — no editions, no per-seat licensing
  • Move your pool between machines as your hardware changes
Usage

Configure it once. Change your mind anytime.

Drive ScaleFS from the scalefs command-line tool.

1

Set your failure tolerance

Choose how many disks can fail, from 0 up to n−1. Raise or lower it later without rebuilding.

2

Optimize for space or speed

Tune the pool toward storage capacity, performance, or an in-between balance.

3

Grow the pool

Add disks to increase capacity, or swap smaller disks for larger ones — live.

4

Shrink or heal

Remove disks to reclaim them, or replace a failed disk. Reduce capacity or fail over on the fly.

Terminal
$ scalefs pool create tank.img --size 1G --label tank $ scalefsd serve $ scalefs daemon attach tank tank.img $ scalefs daemon mount tank /mnt/tank ✓ /mnt/tank ready

Get ScaleFS

Try ScaleFS today; cross-platform, optimized for Linux.

Download ScaleFS View on GitHub