Skip to content

Project status

Last updated: 6 October 2026.

This page exists so that nobody has to guess how far along the project is. It lists what has been done, what is being worked on, and what has explicitly not been tested.

Done

  • The build runs end to end from the project's own repositories on FreeBSD stable/15 and produces an installable ISO.
  • Installation works — verified by an automated acceptance run that installs in a virtual machine and checks the result through the API.
  • Upgrade from official TrueNAS CORE 13.3-U1.2 works and preserves pools, datasets, shares, users, passwords and iocage jails. See Upgrading.
  • Containers work. podman with ocijail runs FreeBSD images and, via FreeBSD's Linux emulation, Linux images too. That required fixing the kernel module list: linux64.ko shipped but could not load because its dependency mqueuefs was never built.
  • A configuration-loss bug was found and fixed. The stock installer copied the preserved configuration into /tmp, which is a 5 MB tmpfs, and copied a 24 MB package database and 93 MB of kernel modules into it before deleting them. The copy failed with No space left on device, and the failure was not checked — the upgrade continued and silently produced a system with a default configuration. The fix checks every copy and verifies the configuration database by checksum.
  • Upstream tracking is automated — four watchers, running weekly.
  • First public build released — see Download, signed with the project key.
  • Installation and upgrade on physical hardware work — network (PXE) unattended install onto the data-array disks, upgrade into a new boot environment with the data pool untouched, verified by checksum.
  • ZFS replication works. The inherited replication engine (periodic snapshots + push over SSH and SSH+NETCAT) was verified between two machines: full stream, incremental runs and checksums all clean.
  • Updates without boot media work. The project runs its own update train: an installed system finds the new build, downloads it, applies it and reboots. Measured end to end on the stand — the system came back on the new version in about three minutes.
  • v0.1.4-dev released (8 October 2026), signed, signature checked against the copy downloaded from GitHub. Before publishing: an upgrade from TrueNAS CORE 13.3 and a clean install in a virtual machine; an update from v0.1.3-dev through the train, the way users receive it; and, on the physical machine by hand at the console, a fresh TrueNAS CORE 13.3-U1.2 upgraded to BSDnas with all 12 data-pool checksums matching, read locally and over SMB. See News.
  • Package-based updates (pkgbase) are in the released image. The train publishes a signed manifest; systems on v0.1.4-dev update by packages into a new boot environment, systems on earlier builds keep receiving updates the old way. Verified end to end on the physical test machine: data pool, partition table and configuration intact.
  • Legal cleanup — files covered by the TrueNAS Enterprise licence were removed at the outset; branding is the project's own.

In progress

  • Containers — podman and ocijail ship in the image and both FreeBSD and Linux containers run, verified on a freshly installed image. What is missing is integration: no web interface, no acceptance coverage.

Not done, and honestly so

  • The published build is a development build rather than a stable release.
  • The repository packages are not signed. The manifest is, and the check is fail-closed: the update does not start without the key or without a signature that verifies. The packages pkg then installs are a different matter — pkg is configured with signature_type: none, so nothing proves they are ours. Until pkg repo -k is part of publishing, their integrity rests on the transport, and the updater refuses a repository that is not on HTTPS.
  • Two test machines. Measurements come from one virtual stand and one physical machine (desktop board, four SATA disks, network install). No varied disk controllers, no real workloads yet.
  • Untested configurations: HA (dual controller) systems, GELI encryption, iSCSI under load, Active Directory authentication, separate cache and log devices.
  • raidz pools — only carried through an upgrade. An existing raidz2 pool came through an upgrade on the physical test machine intact, all checksums matching. Creating a raidz pool, replacing a disk in one and rebuilding it have not been tested.
  • No upgrade path from anything older than 13.3.
  • No security audit. The project inherits whatever the upstream sources contain.

How claims here are made

Every statement on this site is either a measurement or is marked as untested. When something was verified, the project's journal records the command, the output and the date. When a previous claim turned out to be wrong — as happened with the reversibility of zpool upgrade — the correction is recorded next to the original rather than quietly replacing it.

That is deliberate. A fork that overstates what it has tested is worse than no fork at all, because people put data on it.