Installing TrueNAS SCALE: Step-by-Step Guide with ZFS Setup

TrueNAS SCALE has become the default choice for new installations — CORE still receives security updates, but all new features land in SCALE. If you are building a NAS or a small storage server today, SCALE 25.10 Goldeye is the right platform. This guide walks through the complete initial installation: boot media, installer, ZFS pool, datasets, ACLs, and the hardening steps you should take before putting the system into production.

We assume the hardware is already in place. If you are still deciding on a configuration, the TrueNAS configurator helps put together a matching build. Details on the current SCALE release are in the TrueNAS 25.10 Goldeye overview.

Preparation: Boot Media and BIOS Settings

TrueNAS SCALE runs from a dedicated boot medium — separate from the storage pool. Two SATA DOMs or two M.2 SSDs in a mirror are the usual choices. A single USB stick works technically, but we advise against it in production: USB sticks die silently, and without a boot medium the NAS is down. 32 GB is the minimum; 64 GB or 128 GB is more comfortable.

For the installer itself, you need a second USB stick with at least 8 GB. Download the current ISO directly from the TrueNAS download page and write it with Rufus (Windows), balenaEtcher, or dd (Linux). Use Rufus in DD mode — the ISO mode causes boot problems on some boards.

Check three things in the BIOS/UEFI: UEFI boot enabled; Secure Boot can often stay enabled with SCALE 25.10, but disable it if the installer refuses to start. Enable virtualization features — VT-x, VT-d on Intel or AMD-V, IOMMU on AMD — otherwise VMs will not work later. Configure HBA cards in IT mode, not RAID mode. ZFS wants to see the raw disks, with no hardware RAID in between.

ECC RAM: Requirement or Recommendation

ECC RAM is not a formal requirement for ZFS — it runs without. In production environments we still recommend it, because ZFS cannot correct errors that originate in RAM. For homelab and test setups you can start without ECC, but plan for a migration once real data lives on the pool. Details on our recommendations are in the TrueNAS server buyer’s guide.

TrueNAS SCALE Installer: The Actual Setup Process

Boot from the installer stick and select the first menu entry — “Install/Upgrade”. The installer runs in text mode, which feels familiar if you have seen a Debian installer before. SCALE is Debian-based and inherits its installer logic.

Next, choose the boot medium. The installer lists all detected disks. Important: only tick the devices actually intended for the system — never the storage disks that will later join the pool. If you want a boot mirror, mark both boot devices. The installer then creates a boot pool named boot-pool as a ZFS mirror.

Then set the root password. This password is for the local root account and is required immediately on first login to the web UI. Choose a strong password, but remember it — resets without console access are painful.

The boot loader question is usually answered with UEFI — BIOS/Legacy only if your board really cannot do UEFI. The installer defaults to placing SWAP on every disk; leave it as is.

After a few minutes the installer reports completion. Remove the installer stick, reboot the system, and the console menu shown after boot displays the IP address where the web interface is reachable.

First Login to the Web Interface

Point your browser to the IP shown, HTTPS is on by default — the certificate is self-signed, so accept the browser warning. Log in with root and the password you set. SCALE 25.10 greets you with a dashboard that is fairly empty because no pool exists yet.

First steps: under System Settings -> General set the timezone (Europe/Berlin for us), pick your language. Under Network check that the interface is correctly configured — a static IP is clearly recommended for production systems, DHCP only for lab use. Under Network -> Global Configuration verify hostname, DNS servers, and gateway.

Creating the ZFS Pool: Choosing the Right RAID Level and Vdev Structure

The pool is the heart of every TrueNAS installation, and the decisions here are long-term — vdev layouts cannot really be changed after the fact. Under Storage, click Create Pool. Give it a short name without special characters (tank is the classic, but data or pool01 are equally fine).

The central question is the RAID level. The three realistic options for most setups:

Mirror — classic RAID-1 equivalent. Each vdev consists of two disks, capacity is halved. Very fast for random access, uncomplicated resilver. The usual choice for virtualization and databases.

RAIDZ1 — comparable to RAID-5. One parity disk, one may fail. With the large disks common today (12 TB and up), resilver takes so long that the risk of a second failure during that window becomes real. We advise against it for current disk sizes.

RAIDZ2 — two parity disks, two may fail. The sensible choice for classic file server workloads and backup targets. Usual vdev widths are 6 or 8 disks.

For very large pools, RAIDZ3 (three parity disks) can make sense, but that is an edge case. And an important rule: never mix disks of different sizes in a vdev — ZFS uses the smallest disk as reference and wastes the rest.

Special Vdevs and Cache

SCALE 25.10 makes configuration of special vdevs (metadata) and L2ARC (read cache) comfortable in the GUI. A metadata vdev on two mirrored SSDs noticeably speeds up directory operations — useful with many small files. L2ARC is only worthwhile when RAM is tight; with enough RAM (from 64 GB) it is usually unnecessary.

A SLOG (separate intent log) on a fast SSD with power-loss protection accelerates synchronous writes — relevant mainly for NFS with VMs or iSCSI targets. For classic SMB file servers without synchronous writes, a SLOG brings nothing.

If you are unsure about the vdev structure, the TrueNAS configurator helps you play through the options. We also covered the topic in more depth in our ZFS fundamentals article.

Datasets and Shares: A Structure That Will Not Hurt Later

Datasets are ZFS units below the pool. Each dataset has its own properties — compression, recordsize, snapshots, quotas. Think about the structure briefly before creating datasets, because renaming and moving them is possible but involves rework on shares, permissions, and backup jobs.

A proven structure for smaller environments:

tank/
  smb/
    projects
    archive
  nfs/
    vmware
    proxmox
  iscsi/
    hyperv
  backup/
    veeam

Under Datasets, click the pool root, then Add Dataset. Choose the dataset type by purpose: Generic for most cases, SMB if an SMB share follows directly (sets case insensitivity and matching ACL defaults), Apps for container volumes.

Recordsize is a property you should not leave the same across all datasets. For classic file server workloads the default 128K is fine. For databases and VM storage, 16K or 64K. For very large media files (video raw material), 1M can make sense. Leave compression at LZ4 — it costs practically no CPU and often saves 20 to 40 percent of space.

Setting ACLs Correctly

TrueNAS SCALE uses POSIX ACLs (on generic datasets) or NFSv4 ACLs (on SMB datasets). For pure SMB environments, NFSv4 ACL is the right choice because it comes closest to the familiar Windows model. For mixed NFS/SMB environments it gets more complicated — there, the Restricted preset in the ACL editor is a sensible starting point.

Important rule: do not set wild individual permissions for single users at the dataset level. A group-based structure is cleaner: create groups in TrueNAS (or bind Active Directory), assign permissions to groups, put users into groups.

Initial Hardening: What Should Happen Before Production

After the base install, a few steps remain before you attach the system to the production network.

Disable root login: Under Credentials -> Local Users create an admin user with sudo rights, then disable root login via SSH and web UI. The new user gets the TrueNAS Admin role.

Two-factor authentication: Under Credentials -> 2FA enable TOTP. Non-negotiable for production systems — especially because the web UI is usually reachable from the internal network.

Automatic updates — with caution: SCALE can check for updates automatically and send notifications. Automatic installation we do NOT recommend — always apply updates manually, ideally after a snapshot of the boot pool.

Configure snapshots: Under Data Protection -> Periodic Snapshot Tasks set up a snapshot plan for every important dataset. Common: daily with 30-day retention, hourly with 48-hour retention for active working directories.

Enable SMART tests: Under Data Protection -> S.M.A.R.T. Tests schedule short tests weekly and long tests monthly. Configure email alerts under System Settings -> Alert Settings.

Scrub tasks: Under Data Protection -> Scrub Tasks schedule monthly scrubs for every pool. Scrub reads the entire pool and verifies checksums — the core ZFS defense against bit rot.

If you prefer to hand over the whole build: our IT service team takes over planning, installation, and operation end to end — including handover documentation that lets a customer admin run the system independently later.

Next Steps After the Base Installation

With a working pool and the first datasets in place, the foundation is set. What comes next depends on the use case: SMB shares for file server operation, NFS datastores for Proxmox or VMware, iSCSI targets for Windows clusters, replication tasks for offsite backup, or apps and VMs for container workloads.

For backup operations, replication is the standard approach: snapshots are transferred via ZFS send to a second TrueNAS — incrementally, with checksum verification, without the detours of a classic backup tool. We cover that in a separate article.

If you have questions about the specific configuration for your use case or need help with the setup, contact us via the DATAZONE contact page or directly via the TrueNAS configurator with a matching hardware recommendation.

FAQ on the TrueNAS SCALE Installation

Can I install TrueNAS SCALE on a single USB stick?

Technically yes, practically no. USB sticks fail silently, and without a boot medium the NAS is down. For production installations use SATA DOMs or M.2 SSDs in a mirror. For pure test or lab environments a USB stick is acceptable, but should be replaced before production use.

Do I strictly need ECC RAM?

No, TrueNAS SCALE runs without ECC as well. In production environments we clearly recommend ECC, because ZFS cannot correct memory errors once they occur in RAM. For homelab and test setups you can start without ECC and migrate later.

Which RAIDZ level is the right one?

For virtualization and databases, usually mirror vdevs. For classic file servers and backup targets, RAIDZ2 with 6- or 8-disk vdevs. We avoid RAIDZ1 with current disk sizes because the resilver window is too long. RAIDZ3 is an edge case for very large pools.

What happens if I need to replace the boot medium?

With a boot mirror the swap is straightforward: replace one of the two boot disks, resilver runs automatically. With a single-disk boot you have to restore from the config backup — for this, export the config under System Settings -> General -> Manage Configuration beforehand and store it safely.

Can I migrate from TrueNAS CORE to SCALE later?

Migration from CORE to SCALE is generally possible, but is an upgrade process with preconditions (export config, remove jails and plugins). For complex existing systems the migration should be planned. We support such migrations as part of our IT service offering.

How do I keep my TrueNAS up to date?

Check updates under System Settings -> Update and apply them manually. Before every update, take a boot pool snapshot and pull a config backup. Between major releases (e.g. 24.x to 25.10), read the release notes and run tests in a lab environment if needed. We publish current release information regularly in the DATAZONE blog.