Self-Hosting
Proxmox vs TrueNAS vs Unraid: home-server OS
Proxmox VE vs TrueNAS vs Unraid in 2026: hypervisor vs ZFS-NAS vs flexible-array, licensing, hardware, and when to combine them — a which-OS decision map.
Problem summary
These three solve different core jobs, so the decision is a use-case map, not a ranking. Proxmox VE is a Debian-based type-1 hypervisor — VMs, LXC containers, and (its headline strength) free multi-node clustering with live migration and high availability; storage is flexible but there's no first-class NAS-share GUI. TrueNAS (Community Edition, the 25.10 'Goldeye' line — the old CORE/SCALE split was merged) is a ZFS-first storage appliance: data integrity, snapshots, replication, and SMB/NFS/iSCSI sharing first, with Docker apps and experimental VM/LXC second. Unraid (7.x) centers a unique mixed-size parity array (now ZFS-capable in pools) with Docker and VMs and the easiest learning curve — but it's the only paid option (perpetual tiers, with update-renewal on the cheaper ones). A very common answer is to combine them: Proxmox as the host with TrueNAS in a VM (HBA passed through for ZFS). Pick by whether your priority is virtualization, ZFS data integrity, or mixed-drive ease.
State your primary job.
egrep -c '(vmx|svm)' /proc/cpuinfo and check BIOS VT-x/VT-d (or AMD-V/IOMMU)
You know whether virtualization, ZFS data integrity, or mixed-drive ease is the #1 goal.
Don't run ZFS on virtualized disks — give it raw disks (or an HBA) for integrity.
Layer path
Step-by-step runbook
Start here. Do each check in order, compare it to the expected result, and stop when the evidence explains the failure or the safe stop point applies.
Pin your constraints
Check: Write down primary job, node count, drive inventory, budget, and your skill level.
Expected result: You have the inputs that decide the fit.
If not: Skipping this leads to an OS that fights your setup.
Map constraints to a primary OS
Check: Virtualization/cluster → Proxmox; ZFS NAS → TrueNAS; mixed drives/ease → Unraid.
Expected result: You have a primary candidate (a fit, not a ranking).
If not: If you need both virtualization and ZFS NAS, plan the Proxmox+TrueNAS combo.
Check hardware fit
Check: Confirm VT-d/IOMMU for passthrough, an HBA for ZFS, RAM for ARC (>=8 GB for TrueNAS), and drive sizes.
Expected result: The hardware matches the chosen OS's requirements.
If not: Size the box with the mini-PC-by-workload and mini-PC buying guides.
Plan the storage layout
Check: ZFS: choose vdev type (mirror vs RAIDZ) and width. Unraid: pick parity + data + cache pools.
Expected result: You have a redundancy plan that matches your drives and risk tolerance.
If not: Don't run ZFS on virtualized disks; don't exceed Unraid's parity-size rule.
Provision and verify
Check: Install, create the pool/array, set up VMs/containers/shares, and test a drive-failure + restore drill.
Expected result: The server runs your workloads and survives a simulated disk failure.
If not: Back up configs and verify a restore — see /self-hosting/home-server-backup-restore-tested.
Consider clustering / combining later
Check: For HA, add nodes to reach three; for both virtualization and ZFS, add a TrueNAS VM with HBA passthrough.
Expected result: You can grow into HA or a combined stack without rebuilding.
If not: Don't attempt HA on a single node — it needs quorum across three.
Decision tree
If: Virtualization first — many VMs/containers, or clustering/live-migration/HA.
Then: Proxmox VE fits best.
Action: Run Proxmox; for HA build a 3-node cluster (clustering/HA/Ceph are free). Add NAS sharing via a TrueNAS VM if needed.
If: ZFS data integrity and NAS sharing first.
Then: TrueNAS (Community Edition) fits best.
Action: Lay out ZFS vdevs (mirrors or RAIDZ), enable snapshots/replication, share over SMB/NFS; run apps via Docker.
Safe stop: Don't run ZFS on virtualized disks — give it raw disks (or an HBA) for integrity.
If: Mixed-size drives, easiest setup, one-box media/Docker/VM server.
Then: Unraid fits best.
Action: Use your largest drive(s) as parity, mix the rest as data, add SSD/ZFS cache pools; budget the right license tier for your device count.
If: You want serious virtualization AND a real ZFS NAS on one machine.
Then: Combine them.
Action: Proxmox on bare metal + TrueNAS in a VM with a dedicated HBA passed through via PCIe so ZFS gets the raw disks.
Evidence table
| Symptom | Evidence to collect | Likely layer | Next action |
|---|---|---|---|
| Choosing a home-server OS. | Primary job: virtualization vs ZFS NAS vs mixed-drive ease. | Core fit | Proxmox / TrueNAS / Unraid respectively. |
| Need HA / live migration. | Number of nodes available. | Clustering | Proxmox with >=3 nodes (free); single box can't do HA. |
| Drives are mismatched sizes. | Whether you'll grow a disk at a time. | Storage model | Unraid's flexible parity array; ZFS prefers matched vdevs. |
| Budget / update model. | Free vs Unraid's paid tiers + renewal. | Licensing | Proxmox/TrueNAS free; Unraid paid (verify tiers on unraid.net). |
Commands and settings paths
Confirm CPU virtualization + IOMMU for passthrough
egrep -c '(vmx|svm)' /proc/cpuinfo and check BIOS VT-x/VT-d (or AMD-V/IOMMU)
Where: On the candidate host (live ISO or current OS).
Expected: Virtualization extensions are present and IOMMU can be enabled — required for GPU/HBA passthrough.
Failure means: No VT-d/IOMMU means no clean PCIe passthrough — the Proxmox+TrueNAS combo and GPU passthrough won't work well.
Safe next step: Enable VT-x/VT-d (AMD-V/IOMMU) in BIOS, or pick a single-OS path that doesn't need passthrough.
Identify an HBA for ZFS (if combining or running TrueNAS)
lspci | grep -i 'sas\|hba\|raid' — confirm an IT-mode HBA or direct SATA/NVMe
Where: On the host.
Expected: A non-RAID HBA (IT mode) or direct controller exists to give ZFS raw disks.
Failure means: A hardware-RAID controller hides disks from ZFS and breaks its integrity model.
Safe next step: Flash the HBA to IT mode or use direct SATA/NVMe; pass it through to the TrueNAS VM.
Check drive sizes against the Unraid parity rule
Storage UI / lsblk — list each disk size
Where: Before building an Unraid array.
Expected: Your largest drive is (or will be) the parity disk; no data disk exceeds it.
Failure means: A data disk larger than parity isn't protected — Unraid won't allow it.
Safe next step: Assign the biggest drive(s) as parity; plan capacity around that rule.
Verify current Unraid pricing and tier limits
Open https://unraid.net/pricing
Where: In a browser, at decision time.
Expected: The tier you need (device count + update window) matches your build and budget.
Failure means: Pricing/tiers change; the Starter tier's device cap or the renewal fee can surprise you.
Safe next step: Pick the tier for your device count; budget the annual extension if not Lifetime.
Hardware and platform boundary
Change only when
- Choose Proxmox when you'll run many VMs/containers or want free clustering/HA/live-migration (build three nodes for HA).
- Choose TrueNAS when ZFS data integrity and NAS sharing are the point — checksums, snapshots, replication.
- Choose Unraid when you have mixed-size drives, want the easiest setup, and a one-box media/Docker/VM server.
- Combine Proxmox + TrueNAS VM (HBA passed through) when you want both virtualization and a real ZFS NAS.
Evidence that matters
- VT-d/IOMMU support for GPU/HBA passthrough; an IT-mode HBA for ZFS.
- >=8 GB RAM (more helps ZFS ARC); matched drives for RAIDZ, or any sizes for Unraid.
- The right Unraid license tier for your device count (if you pick Unraid).
Evidence that does not matter
- ECC RAM as a hard requirement — it's better but not mandatory; 'ZFS needs ECC' is a myth.
- The '1 GB RAM per 1 TB' rule — a rule of thumb, not a requirement for home pools.
- Maximizing core count — home-server load is usually I/O and RAM bound, not CPU bound.
Avoid
- Running ZFS on virtualized disks or a hardware-RAID controller (breaks ZFS's integrity model).
- Planning HA on a single node — Proxmox HA needs three for quorum.
- Exceeding Unraid's parity-size rule (a data disk larger than parity isn't protected).
Related tool
Use the linked tool to turn this runbook into a guided check for your exact setup.
NAS setup plannerRelated problems
Last reviewed
2026-06-23 · Reviewed by HomeTechOps. Built from June-2026 research verified against the Proxmox VE wiki (hypervisor, cluster/HA, PCI passthrough), TrueNAS docs (Community Edition / 25.10 Goldeye, ZFS, 8 GB RAM guidance) and Unraid's official pricing/array docs (tiered perpetual licensing, mixed-size parity array). Framed as a which-OS use-case map across virtualization/storage/ease/licensing, never a 'winner'; version numbers and Unraid pricing flagged as volatile to re-verify before relying on exact figures.
Sources/assumptions
- Versions move fast: Proxmox VE 9.x (9.0 on Debian 13), TrueNAS Community Edition 25.10 'Goldeye', Unraid 7.x are the 2026 lines — verify the exact point release before relying on it.
- Unraid pricing (Starter/Unleashed/Lifetime tiers + the annual update-extension on the non-Lifetime tiers) is the most volatile fact here; figures are described as a model — confirm current prices on unraid.net.
- This is a which-OS decision map across virtualization, storage model, ease, and licensing — not a winner; the right pick (or combination) depends on your priority.
Source-backed checks
HomeTechOps turns official docs and conservative safety rules into a shorter runbook. These links are the source trail for the page direction.
Planning a purchase?
We keep a source-backed, price-free comparison so you can buy once and right. No star ratings, every spec cited.
Mini PC for a Home Server: N150 vs MS-01 vs Tiny →