What to Do When Your Unraid Server Hardware Fails

If your Unraid server’s motherboard, power supply, or storage controller fails and the drives are still good, start by recovering the original configuration. On a typical USB-booted system, that means reusing the healthy Unraid boot USB or restoring a recent backup of it. Keep the array stopped until you’ve matched every data disk, parity disk, and pool member to its expected slot.

Mike Wieger and I talked about this on the Home Gadget Geeks tenth anniversary episode. At about 23 minutes, I described getting my system back with the original USB drive. Mike had moved his Unraid setup twice after hardware failures and pointed out that the replacement hardware could need new network mappings. That’s the 2020 story. The recovery workflow below follows Unraid’s current documentation.

1. Record the old configuration before moving anything

If the old system is still accessible, stop the array, disable array auto-start in Settings → Disk Settings, and capture the Main tab. Record full drive serial numbers and slot assignments for Parity, Parity 2, every numbered data disk, and every pool member. Save a boot-device backup off the server, then shut down cleanly. If the old host is already dead, use your last backup, screenshots, or drive labels. Pay attention to disks added or replaced after the backup was made.

The boot device’s config directory holds your settings. Unraid identifies config/super.dat as the array assignment file. Back up the entire boot device and keep a separate copy of anything you’ll need to unlock encrypted storage. The boot backup covers configuration. Keep a separate backup of the files, appdata, and VM disks you need to recover. See Unraid’s boot-device backup instructions and its data recovery guide.

2. Bring up the replacement host with the array stopped

On a conventional USB installation, use the healthy original boot device and select it in the replacement machine’s BIOS or UEFI. If you couldn’t disable auto-start on the old server, leave the storage drives disconnected for the initial configuration boot, turn auto-start off, then shut down before reconnecting them. Avoid saving any new disk assignments during that first boot. If the web interface doesn’t answer at the old address, check the local console and the router’s DHCP leases. A new NIC can change the interface mapping or invalidate an old DHCP reservation.

Connect drives to a controller that presents them individually. Unraid recommends HBA/IT mode for supported controllers, so the OS can see each disk directly. In Main, compare the reported identifiers against your records. Linux names such as /dev/sdb can change. A missing drive or unexpected identifier is a reason to check power, cables, controller mode, and device visibility before touching assignments. See Unraid’s startup troubleshooting and hardware FAQ.

3. Check the assignments before starting the array

• Every drive matches: Confirm both parity positions if you use dual parity, then verify numbered data slots and pool membership. Review the array’s proposed action before starting. For the first validation, leave Docker and VM services disabled if practical. Check that the expected filesystems mount, open a few known files from important shares, and confirm appdata and VM storage paths before bringing services back.
• Something doesn’t match: Stop and collect Tools → Diagnostics. Resolve missing disks, unexpected assignments, or a proposed clear or rebuild before proceeding. Assigning a data disk to a parity slot can overwrite it when parity is built. A routine host move with intact assignments doesn’t require Tools → New Config.
• A previously working disk is unmountable: Leave Format alone. It creates an empty filesystem and, for a protected array disk, updates parity to reflect that change. Follow the filesystem troubleshooting guide for the actual filesystem and storage type before attempting repair.

If the boot USB failed too

Restore a recent boot backup to new media using the official replacement procedure. Unraid Connect users can retrieve an automated flash backup if they previously enabled it. Confirm the replacement boots before transferring a USB-bound license via Tools → Registration → Replace Key. The transfer blacklists the old device. Internal boot and TPM-based licensing have their own rules, so follow the licensing documentation for your setup.

If you have no usable configuration backup and can’t identify the parity disks with confidence, stop before assigning parity or starting a normal array. Unraid has a separate lost boot device recovery procedure. Get a second set of eyes on the assignments if they’re unclear. Leave “Parity is already valid” unchecked if you had to infer the positions of two parity disks.

What I’d save before the next failure

Keep an off-server boot backup and a dated screenshot of disk assignments. Update both when you swap a drive or change a pool. Keep a separate copy of critical data and test opening a few files from it. That gives you a way to recover the server configuration and a way to recover the files if the hardware failure turns out to be more than a motherboard.

For the original hardware-move discussion, start at [23:07] in HGG471’s transcript. For the other side of the problem, see what to do when the drive itself fails and the warranty replacement is out of stock.