TerraMaster Forum

Subject: F4-425 Pro — boot failure after simple network IP change (brand new unit, no data yet)
Model: F4-425 Pro
TOS Version: Factory TOS7 Pro unit 7.0.0750
BIOS - BJFX-TWLX-108
Disks: 1x SSD/NVMe (256GB, used for OS + Docker), 6TB HDDs not yet installed
RAID Configuration: None yet — single SSD, no array built
Issue:
Brand new unit, freshly initialized. After successfully installing Docker/Portainer and migrating several containers over (working fine), I made a simple network change — changed the static IP address via Control Panel → Network — and rebooted. The unit failed to boot afterward.
Boot failure details:

HDMI console shows repeated mdadm: error while loading shared libraries: libdl.so.2: cannot open shared object file: No such file or directory
Followed by No root device specified. Boot arguments must include a root= parameter.
Drops into BusyBox v1.30.1 (Ubuntu 1:1.30.1-7ubuntu3.1) rescue shell
In this rescue shell, most tools (mdadm, lsblk, fdisk) are missing or broken with the same libdl.so.2 error

Steps taken so far:

Performed the documented RESET button procedure (hold reset, tap power while holding, wait for 3 beeps, release) — unit then booted to TNAS initialization page
With SSD removed, initialization page shows "no disks detected" (expected, since SSD wasn't installed)
Hot-plugged SSD back in while TOS initialization page was active — initialization still does not detect/list the SSD as available storage
From the boot menu, used the "c" command-line option which appears to be a GRUB prompt
In GRUB, ran ls — shows (hd0), (hd0,gpt1) [internal boot flash, ~240MB], and (hd1) with (hd1,gpt1) through (hd1,gpt4) — presumably the NVMe SSD
ls (hd1,gpt1)/ shows boot and EFI folders (readable)
ls (hd1,gpt2)/, ls (hd1,gpt3)/, ls (hd1,gpt4)/ all return "unknown filesystem" — GRUB cannot read any of these partitions

Current state:

SSD is physically installed
GRUB can see the NVMe device and its 4 partitions, but only the boot/EFI partition (gpt1) is readable
TOS initialization wizard does not detect the SSD as available storage
Console/BusyBox rescue environment cannot run basic diagnostic tools due to library errors

Question:
Is this recoverable, or does this indicate a failed/corrupted NVMe SSD or a deeper TOS firmware issue? Given this happened on a brand new unit after a simple, routine network IP change (no disk/RAID operations were performed beforehand), I'm concerned this may be a hardware or firmware defect — just need to understand if this unit is faulty and needs RMA, or if there's a recovery path I'm missing.

*** Update

Mounted the SSD in Proxmox to check disk
had to use both mdadm and then LVM to see folder/file structure

Able to see all folders so SSD is good.

btrfs check --readonly /dev/vg0/lv0
Opening filesystem to check...
Checking filesystem on /dev/vg0/lv0
UUID: 60e683e0-34fda5-451c-a8c4-95ffa186e5ad
[1/8] checking log skipped (none written)
[2/8] checking root items
[3/8] checking extents
[4/8] checking free space tree
[5/8] checking fs roots
[6/8] checking only csums items (without verifying data)
[7/8] checking root refs
[8/8] checking quota groups skipped (not enabled on this FS)
found 23573684224 bytes used, no error found
total csum bytes: 22849592
total tree bytes: 175702016
total fs tree bytes: 140509184
total extent tree bytes: 5111808
btree space waste bytes: 28575686
file data blocks allocated: 33843269632
referenced 23548944384

I did not get any useful information from support on this except to remove SSD and install on Spinning Disks.


For anyone that has TOS OS issues, and boot failures , and want to retain Working Partitions here's what I did.

Removed SSD and Mounted in Caddy
Attached to Proxmox,
Was able to Mount with mdadm and LVM
There are 4 Partitions sda1,sda2,sda3,sda4
sda4 is The Large drive with you data on it.
I wiped on sda1 & sda2
Installed SSD back in F245.
The initialisation wizard came up, I did manual install, and there were thankfully no messages about wiping DISK and losing data.
It did the full TOS install again, and upon boot I have recovered my full data, docker, portainer, and 10 containers.

I will try a few reboots to make sure its actually booting again, and probably clonzilla the SSD as a backup , as zero faith so far given use experience so far.

stuzza21 wrote: ↑01 Jul 2026, 06:49

Thank you for sharing this information! Based on your troubleshooting records, it appears that some files in the rootfs are either corrupted or missing. This is unrelated to the IP change — modifying the IP does not affect the rootfs. Our best guess is that the initial system installation didn't complete successfully, and some files failed to write to the SSD. After you wiped the first two partitions and reinstalled, the second attempt went through fine. We still don't know exactly why the first installation failed, and we haven't been able to reproduce the issue so far, so further investigation is needed. One more thing: an NVMe SSD inserted after the system has booted will not be detected. Please shut down the system, insert the SSD, and then power it back on.

Could you try reinstalling Docker/Portainer, reboot, and check if the same issue happens again?

All going well since re-install thanks

I did many shutdown, power pulls and testing before progressing further.
All containers back working.

BIGGEST issue i face now is I can't move the /Volume1/homes (256Gb) folder to Volume2/homes (18TB)
I did disable ControlPanel>User>Settings>EnableUserHomesDircectory

also copied via SSH

But when I enable it again, it recreates on Volume1 which will fill up in about 3 mintues once I pull all the user data accross.
THIS NEEDS to be an Option.

How do I changed Default Location for homes

stuzza21 wrote: ↑02 Jul 2026, 12:05

Glad to hear that most things are working well after the re-installation.

Regarding the homes folder location: Currently, the TOS system defaults to mounting the user home directory on the primary volume (Volume 1). Changing the default path for the homes directory is temporarily not supported, which is why it automatically recreates on Volume 1 once re-enabled.

Hi and thanks for the response.

When will it be supported, as this seems quite a large issue for those mounting OS on SSD for speed and having Spinning Disk as mass storage.
There should be an Option on which Volume to mount the homes directory.
I hope this issue is fixed, along with none of the Apps starting on reboot in next release.
I cant even replicate it on a per user basis, as Permissions are granted only at the highest level.
I would need to Create 1 Folder Per user at Volume level for Permissions.

Is there no config file that can be modified to create on /Volume2, Its stopping me from moving users across and shutting down old NAS ?