[Help] F8 SSD Plus - Frequent Crashes with Unsafe Shutdowns (503+ on boot drive)

[alienable990]

**System Hardware:**
- Model: TerraMaster F8 SSD Plus
- BIOS Version: MADN0301.V07 (07/29/2025)
- Drives: 4x Kingston KC3000 NVMe
- nvme0: KINGSTON SKC3000S512G (512GB) - BOOT DRIVE
- nvme1: KINGSTON SKC3000D4096G (4TB)
- nvme2: KINGSTON SKC3000D4096G (4TB)
- nvme3: KINGSTON SKC3000D4096G (4TB)
- OS: Proxmox VE 9.1.4

**Problem:**
System experiencing frequent, sudden crashes with complete power loss. No error messages in logs before crash - system simply stops and reboots.

**Critical Evidence:**

1. **Unsafe Shutdowns (Abrupt Power Loss):**
- nvme0 (boot drive): 503 unsafe shutdowns (was 18, now 503!)
- nvme1: 26 unsafe shutdowns
- nvme2: 179 unsafe shutdowns
- nvme3: 361 unsafe shutdowns

2. **All NVMe Drives Show Same Error Pattern:**
```
nvme0 (most recent):
Num ErrCount SQId CmdId Status PELoc LBA NSID Message
0 102 0 0x4005 0x4004 0x028 0 0 Invalid Field in Command
1 101 0 0x0005 0x4004 - 0 0 Invalid Field in Command

nvme1 (hottest drive at 81°C, showing thermal events):
Num ErrCount SQId CmdId Status PELoc LBA NSID Message
0 136 0 0x501b 0x4004 0x028 0 0 Invalid Field in Command

nvme2, nvme3: Show identical error pattern
```
- All 4 drives show identical error type
- Errors increase on every crash
- SMART overall-health still shows "PASSED" on all drives

3. **No Crash Logs:**
- System stops logging completely before crash
- No kernel panic, no OOM killer, no thermal shutdown messages
- Last log entry is always normal authentication event
- Next log entry is from boot after crash

4. **Typical Crash Pattern:**
- Uptime varies: 1 minute to several hours
- Some crashes occur back-to-back (1-3 minutes apart)
- No correlation with CPU load, disk I/O, or specific VMs

**Troubleshooting Already Done:**

1. Tested multiple Proxmox kernels (6.17.4-2, 6.17.2-1)
2. Added kernel parameters: `nvme_core.default_ps_max_latency_us=0 pcie_aspm=off`
3. Checked temperatures: All drives within normal range (32-51°C, nvme1 at 81°C with thermal events)
4. Checked SMART data: All drives report healthy
5. Verified memory usage: Plenty available RAM
6. Power supply: Denmark has very stable grid power
7. No USB passthrough or GPU passthrough to VMs

**System Logs Excerpt (Before Latest Crash):**
```
Jan 25 18:47:29 pve kernel: ACPI Error: Aborting method \_SB.PC00.XHCI.RHUB.HS02._PLD
Jan 25 18:47:31 pve smartd[768]: Device: /dev/nvme0, NVMe error count increased from 100 to 102
Jan 25 18:47:31 pve smartd[768]: Device: /dev/nvme1, NVMe error count increased from 134 to 136
-- CRASH --
Jan 25 19:04: pve kernel: Linux version 6.17.2-1-pve
```

**Questions for TerraMaster:**

1. Is there a BIOS/firmware update newer than MADN0301.V07 for F8 SSD Plus?
2. Are there known issues with Proxmox/kernel 6.17 on F8 SSD Plus?
3. Are there known incompatibilities with Kingston KC3000 NVMe drives on this hardware?
4. Could this be a PSU issue? (Power cuts are happening 503+ times on boot drive)
5. Could this be NVMe controller/chipset issue on motherboard?
6. Is there a way to enable better hardware monitoring (IPMI/BMC)?
7. Should I try different kernel parameters specific to TerraMaster hardware?

**Additional Info:**
- No monitor/keyboard normally attached (headless operation)
- 2 VMs running: TrueNAS (8GB RAM, 2 cores) and Docker (4GB RAM, 3 cores)
- 3 SATA drives passed through to TrueNAS
- System runs cool, no overheating symptoms
- No physical signs of hardware failure

Any help would be greatly appreciated. The 503 unsafe shutdowns on the boot drive are very concerning, and I'm worried about data corruption/damage to drives.

[CursaYang]

Your current BIOS version does not need an update. Please confirm if installing the SSD in the 5/6/7/8 NVMe slots will also result in the same issue? If not,you can refer to this post to update the firmware of the PCIe expansion chip:
viewtopic.php?t=8961
The Kingston KC3000 SSD is compatible.
Regarding whether it could be a power issue, were the status indicators for the hard drive, LAN, and power on the front panel normal at the time the fault occurred?

[alienable990]

I switched the 4 nvme to 5/6/7/8. Started it up. It's been running from 19:41 to around 22:34 (less than 3 hours) but then crashed. It's crashed 3 times in the last 24 hours.

[CursaYang]

I switched the 4 nvme to 5/6/7/8. Started it up. It's been running from 19:41 to around 22:34 (less than 3 hours) but then crashed. It's crashed 3 times in the last 24 hours.

Thank you for your cooperation. The initial assessment suggests that it may be a power supply issue, as you mentioned that the device suddenly lost power without overheating. Is the power adapter functioning normally? If you have a power adapter with a similar wattage, you could try replacing it for testing.
Additionally, you can refer to this post to modify the BIOS, and use a single hard drive to install the system and run it to confirm whether insufficient power is the cause.
If the same issue occurs with a single hard drive, please send your purchase order information along with the link to this post to our technical support email ([email protected]), and we will arrange for a sales representative to contact you.

[alienable990]

The power adapter seems to be functioning normally.

Tried checking the firmware version of the PCIe (see viewtopic.php?t=8961) but got this:

root@pve:~# chmod 755 218xfwdl
root@pve:~# ./218xfwdl /s
./218xfwdl: line 1: syntax error near unexpected token `<'
./218xfwdl: line 1: `<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"><html><head><title>403 Forbidden</title></head><body><h1>403 Forbidden</h1><p>You don't have permission to access the URL on this server.</p><p>denied by UA ACL = not in whitelist</p><hr/>Powered by Tengine <br/>CDN Request Id: a3b5839517697171733918648e</body></html>'

I will try removing my 3 x 4TB drives and just have the 1 x 512GB and see how that works.

[TMjack]

Hello, we recommend that you upgrade directly to TOS 6.0.794. You must use this method and should not update via any other path.
Please execute ./218xfwdl -S

[alienable990]

I'm not sure I understand "upgrade directly to TOS 6.0.794". I'm not using TOS, I'm using proxmox.

[Felix]

Installing third-party operating systems (such as Unraid, TrueNAS, Proxmox, etc.) on TNAS devices may carry risks. For F8 SSD/F8 SSD Plus models, it is known to cause the device to fail to boot or repeatedly restart. This requires upgrading the BIOS and PCIe chip firmware according to the official guide to resolve the issue. Upgrade instructions are available here: viewtopic.php?t=8961. We cannot provide technical support for third-party systems. If you have higher requirements for certain TOS functions, please provide feedback through official channels, and we will continue to improve the system.