4 TiB WD SN700 drives drop off of Sonnet Card

[Jani Partanen]

Hi, if you are running them with 4K LBA format, switch back to 512k. 4K 
LBA cause issues to many WD drives. WD also know this, its been reported 
years ago and it never got firmware fix so its quite safe to say that 
they have harware flaw what cause it and cannot be fixed by firmware.

If you are running your drives with 512k LBA, then this is some other issue.


// Jani

On 01/03/2026 13.39, Justin Piszcz wrote:
> Hello,
>
> Kernel: 6.12.73
> Distribution: Debian Stable Trixie
>
> There are long outstanding issues with these WD Red SN700 4TiB drives
> with an ASUS FlahsStor Gen1.
>
> I have 16 drives on two x8 Sonnet boards and the same issue is
> occurring across two different platforms.  Are these drives completely
> defective when used in a software raid configuration?
>
> Does anyone have a specific kernel parameter to make these drives work
> without dropping off the bus— or is it time to retire/replace these
> drives?  On various forums the below parameters had been tested
> previously and did not help.
>
> [Sun Mar 1 00:30:58 2026] nvme nvme0: controller is down; will reset:
> CSTS=0xffffffff, PCI_STATUS=0xffff
> [Sun Mar 1 00:30:58 2026] nvme nvme0: Does your device have a faulty
> power saving mode enabled?
> [Sun Mar 1 00:30:58 2026] nvme nvme0: Try
> "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off"
> and report a bug
> [Sun Mar 1 00:31:03 2026] nvme 0000:03:00.0: Unable to change power
> state from D3cold to D0, device inaccessible
> [Sun Mar 1 00:31:03 2026] nvme nvme0: Disabling device after reset failure: -19
> [Sun Mar 1 00:31:03 2026] md/raid:md0: Disk failure on nvme0n1p1,
> disabling device.
>
> Justin
>