Aoostar WTR MAX NVMe Slot Issue

[Enquiries]

Hi Robbie,

Thanks for all the work you do—your reviews have been a great help.

I’m seeing a repeatable issue on the Aoostar WTR MAX with two Samsung 980 PRO NVMe drives in a mirrored ZFS pair. When both drives are in the ×2 tray slots, one often drops and is marked REMOVED, leaving the mirror degraded. Moving a drive to the furthest ×1 slot seems to help, but this isn’t confirmed.

I’ve tried the Linux/Unraid PCIe power tweaks (nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off), but the issue remains. A friend with the same hardware sees the same behaviour. Aoostar say they will ship an updated M.2 bracket in November, hinting at a design fault, and also sent me an untested BIOS, which I avoided as they could not confirm its details—only that it targeted NVMe issues.

Since you reviewed this unit, could you check if this behaviour can be reproduced? I can share logs and steps if useful.

Your independent insight would be very helpful.

Kind regards,
Ben

[ed]

Yes, this behaviour can be reproduced based on what others report. It seems Aoostar is working on a fix, and in many cases the fix (BIOS + bracket) mitigates or resolves the issue. But until you have the improved hardware and verified stability under your exact workload, there is a risk of seeing drive drops.


What I suggest you do

Here are steps you can take now and short-term / long-term fixes:
• Apply the updated BIOS from Aoostar if you haven’t already. It may resolve many of the issues.
• Monitor detailed logs of when the drop happens. Note which slot the failing NVMe is in, and whether temperatures / load are high.
• Request the unmodified / improved M.2 bracket from Aoostar once it becomes available (they have committed to shipping it to affected customers).
• As a test, try configurations with one drive in a ×2 slot and one in a ×1 (if possible) to see if stability improves.