Hi, really hope some can help.

 

I've recently built a new Unraid server and have configured 2 x 2TB Crucial P3 NVME drives as a mirrored cache using ZFS. Unfortunately the drives seem to drop constantly (as seems to be common NVME drive issue searching on the forums).

 

I have tried appending both "nvme_core.default_ps_max_latency_us=0" and "pcie_aspm=off" to the Syslinux config but this has not made any difference.  I have also tried connecting the drives using PCIe to NVME adapters which also hasn't fixed the issue, as well as tested the memory to rule that out.  Bios is up to date and there are no availble Firmware Updates I can see for the Crucial P3 drives.

 

A common suggestion is to change the NVME drives which I would rather not do, but if it's the only solution then I guess I'll have to. Unfortunately It's not clear which drives would work, are there any recommended NVME drives that people have had success with for this particular motherboard? 

 

I assume this is a common Linux issue and not unique to Unraid?

 

My build is as follows:

 

Gigabyte c246m-wu4

Intel i3-9100

8GB DDR4 Non ECC

2 x 2TB Crucial P3 NVME in ZFS Mirror

3 x 16TB Toshiba MG08 in xfs Single Parity Array 

 

Below is an example of what I will see in Syslog:

 

Oct 31 06:10:48 NAS kernel: nvme nvme1: I/O 430 (I/O Cmd) QID 4 timeout, aborting
Oct 31 06:11:19 NAS kernel: nvme nvme1: I/O 430 QID 4 timeout, reset controller
Oct 31 06:11:32 NAS kernel: nvme nvme1: I/O 5 QID 0 timeout, reset controller
Oct 31 06:14:30 NAS kernel: nvme nvme1: Device not ready; aborting reset, CSTS=0x1
Oct 31 06:14:30 NAS kernel: nvme nvme1: Abort status: 0x371
Oct 31 06:16:38 NAS kernel: nvme nvme1: Device not ready; aborting reset, CSTS=0x1
Oct 31 06:16:38 NAS kernel: nvme nvme1: Removing after probe failure status: -19
Oct 31 06:18:46 NAS kernel: nvme nvme1: Device not ready; aborting reset, CSTS=0x1
Oct 31 06:18:46 NAS kernel: nvme1n1: detected capacity change from 3907029168 to 0
Oct 31 06:18:46 NAS kernel: zio pool=nvme vdev=/dev/nvme1n1p1 error=5 type=2 offset=567128489984 size=4096 flags=180880
Oct 31 06:18:46 NAS kernel: zio pool=nvme vdev=/dev/nvme1n1p1 error=5 type=2 offset=1736266567680 size=4096 flags=180880
Oct 31 06:18:46 NAS kernel: zio pool=nvme vdev=/dev/nvme1n1p1 error=5 type=2 offset=1685769904128 size=131072 flags=180880
Oct 31 06:18:46 NAS kernel: zio pool=nvme vdev=/dev/nvme1n1p1 error=5 type=1 offset=1640755515392 size=131072 flags=180880
Oct 31 06:18:46 NAS kernel: zio pool=nvme vdev=/dev/nvme1n1p1 error=5 type=5 offset=0 size=0 flags=100480
Oct 31 06:18:46 NAS kernel: zio pool=nvme vdev=/dev/nvme1n1p1 error=5 type=5 offset=0 size=0 flags=100480
Oct 31 06:18:46 NAS kernel: zio pool=nvme vdev=/dev/nvme1n1p1 error=5 type=5 offset=0 size=0 flags=100480
Oct 31 06:40:06 NAS kernel: nvme nvme0: controller is down; will reset: CSTS=0xffffffff, PCI_STATUS=0x10
Oct 31 06:40:06 NAS kernel: nvme nvme0: Does your device have a faulty power saving mode enabled?
Oct 31 06:40:06 NAS kernel: nvme nvme0: Try "nvme_core.default_ps_max_latency_us=0 pcie_aspm=off" and report a bug
Oct 31 06:40:06 NAS kernel: nvme 0000:01:00.0: enabling device (0000 -> 0002)
Oct 31 06:40:06 NAS kernel: nvme nvme0: Removing after probe failure status: -19
Oct 31 06:40:06 NAS kernel: nvme0n1: detected capacity change from 3907029168 to 0
Oct 31 06:40:06 NAS kernel: zio pool=nvme vdev=/dev/nvme0n1p1 error=5 type=2 offset=567281975296 size=69632 flags=180880
Oct 31 06:40:06 NAS kernel: WARNING: Pool 'nvme' has encountered an uncorrectable I/O failure and has been suspended.

 

After this, the system either becomes unresponsive or the CPU's are in full use, with a hard power off the only solution, resulting in a very long 2 day parity check for my array.

 

Attached are my diagnostic files.

 

Thanks in advance for any assistance.

nas-diagnostics-20231031-1435.zip

After digging around in the Bios a bit more, I noticed that if I disable "CEC 2019 Ready" in Power Management, I can disable the various ASPM options that were grayed out under "Power Platform Management". I've been up for over 24 hours now, which is longer than it has been running for a while. I would however like to persevere with getting ASPM working in some form, as I want to be running as efficiently as possible. If there are any good guides that anyone can point me in or has any recommendations it would be appreciated.

 

Cheers

You found a solution?