Over the past week I've gotten hundreds of warnings in Unraid saying my Write Cache NVME has been disconnected. A reboot will bring it back up temporarily, but anywhere from a few minutes to a day or two later it's back down again. 

 

I thought it might have been the NVME slot it's in, so I swapped it from slot 2 to 3 with no luck. 

 

Secondly, I replaced the SSD (which is a less-than-year-old Sabrent Rocket 4.0 2TB) with the old CS3030 that it replaced, which I know to be good. Both show up fine in a Windows machine, and the SMART data on both are fine. 

 

Now that drive dropped out today. Going to the attributes tab shows it can't be read, but I don't see any warnings in Unraid yet. Looking at the logs though, there's a ton.

 

I really hope it's not my motherboard, but that's what I'm leaning toward at this point...

 

nanolith-diagnostics-20230405-2100.zip

Edited April 6, 20233 yr by RandomPedestrian

This can sometimes help:

 

On the main GUI page click on the flash drive, scroll down to "Syslinux Configuration", make sure it's set to "menu view" (top right) and add this to your default boot option, after "append initrd=/bzroot"

nvme_core.default_ps_max_latency_us=0 pcie_aspm=off

e.g.:

append initrd=/bzroot nvme_core.default_ps_max_latency_us=0 pcie_aspm=off

Reboot and see if it makes a difference.

No luck unfortunately. I updated the BIOS (which was admittedly two years outdated), found that one hard drive was dying and downgrading the link to 1.5Gb/s and replaced it, but after a week of it being fine I’m getting the error again. 

Best bet then would possibly be to use a different model NVMe device, or a different board.

For what it's worth I had the same issue with my motherboard (Aliexpress BKHD-1264 with Intel N100). This MB has 2 M.2 slots, and the NVMe on M.2 slot0 would disconnect and reconnect periodically. M.2 slot 1 had zero issue. I tried initrd=/bzroot nvme_core.default_ps_max_latency_us=0 pcie_aspm=off and no luck.

 

The only difference between both M.2 slots is slot 0 is also connected to the Intel SATA controller (slot1 is only NVMe). Hotplug was enabled as well so I suspect the SATA controller was messing up with the NVMe SSD

 

So I disabled SATA port connected to the M.2 slot0 on the BIOS, and so far so good.

Well, no luck, after a while it disconnected again.

 

I disabled the ACPI in BIOS, it's been stable 1 day.... (versus disconnects every hour).

On 4/6/2023 at 3:05 PM, JorgeB said:

This can sometimes help:

 

On the main GUI page click on the flash drive, scroll down to "Syslinux Configuration", make sure it's set to "menu view" (top right) and add this to your default boot option, after "append initrd=/bzroot"

nvme_core.default_ps_max_latency_us=0 pcie_aspm=off

e.g.:

append initrd=/bzroot nvme_core.default_ps_max_latency_us=0 pcie_aspm=off

Reboot and see if it makes a difference.

This worked for me; I have two of the below drives BTRFS Raid 1, the second drive kept going offline; a reboot would fix it, but it kept dropping offline after a day. It has been a few days now, which looks promising. 

Samsung SSD 990 PRO with Heatsink 2TB

On 1/20/2025 at 12:30 PM, presidential-visor5807 said:

This worked for me; I have two of the below drives BTRFS Raid 1, the second drive kept going offline; a reboot would fix it, but it kept dropping offline after a day. It has been a few days now, which looks promising. 

Samsung SSD 990 PRO with Heatsink 2TB

Did it solve your problem over longer time? I have the Samsung SSD 990 Pro too but still have randomly dropouts (sometimes after 6 days of running). Cannot find the fix for it. 

I'm having the same issue on unraid 7.1.4. My 990 PRO 2TB disconnects after a few days. A reboot usually fixes it but it only lasts a few days.

Adding the following to Syslinux configuration fixed my issue with the SSD disconnecting every couple of days.

label Unraid OS

menu default

kernel /bzimage

append initrd=/bzroot nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off

Edited July 14, 2025Jul 14 by PELLETT655

i got two 990 pro 2TB.

One of them fluked out the other day.. that drive is going back, I am not playing around with my cache pool and data.

replacing them with a wd sn850x, with them, i had no issues so far.

The NAND is fine though, I ran badblock for 48 hours straight without any issue.