Help with SDD cache and TRIM failure. "FITRIM ioctl failed: No data available"

[in_trauma]

I searched around a bit and couldn't find this same problem discussed. 

For past 5-6 weeks I get this email message after TRIM is set to run. (2AM every sunday)

SUBJECT:  

cron for user root /sbin/fstrim -a -v | logger &> /dev/null

BODY:

fstrim: /mnt/cache: FITRIM ioctl failed: No data available

 

I have Dynamix SSD TRIM installed and set to run every Sunday morning. My cache pool is 1x Samsung 830 256GB (2.5" connected to motherboard SATA), 1x Kingston A2000 M.2 250GB (pcie NVME connected to a PCIE adapter card) The Kingston was add a few months ago and shortly after I started getting the email message about TRIM. My base system is an HP Proliant Microserver Gen 10 Im pretty maxed out on what I can put in this system The PSU doesn't offer me the ability to power an PCIE SATA card.

I've had my system running unraid for 8yr now and done simple drive swaps, upgrades to hardware and updates to dockers, plug-ins and such but my knowledge is still quite lacking on the more complicated stuff. I added the Kingston in July fearing a failure of the Samsung and relying more and more on my Cache drive as I use the server now more than ever. 

 

My logs have MANY BTRFS errors that I don't fully understand. (see screenshot)

 

The SMART health for the Kingston drive currently states FAILED and when I attempt to run a SMART test (short or extended) nothing happens. It also cant read it's capabilities

 

The Samsung drive notes a CRC error count issue. It's said this for a long time, I tried swapping SATA cable a few times as well as I read this could solve it.

 

Is there a simple fix for this? Do I just have something configured wrong? Is my PCIE adapter just crap?
 

Edited October 4, 2020 by in_trauma

[JorgeB]

14 hours ago, in_trauma said:

when I attempt to run a SMART test (short or extended) nothing happens.

NVMe devices don't have SMART tests.

 

Please post the diagnostics: Tools -> Diagnostics

[in_trauma]

8 hours ago, JorgeB said:

Please post the diagnostics: Tools -> Diagnostics

Here they are.

tower-diagnostics-20201005-0946.zip

[JorgeB]

Syslog is spammed with errors and can't see the beginning of the problem, please reboot and post new diags once the errors begin.

[in_trauma]

Rebooted. BTRFS errors immediately like before pointing to nvme storage.

tower-diagnostics-20201007-0758.zip

[JorgeB]

Looks like a hardware problem, possibly there's some issue with the NVMe device or it's a compatibility issue, if you can try a different NVMe device or try that device in another computer/OS.

[in_trauma]

Im betting it's my nvme adapter. It's quite cheap. My only other option that I have available at the moment would be to use a usb to nvme enclosure until I buy a better pci-e to nvme adapter.  Would you be able reccomend one?

[JorgeB]

Don't recommend USB for array/pool devices, they tend to be very bad at error handling and drop devices frequently.