I'm running unraid on an HP ML10v2, using a a 1TB SSD as a cache drive.  The other day, i started seeing errors, and someone on this forum suggested i look at the cabling on the SSD.  I was able to get unraid to see the cache drive again, but soon I got errors about it again.  I've since removed the cache drive physically.  I installed a NVMe2 500 GB SSD to use for a new cache drive.  I reboot unraid and see the NVMe drive in the unassigned devices, but the format button is grayed out.  I tried to stop the array to format, but I'm getting a message about trying again after removing disk share(s)...  I can't seem to find the issue.  I've attached the logs, hope they help and thanks in advance for the help!

tower-diagnostics-20200911-0924.zip

Syslog is spammed with ACPI errors, see here for the fix, then reboot and post new diags.

Here is the new one, hopefully the spam is gonetower-diagnostics-20200911-1155.zip

I just tried to stop the array for fun, and it worked now that I added that line in /config/go

 

Not sure what happened?

If it happens again post new diags, that fix should be unrelated to array stopping issues.

I was able to stop the array, and assign the NVMe drive to cache, and see that the format button is available, but it doesn't actually format anything and I still have "No Pool uuid"  for the drive.  Mover is running, but not sure why.   I didn't do anything in unraid when I removed the 1 TB cache drive, and just added this one in the same slot in the configuration.  Could that be causing problems?  Thanks again for the assistance. 

I don't see any format attempt on the diags posted.

This is from the disk log information:

 

     Sep 11 12:17:35 Tower kernel: print_req_error: I/O error, dev nvme0n1, sector 0
     Sep 11 12:17:35 Tower kernel: Buffer I/O error on dev nvme0n1, logical block 0, lost async page write
     Sep 11 12:17:35 Tower emhttpd: error: mkmbr, 1967: Input/output error (5): write: /dev/nvme0n1

 

Does this shed any light?

Please post the full diags.

tried to stop the arratower-diagnostics-20200911-1332.zipy again, having same issue again with not being able to stop array.  Here are the new diagnostics

Zip is empty.

this is all I get when I click the icon next to my cache drive:

 

Sep 11 14:44:02 Tower emhttpd: WDS500G3X0C-00SJG0_203123802818 (nvme0n1) 512 976773168
Sep 11 14:44:02 Tower emhttpd: import 30 cache device: (nvme0n1) WDS500G3X0C-00SJG0_203123802818
Sep 11 14:47:16 Tower emhttpd: writing MBR on disk (nvme0n1) with partition 1 offset 64, erased: 0
Sep 11 14:47:16 Tower emhttpd: error: mkmbr, 1967: Input/output error (5): write: /dev/nvme0n1
Sep 11 14:47:16 Tower kernel: print_req_error: I/O error, dev nvme0n1, sector 0
Sep 11 14:47:16 Tower kernel: Buffer I/O error on dev nvme0n1, logical block 0, lost async page write

 

I've also attached the diag, after attemepting to format the NVMe drive again after a reboot.  I'm still having issues stopping the array again, though.

tower-diagnostics-20200911-1448.zip

See if you can wipe it clean with:

blkdiscard /dev/nvme0n1

If command is successful try to format after that.

On 9/12/2020 at 2:48 AM, JorgeB said:

See if you can wipe it clean with:

blkdiscard /dev/nvme0n1

If command is successful try to format after that.

Thank you for this.  I am at work, and took the drive out of my system to see if i could format in my windows machine, but I was getting errors about being misnamed or mismatched?  I might be wrong about the error, as it was Friday evening, but when I looked at the device settings in windows, it said device was working properly, but the log showed something about not being matched or named.  When I get home tonight, I will reinstall the drive and try this.   Am I correct in that I just open the command prompt from the GUI and type that command?

34 minutes ago, PaDadof2 said:

Am I correct in that I just open the command prompt from the GUI and type that command?

Yes, but if you were also getting errors with Windows it's likely a device problem.

On 9/14/2020 at 9:56 AM, JorgeB said:

Yes, but if you were also getting errors with Windows it's likely a device problem.

I was finally able to reinstall the drive, and used the command above, I'm not seeing that it did anything though.  I stopped the array, and I can see the drive in the unassigned devices, but the format button is still grayed out.  Guessing this drive needs to be replaced.  Smart report and new diagnostic are attached, not sure if they say anything new?

tower-smart-20200917-1042.zip tower-diagnostics-20200917-1046.zip

Yeah, probably a device problem, you can try formatting assigned as cache but would except same error.

I was able to assign it as a cache drive, and see the format button, but again, it does nothing.  Will be returning it.  Thanks for helping out.

any reason why I shouldn't use this Portable SSD for a cache drive?  I was able to pick up three of them brand new for 15 each.  Using one for my android tv box, still have two to use.

18 minutes ago, PaDadof2 said:

any reason why I shouldn't use this Portable SSD for a cache drive? 

It's USB, so not good for Unraid.

i figured as much