So I was moving some files to that disk that I very recently moved data AWAY from that disk.

While that move was taking place (using UNRAID webui), it reported read errors (not write errors?), 1280 of them and then it disabled that disk and shows it emulated.

The good thing is that the data I was copying to the disk (and the few data it already had inside), are easily identifiable and fixable/replaceable.

So my issue is what to actually DO as a process.

Before attempting to replace the disk (which is a bit old but not the first in line to replace yet), I want to make sure if the disk is maybe ok.

NOTE: Do not explain me how to replace the disk. I know that and have done this process many many times. What I need is how to first check if the drive is actually ok and that was a system "fart" (maybe shift some cable or whatever).

I tried to do a full SMART test, but it stops immediately after clicking and I didn't get any message if the test happened (or is still happening). SMART logs and history, show nothing.

Can someone describe the process of what to do next?

Should I stop the array? Should I restart it in maintenance mode? Should I somehow check the disk? ( how?) How do I re-enable the disk. Is it possible that SMART doesn't do anything because the disk is emulated and maybe the filesystem doesn't actually see the physical disk? (something I think I can only verify if I stop the array)

Help?

(read older such threads but there was some conflicting info, so I am making this fresh thread)

Edited October 1, 2025Oct 1 by NLS

Please post the diagnostics.

44 minutes ago, JorgeB said:

Please post the diagnostics.

Here... thanks.

quasar-ultima-diagnostics-20251001-1810.zip

Disk dropped offline. This is most often a power/connection issue, but that's also why SMART doesn't work. Check/replace cables and post new diags after array start.

1 hour ago, NLS said:

it reported read errors (not write errors?)

P.S. a disk is only disabled if there are write errors:

Oct 1 15:44:19 quasar-ultima kernel: md: disk10 write error, sector=4716650520

So I will simply shutdown, as is, check/shift cables etc. and turn back on.

If disk is back, at least it will report as existing - but I am not sure if it will go in the array or what.

Maybe I need to stop array and remove it from array first and then shutdown?

This way if disk is back to life after reboot and I re-add it, it will rebuild it, right?

(will probably do the second thing, not just shutdown - a rebuild will be safer anyway, I don't know what didn't go ok while writing - plus it will be a full surface test for the disk)

BTW the disks are all connected to my 16i controller, with 3 x 1-to-4 cables.

Edited October 1, 2025Oct 1 by NLS

EDIT: Noticed another thing though. The disk shows in unassigned devices! Even now, before rebooting.

So I think it actually just briefly dropped offline.

I will remove from array, shuffle cables and reboot anyway.

EDIT: ...and did it. It is rebuilding on same disk so I guess if the disk has a problem it will hit again will trying to write it.

Edited October 1, 2025Oct 1 by NLS

1 hour ago, NLS said:

Noticed another thing though. The disk shows in unassigned devices! Even now, before rebooting.

Yes, that's normal; it dropped and then reconnected with a different identifier.

18 minutes ago, JorgeB said:

Yes, that's normal; it dropped and then reconnected with a different identifier.

Not sure it had any different identifier, looks the same to me in every way.
I may look at the logs I sent you, because I suspect it lists it there and I can probably verify.

In any case, doesn't matter much, the good thing is that it is still rebuilding with no errors. In a few hours I will know if it was just a hiccup.

17 minutes ago, NLS said:

Not sure it had any different identifier, looks the same to me in every way.

It was sdl

Oct 1 15:44:19 quasar-ultima kernel: sd 1:0:10:0: [sdl] Synchronize Cache(10) failed: Result: hostbyte=0x01 driverbyte=DRIVER_OK

It reconnected as sdn

Oct 1 15:44:58 quasar-ultima kernel: end_device-1:13: add: handle(0x001a), sas_addr(0x300605b0121272a5)

Oct 1 15:44:58 quasar-ultima kernel: sd 1:0:13:0: [sdn] 7814037168 512-byte logical blocks: (4.00 TB/3.64 TiB)

Oct 1 15:44:58 quasar-ultima kernel: sd 1:0:13:0: [sdn] 4096-byte physical blocks

Oct 1 15:44:58 quasar-ultima kernel: sd 1:0:13:0: [sdn] Write Protect is off

Oct 1 15:44:58 quasar-ultima kernel: sd 1:0:13:0: [sdn] Mode Sense: 6b 00 10 08

Oct 1 15:44:58 quasar-ultima kernel: sd 1:0:13:0: [sdn] Write cache: enabled, read cache: enabled, supports DPO and FUA

Oct 1 15:44:58 quasar-ultima kernel: sdn: sdn1

Oct 1 15:44:58 quasar-ultima kernel: sd 1:0:13:0: [sdn] Attached SCSI disk

Oct 1 15:44:59 quasar-ultima unassigned.devices: Disk with ID 'WDC_WD40EFAX-68JH4N1_WD-WX42D410YUKV ()' is not set to auto mount.

Right. After reboot it returned to sdl.
Thanks for the help - actually the hand-holding, as I wanted an expert's advice to "enhance" my own (I do use UNRAID since almost it existed, way before nice GUI and plugins etc.). It was not even the first time this happened over the years, but I didn't remember exactly how I handled it before (i.e. keeping same disk in), so you helped me take more careful steps.

EDIT: Rebuild finished fine, I am actively using the disk again now.

Edited October 2, 2025Oct 2 by NLS