I had a 12 TB Parity drive start to show some errors. It was still under warranty so I powered down the array and swapped in a replacement 12 TB drive to start a parity sync. Now that the parity sync is in-progress, I'm seeing a massive amount of read errors on Disk 3 in the array (over 400k errors).

I'm not sure what the best case scenario is at this point. I have the old parity drive still (have not erased data on it).

Should I allow the parity sync to complete? Should I stop the sync, replace the old parity drive, then replace the failing Disk 3?

For the most part, Unraid is still working fine except the syslog is getting hammered with lines like this:

md: disk3 read error, sector=4472505808

Edited July 16, 2025Jul 16 by ColonelRhodes

Please post the diagnostics.

Just now, JorgeB said:

Please post the diagnostics.

tower-diagnostics-20250716-0836.zip

Syslog has already rotated and SMART for disk3 is not present, check/repalce cables for that disk and post new diags after array start.

@JorgeB Ok, I opened the case back up and made sure all of the SATA connections are snug and fit. Booted UnRaid back up and see that Disk3 is showing "Unmountable: wrong or no file system". Attaching diagnostics here. All of my disks are formatted for btrfs if that matters.

tower-diagnostics-20250717-1033.zip

Also if it helps, here is the output of wipefs for that device:

# wipefs /dev/sde
DEVICE OFFSET TYPE UUID LABEL
sde 0x200 gpt
sde 0x3a3817d5e00 gpt
sde 0x1fe PMBR

Edited July 17, 2025Jul 17 by ColonelRhodes

Emulated disk is failing to mount due to filesystem corruption:

Jul 17 10:32:07 tower kernel: BTRFS error (device md3p1): bad tree block start, mirror 1 want 4025951944704 have 4111999304343735503

Recomend seeing if the actual disk mounsta with UD, you will have to unassigned it first.

1 minute ago, JorgeB said:

Emulated disk is failing to mount due to filesystem corruption:

Jul 17 10:32:07 tower kernel: BTRFS error (device md3p1): bad tree block start, mirror 1 want 4025951944704 have 4111999304343735503

Recomend seeing if the actual disk mounsta with UD, you will have to unassigned it first.

Looks like it does mount within UD.

Seems like all the data is still there. Should I try to mount inside the array again?

30 minutes ago, ColonelRhodes said:

Should I try to mount inside the array again?

Nope, the disk is being emulated, and if you rebuild on top, it will overwrite the actual disk.

Assuming the disk is OK, and SMART looks mostly fine, you can do a new config instead: Tools - New Config - Keep all - Apply

Reassign disk3, check parity is already valid, and start the array. If all looks good, run a correcting parity check.

10 minutes ago, JorgeB said:

Nope, the disk is being emulated, and if you rebuild on top, it will overwrite the actual disk.

Assuming the disk is OK, and SMART looks mostly fine, you can do a new config instead: Tools - New Config - Keep all - Apply

Reassign disk3, check parity is already valid, and start the array. If all looks good, run a correcting parity check.

I just did all of this, but before starting the array I see this message next to the parity disk:

"All existing data on this device will be OVERWRITTEN when the array is started"

Should I still continue? This is a new parity drive since originally I was swapping this one out for a new disk.

Edited July 17, 2025Jul 17 by ColonelRhodes

29 minutes ago, ColonelRhodes said:

Should I still continue?

Yes, but make sure that "parity is already valid" is checked. The warning will still be there, but it won't resync parity. Then run a correcting check.

Thank you for all the help! Seems like I'm back up and running. Doing the correcting parity check now!

I just got an alert during parity check that disk7 now has errors (seems similar to what happened on disk3)

tower-diagnostics-20250717-1421.zip

Edited July 17, 2025Jul 17 by ColonelRhodes

Looks like a power/connection issue.