So I think I really messed up, need some feedback on my options.

I am currently trying to upgrade a 4tb drive with an 8TB. I swapped out the drives and the server was rebuilding, but i noticed very slow rebuild time. It looks like another drive i have (also 8TB) is failing and giving me lots of errors when trying to rebuild. I have tried re-connecting the sata cables and rebooting several times. The rebuild will start fast at 100MBs but then creep to a halt 37kb (1300 day rebuild time) once the errors start to pick up on the other drive.

SO i tried to put the 4tb drive back in but it wont recognize it (maybe a few files were added during my rebuild). So now it looks like i'm stuck with a failing drive and i can't rebuild it (i only have 1 parity drive).

Any ideas? TIA

For a better idea of the current status, post the diagnostics, ideally after there were errors during the rebuild.

Thanks @JorgeB Attaching Diagnostics

mattos-diagnostics-20251211-1330.zip

So currently i have replaced the original 4tb drive back in the rig, any ideas how to get it back on the array as it originally was, so i can replace the failing drive?

WDC_WD80EFAX-68KNBN0_VDJ7YRKK (sdi) is the drive that is giving errors in the rebuild

So here is what im thinking, if someone can sanity check this for me. Currently ill leave the 4tb drive as emulated, ill add the new 8TB drive as a new drive, then i can clone the 8TB drive that is failing? Is this possible. I have 1 extra slot (i have 7 drives currently) Then once the failing drive is cloned and replaced i can get a new drive to rebuild the emulated 4tb.

Edited December 12, 2025Dec 12 by MattOS

Disk3 errors are logged as a disk problem, but SMART looks mostly OK, run an extended SMART test on that disk and post new diagnostics.

Its been running about 24 hours, stuck on 10%

mattos-diagnostics-20251213-0947.zipmattos-smart-20251213-0946.zip

Edited December 13, 2025Dec 13 by MattOS

Sata cable replacements arrived today, going to shut her down and change out the cables to see if that helps.

OK i have a 2nd drive arriving today, i think the best path is to add a 2nd parity drive, then i can emulate both the drive i was trying to replace and the failing drive, then i can rebuild with new drives. would you agree this is the best path for now? once i rebuild on a new drive i can do a preclear/diagnostics on the failing drive and see if i can get it back in service. My biggest concern right now is just to get the array back in good shape. @JorgeB

quick smart test on failing drive after replacing sata cable

mattos-smart-20251213-1444.zip

well, new sata and different sata port did not help. rebuild still not running as expected with errors on the drive. Thinking about it, i don't think ill be able to add a 2nd parity drive either if it can't read the drive, i guess my only option is to try to clone the drive with ddrescue unless there is an alternative.

With one invalid and another bad drive, the options are limited.

Do you still have the old disk4 intact? If so, and assuming nothing else has been written to the array, you may try to use that to rebuild disk3, but keep old disk3 intact for now; that should be mostly salvageable using ddrescue if needed.

@JorgeB Yes old disk 4 is still intact, but i guess i had new data on the array to disk 4 since it now exceeds 4tb (4.01) really annoying unraid will write new data to an emulated disk. That was my first plan of action was to get the 4tb disk 4 back so i could rebuild disk 3 but that ship sailed. I have an 8th disk connected now, so ill try to use unbalanced to move the data off the bad disk. Or do you think i should just go straight to ddrescue on disk 3? OR should i try to use unbalanced to move the data off disk 4, then i can use parity to rebuild disk 3, is that possible?

Edited December 14, 2025Dec 14 by MattOS

im thinking this is my best option:

Copy emulated Disk 4 data manually: While array runs (Disk 4 emulated), use Unbalance plugin or rsync to move files from Disk 4/share to other disks. Then New Config to unassign Disk 4 → shrink array temporarily → add 8TB later.

Here is what i don't understand, if disk 4 is now emulated, that means its generated by disk 3 (which is giving lots of errors) then my only option is to try to use ddrescue to recreate disk 3 on a new drive that doesn't have errors. because disk 4 has data written to it after it was emulated i can't just put my old disk 4 back in the array and preserve parity, and i cant create a new parity because disk 3 will throw errors. I feel stuck here an don't want to do more damage than i already did. If i try to use UNbalance to move data off the emulated drive, im assuming the errors on disk 3 will prevent that from working. But just maybe i can run ddrescue on disk 3 hoping the errors are disk related and not data related, but im not sure never been here before.

Edited December 15, 2025Dec 15 by MattOS

@JorgeB OK here is my last question, if i can't get dd rescue to run because disk 3 is too corrupted, what can i do here to move forward with the least amount of damage, assuming disk 3 is not recoverable and im out of parity with disk 4, (but i have disk 4 intact except for the additional data written to the array)? any options here to force disk 4 back on the array and just lose the new data? Or is my whole array just toast now?? Can I just move forward with 2 failed disk and purge the corrupt data and keep the rest of the array?

Thanks!

Edited December 15, 2025Dec 15 by MattOS

If disk3 is so bad that ddrescue can't recover most of it, I'm afraid you won't have many good options. You can try and copy the new data that was written to disk 4 somewhere else, then use the old disk 4 and whatever can be salvaged from the disk3 clone to rebuild the array.

OK im going to just have to lose the data on disk 3. Whats the best way to move forward, go in and delete the files directly off the drive to zero it out then remove from the array with a new config?

Update, i just had to kill disk 3 and 4 and reconfigure the array, lesson learned.