I have three HDDs in a DAS attached to my Unraid server via USB. Not ideal I know but it's what I have to work with.

The array was setup like this:

• Parity: 8TB HDD1

• Data: 8TB HDD2

• Data: 4TB HDD

• Pool: 1TB SSD

Last week the 8TB data drive failed completely. I removed it and shutdown the server. Today the replacement drive arrived and I installed it in the DAS then booted the server. As a side effect of using the DAS Unraid will sometimes get upset that drives are not in the right slots even though they are. So when the Parity drive had a red X beside it and the server complained of too many errors I did what I've done before and hit new Config, assigned the disks to their proper slots, checked Parity is Valid, and started the array.

Here is where I'm pretty sure I majorly messed up.

The new 8TB data drive appeared as unformatted instead of emulated like I expected. In a rush to start rebuilding I formatted the drive. I read the warning but somehow didn't process that it was going to apparently wipe parity?

So, I know I'm an idiot but have I lost all the data on the array now?

There is a way you could have made the new disk emulated, and we could have told you how if you had asked before doing anything. But now parity agrees it has been formatted

Are you sure the original disk is unreadable?

Parity doesn't contain any of your data of course. It just allows the data from a missing disk to be calculated from all the other disks. And format doesn't completely wipe out every trace of the original disks data from the array. It just writes an empty filesystem to the disk, and parity agrees the disk has an empty filesystem after that.

But maybe if we make it emulated and rebuild it from parity, even though the result would be an empty filesystem, all the rest of the bits from the original data would still be there and UFS Explorer might be able to recover some of it.

The original disk is gone unfortunately. I was only able to get a replacement by returning it.

I'm up for trying anything at this point if you're willing to walk me through it. The array is currently stopped.

Have you used the array at all since you formatted?

No, once I realized my mistake I stopped the array but left the server running.

Unassign the formatted disk. Start the array without it. Post diagnostics.

Just now, trurl said:

Start the array without it

In normal (not maintenance) mode

unraid-diagnostics-20260502-1656.zip
Diagnostics I pulled just after my mistake in case you want those. Will pull again once array starts

Those show disk1 empty as expected,.

But disk2 also appears empty. Is that expected also?

Array started. New diagnostics attached.unraid-diagnostics-20260503-1455.zip

I don't believe the server every put anything on disk 2

Edit: Due to using High-water allocation I believe.

Edited May 3May 3 by jparrott3
Added info

Reassign disk1, start the array to begin rebuilding disk1. The result will still be an empty disk, but it will be as if you accidentally formatted the disk after you rebuilt it with its data. Then you can try something to help recover deleted data from a disk, such as UFS Explorer.

Do you have backups? You must always have another copy of anything important and irreplaceable. Parity is not a substitute for backup, whether Unraid or any other system. Plenty of more common ways to lose data besides a failed disk, including user error.

The array is almost entirely media ripped for Plex use so I have backups in the sense that I still have the DVD/Blu-Rays in storage. Would really like to not have to rip all of them again but nothing on here is irreplaceable.

Array restarted and Data-Rebuild in progress. Time to complete started at 12 hours but now says about 9.

With something like UFS Explorer I assume I would need at least 1 8TB hdd to recover to, is that correct?

Also, if I find myself in a similar situation in the future where the array thinks the Parity drive is wrong but I also am trying to replace a failed disk, what steps should I have taken?

Not sure if related to what we're trying to do or if my 4TB drive (disk 2) is having trouble now, but the 4TB drive is showing a few million errors and rising.

Any time you are unsure what to do, you should ask on the forum and post your diagnostics. Since you are using USB, I wouldn't be surprised if this doesn't happen any time you need to replace a disk.

Actually getting Unraid to rebuild a disk after New Config requires several steps that have to be followed exactly, with some checks along the way to decide what needs to be done next.

I've never needed to use UFS Explorer. But it isn't going to be like rebuilding a disk. It is going to scan the disk and try to find files that it can recover individually. The trial edition will let you see what files it can recover but you will have to pay to actually recover anything. Standard edition will be fine for recovering an Unraid array disk since it is just a single Linux filesystem.

Just now, jparrott3 said:

Not sure if related to what we're trying to do or if my 4TB drive (disk 2) is having trouble now, but the 4TB drive is showing a few million errors and rising.

You can't rebuild disk1 unless ALL other array disks are working well.

Post new diagnostics.

The real solution to your problems is a build with individual SAS/SATA connections for each assigned disk.

New diagnostics

unraid-diagnostics-20260503-1531.zip

2 minutes ago, trurl said:

The real solution to your problems is a build with individual SAS/SATA connections for each assigned disk.

Yeah, unfortunately for now I'm stuck with a Minisforum mini PC and Mediasonic Probox. The Minisforum is great little machine but not really suited to this.

Might as well stop the rebuild. Maybe try reseating all drives, checking all connections. You will have to start the rebuild over.

Did you every successfully build parity with this hardware?

Even though disk2 is empty, all sectors of disk2 (and all sectors of parity) must be reliably read to reliably rebuild disk1.

Given the unreliable nature of drive identification with my setup would a disk spin down and hot re-seat be safer than a shutdown?

About two weeks ago, before the old drive failed, it had successfully built parity and was working fine until the drive failure.

Just now, jparrott3 said:

would a disk spin down and hot re-seat be safer than a shutdown?

I would expect it to be less safe, except for the fact that nothing is working anyway.

And whether you shutdown or not, I wouldn't be surprised if it was back to this

22 hours ago, jparrott3 said:

Unraid will sometimes get upset that drives are not in the right slots even though they are.

DId you always have a disk2 when you set this up? Or was it added later?

Original setup:
Parity - 8TB Barracuda SMR

Disk 1 - 4TB Ironwolf Pro

I swapped the Barracuda out for an 8TB Exos and once that process was done added a second 8TB Exos data drive so the setup was:
8TB Exos (Parity)

8TB Exos (Disk 1)

4TB Ironwolf Pro (Disk 2)

I remember somehow moving the data from Disk 2 to Disk 1 but don't remember why or how I did it now. Until Disk 1 failed I had multiple successful parity checks.