I just set up my NAS a week or so ago, moving from Synology and rebooted it today, only to find that the array won't start back up again

Admittedly the drives are very old, but aren't showing any SMART errors

My Parity drive shows in Unassigned Devices below

Thankfully I have backups of all the important stuff (first thing I set up thankfully) but wondering (hoping) if there's any way I might be able to recover from this?

The disk logs are essentially showing no partitions for all drives

tower-diagnostics-20251214-1854.zip

I mounted one of the drives to my computer and it does indeed show no partitions :(
My cache pool still seems to be okay, so I'm seriously wondering if these drives have finally given out (I think I bought them at the same time...), however as I mentioned before, they haven't shown any SMART errors or warnings (other than one of them was a little hotter than the rest)

EDIT: just to clarify, I'm not asking for a miracle, I can rebuild the array if it's utterly unrecoverable - not the end of the world, but wondering if there is a way to recover this...

Edited December 14, 2025Dec 14 by conjuror3726
clarified title a little

Any chance you have the diags pre reboot? Also, were the devices formatted outside the array, like using the UD plugin? Please try gibing as many details as possible on how the disks were added.

The record size is completely different from what it should be, e.g., for disk2

[size] => 4241706

[sizeSb] => 7814026532

I don't unfortunately, as all was looking fine pre-reboot

I initially formatted the drives using UD plugin to try and clear them out in advance, but had some weirdness with partition sizes which caused the drives to look half as big as they actually were when adding to the array (showing 4TB when the drives are 8TB)

I resolved this by formatting each drive by setting it as the parity before properly setting up the array, then properly initializing the array - then formatting the drives as part of array initialization. That gave me the full 8TB with GPT: 4KiB-aligned on each

However... looking at the partition size of one of those drives, it does look the same as before, which matches what you're seeing there

I wonder if the partition tables didn't properly reset from Synology

Edited December 15, 2025Dec 15 by conjuror3726

Actually just thinking back, I did attempt to format them with my Startech HDD eraser after the 50% size issue, but it showed a warning 'No Device Exist' and wouldn't continue (so this did nothing to the drives) - didn't think anything of it, but maybe that should have been some bad omens from the start

Edited December 15, 2025Dec 15 by conjuror3726

You can try recreating the standard Unraid partition layout. Try with disk2 first, for example, which is showing no partitions:

sgdisk -o -a 8 -n 1:32K:0 /dev/sde

Confirm it's still sde, identifiers can change with a reboot, then reboot and see if disk2 appears correct.

Amazing! that seems to have recovered the drive back to normal thank you!!

Just before I go all gung ho and do the same to the other drives, is that the right approach?

Even with the Parity drive which is showing as Missing instead of just 'Wrong'?

Diagnostics after doing that:

tower-diagnostics-20251215-1106.zip

Disk 1 looks to have partitions in GPT: 4KiB-aligned, however Disk 3 (and the Parity Drive I think) has no partitions like with Disk 2

Edited December 15, 2025Dec 15 by conjuror3726

I just tried mounting the drives with valid partitions via Unassigned Devices and can see all the data there and the correct partition capacity, however when I moved the drive back into the array, it's no longer showing as Wrong again and has the incorrect partition size

I can also mount the drive on my computer too, which is definitely a big step forward from before

I ran sgdisk -o -a 8 -n 1:32K:0 /dev/sdf as well on the last drive that had no partitions and that's now readable, however was still showing as wrong with the 4GB partition size, the only one remaining is the Parity which I'm not sure if that needs to be treated differently

tower-diagnostics-20251215-1216.zip

Edited December 15, 2025Dec 15 by conjuror3726
added latest diagnostics as state has slightly changed

Do you mean just unassigning the devices from the array and reassigning them show as wrong again? Looks like it's detecting the small partitions again, which is very strange.

Please reboot once more, with the drives assigned and post new diags after the reboot.

Yeah just unassigning, mounting via UD to check the data and re-assigning back to the same slot in the array
I think the last diagnostic I sent was after a reboot but I've rebooted again just to be sure

tower-diagnostics-20251215-1357.zip

The actual partition looks fine for the data disks, not sure where it's getting the small one, try this:

Tools - New Config - Keep all - Apply

Start the array without parity for now; confirm all the disks mount normally.

Also, was parity already synced/valid before this happened?

I believe the parity was valid before this happened. I haven't run the parity check since initialising it though, I haven't manually run a sync but nothing said it was out of sync (apologies I'm a noob at unraid, coming from synology where it is all done for you)

Last checked on Sunday, 7 December 2025, 14:57 (eight days ago)

I did have reconstruct write enabled though previously.

I've done that and everything seems to have mounted fine with the full partition size

it's understandably complaining that the shares aren't protected

The parity drive still has no partitions currently, would it be an idea to run sgdisk on it again or just rebuild parity? I guess that's where your question about parity is leading to?

Before doing any parity checks, would it be worth trying a restart just to make sure the partitions stuck this time?

If parity was valid, you can do do a new config again Tools - New Config - Keep all - Apply, but this time back on main, also assign parity, check "parity is already valid", start array. (this will create the partition for parity.)

A correcting parity check will need to be done, since there will be a few sync errors, typically in the beginning only, but before doing that, recommend rebooting to confirm the array remains valid, since I'm not sure what is causing the issue; I don't remember seeing a similar problem.

That seems to have stuck this time after a reboot! It looks like there are quite a few sync errors corrected - 151 so far, 0.7% completed

thank you so much for your help! 😀

Edited December 15, 2025Dec 15 by conjuror3726

24 minutes ago, conjuror3726 said:

It looks like there are quite a few sync errors corrected - 151 so far, 0.7% completed

It's normal, this was for mounting the filesystems without parity, but it should not find many more after those.

Finished the parity check with 9503 errors fixed, wonder if that might have been part of the problem. Tried restarting again and the array persisted, so hopefully all looking good now 🤞

That's still a relatively small amount of errors; if the config is persisting, it should be resolved. I would like to find out what was causing the issue, but that would require more testing, and at least the issue is resolved for you now.