TLDR: Server crashed, OS USB corrupted, restored from old backup before I added new parity drives, when array was started, server put back wrong drives as parity (that were now data drives) and unmounted new parity drives

I had a power outage 3 days ago and the OS USB drive got corrupted. My last backup was 10 months ago, BEFORE I got two new larger 24tb drives and made those parity, copying the parity data off the old 16tb & 20tb parrity drives, making the 16 & 20tb drives data drives. (I also had one data drive die after the backup was made, , slot 4 if that matters)

When I recovered the OS from the backup and started the array, updated the OS and dockers. I nioticed some data was missing (some seasons of a show, some movies).

Then I notices that the array made the 16tb and 20tb drives PARITY drives again, probably wiping any data on them. The two 24tb drives were listed as 'unassigned'.

1) Just checking that there was NO WAY I could have only set ONE of the 24tb drives as parity, with the 20tb as 2nd parity, and had the other 24tb as a data drive. (I definitely recall copying the 16tb parity data to a 24tb, just not sure if I did both 24's)

2) If both 24tb drives HAD TO BE parity drives, is there any way to set them as parity drives again, set the 16tb and 20tb drives as data again, and rebuild the array from the 24tb parity?

FWIW, here's diagnostics from 2 days ago (soon after crash, 0708) and from today (0710)

Thanks in advance.

EDIT: The two 24's, and the 16tb and 20tb, all seem to be 'parity' when trying to add them as data drives and starting the array.

i7-77-diagnostics-20260710-0842.zip

i7-77-diagnostics-20260708-1054.zip

Edited 3 hours ago3 hr by YaSo
tldr on top

TLDR: Server crashed, OS USB corrupted, restored from old backup before I added new parity drives, when array was started, server put back wrong drives as parity (that were now data drives) and unmounted new parity drives.

Is there any way to see if last week's 24tb parity drives can restore data to the 16tv & 20tb data drives that the OS thought were parity and (probably) messed up?

Did you start the array with the wrong parity drives?

2 hours ago, YaSo said:

and started the array

I see that you did. This will make it very difficult to recover, but there's a small possibility if you know which drive was parity and which wsa parity2 and the correct order for all the data disks, and parity is still mostly valid, that is, not much was written since the wrong start.

If you want to try, that would require doing a new config:

-Tools -> New Config -> Preserve current assignments: All -> Apply

-Check all assignments and assign any missing disk(s) if needed; both correct parity drives need to be correctly assigned as they were; they are not interchangeable, as well as the remaining data disks in their original order, or it will never work.

-IMPORTANT - Check both "parity is already valid" and "maintenance mode" and start the array (note that the GUI will still show that data on parity disk(s) will be overwritten, this is normal as it doesn't account for the checkbox, but it won't be as long as it's checked)

-Stop array

-Unassign both disks that were wrongly assigned as parity before

-Start array (in normal mode now), and post new diagnostics

5 minutes ago, JorgeB said:

Did you start the array with the wrong parity drives?

The situation OP is currently in is all of their disks are shown an "Unmountable" and they are not sure which two are the parity disks.

The plan was to attempt to identify which two disks were going to be reported as unmountable to determine which two disks were parity, assign them back to P1 and P2 slots. However we ran into a snag, due to issues with disk assignments missing (old super.dat) it wasnt possible to start the array to use unassigned devices to mount the disks with the RO flag.

Plan B was to invaldiate parity by mounting all the disks to the array slots only, after running tools > new config, and then simply identify which two disks didnt mount/have a valid filesystem. Ran into another issue where all of their disks are showing as unmountable.

The goal now is to fix the filesystems on the disks that had filesystems and rebuild parity as parity was invalidated.

Disks, 1,2,3,5,6 are known XFS disks,

Disks 4,7,8,9 are the unknown disks.

Note that the 16TB and 20TB were old parity disk assignments and one of these has since been re-assigned as a data disk by OP at some point during a parity upgrade.

Edited 2 hours ago2 hr by MowMdown

The only option that I can see is if it's possible to find out how the disks were assigned before, by looking for some old digs, or a flash backup, and even so, it may fail because the array was started and used with the wrong assignments, so parity will never be 100% valid, even with the disks correctly assigned, but it's worth a try as described above; if the assignments are known, there's nothing to lose at this point, but I don't see a way to find the correct assignments if old diags, backup, screenshot, etc., are not available

1 minute ago, JorgeB said:

The only option that I can see is if it's possible to find out how the disks were assigned before, by looking for some old digs, or a flash backup, and even so, it may fail because the array was started and used with the wrong assignments, so parity will never be 100% valid, even with the disks correctly assigned, but it's worth a try as described above; if the assignments are known, there's nothing to lose at this point, but I don't see a way to find the correct assignments if old diags, backup, screenshot, etc., are not available

I attached screenshots of OP's setup in my previous comment, all 14TB disks are known data disks, it's a matter of finding out which of the two 24TB, the 20TB, and the 16TB disks were data disks or parity disks.

The 14TB disks however are currently unmountable, how do we correct these so we can mount them?

From there how can we identify, of the 4 remaining disks, which two had a valid filesystem vs just being a parity disk?

Edited 1 hour ago1 hr by MowMdown

54 minutes ago, JorgeB said:

Did you start the array with the wrong parity drives?

Yes, but the 'old' Parity drives (from before the crash) were unassigned. The backup was from before I got the new, larger 24tb drives.

4 minutes ago, MowMdown said:

I attached screenshots of OP's setup in my previous comment, all 14TB disks are known data disks, it's a matter of finding out which of the two 24TB, the 20TB, and the 16TB disks were data disks or parity disks.

The 14TB disks however are currently unmountable, how do we correct these so we can mount them?

From there how can we identify of the 4 remaining disks had a valid filesystem vs just a parity disk?

If I am right, between the 16tb, the 20tb, and the two newer 24tb drives, I could not have had anything but the two 24tb drives as parity, right? Largest size drives have to be parity, right?

Yes, but you still need to know which one was parity and parity2

8 minutes ago, JorgeB said:

Yes, but you still need to know which one was parity and parity2

Not sure, but here's a thought...

I originally had 8 drives, (16+20tb parity drives, 5 good 14tb drives,and a dead 14tb drive)

I added two new drives. two 24tbs. 1st 24tb was added in same case location as the dead 14tb drive (slot 4 in Unraid) and 2nd 24tb was added as an extra drive with new data and power cables.

When I would have added the two 24tb drives as parity, would Unraid likely show the slot 4 drive as the first of the two available unassigned drives, and the new addition (9th) as the lower of the two options in the dropdown list when I was adding them? I can check drive serial number and go off of that

ALSO, is there any way to add the two 24tb drives as parity, not knowing which was in slot 1 or 2, and start in safe mode to not screw stuff up, just to check if the order is correct?