Hello! I set up a new Unraid server for the first time and encountered an issue today where two of my disks are showing "Wrong" after a reboot.

The system has been working smoothly since adding these disks (connected via SATA directly to the motherboard), however today I shutdown the server using the button in the webgui so I could install an NVME SSD to be used as a cache drive. The system didn't move or have any other physical changes. After booting the server back up they show as "Wrong".

These disks had their previous partitions erased using destructive mode and the Unassigned Devices plugin. They were then added to the array with the default file system (xfs). I then started the array and selected the option to format the drives.

Based on other forum posts around this topic, I was able to check my logs (unfortunately no pre-reboot logs) and see two potential issues:

1. In syslog.txt I can see the devices for disk8 and disk9 are showing up as having "no partitions":
Dec 19 09:00:24 Tower emhttpd: online: WDC_WD180EDGZ-11BLDS0_8LJ34RVL (sdg) 512 35156656128 => no partitions
...
Dec 19 09:00:24 Tower emhttpd: online: ST6000NM0004-1FT17Z_Z4D43600 (sdc) 4096 1465130646 => no partitions

2. In vars.txt I can see that [size] and [sizeSb] do not match for disk8 and disk9:
[name] => disk8
[size] => 5860522532
[sizeSb] => 5860388864
...
[name] => disk9
[size] => 17578328012
[sizeSb] => 17578194944

Based on this post for a very similar issue, I'm thinking of trying to rebuild the standard Unraid partition layout using the following for the respective drives:
sgdisk -o -a 8 -n 1:32K:0 /dev/sdc
sgdisk -o -a 8 -n 1:32K:0 /dev/sdg

I am not completely sure how to proceed without data loss at this point, although I have dual parity and the data is technically non-essential.

Would love some advice!

tower-diagnostics-20251219-1000.zip

The disks don't have partitions. Any chance you have the diags/syslog before the reboot? Also, do you know where the disks came from, i.e., where they were originally partitioned?

2 minutes ago, JorgeB said:

The disks don't have partitions. Any chance you have the diags/syslog before the reboot? Also, do you know where the disks came from, i.e., where they were originally partitioned?

Unfortunately no diags/syslog from before the reboot. They were originally partitioned as part of a Windows Storage Spaces pool before I moved them to Unraid, where I performed the steps listed above.

I have a good guess what the starting sector is for the partitions but cannot be certain because the previous partitions had some slack, and Unraid only records the expected size, not where it started, so it's not possible to exactly know with just the diags with the syslog-previous also present (recommend enabling that).

If the data is important, and since you have dual parity, I would recommend rebuilding both. Assuming parity can emulate them, and it should if it's valid, you can test beforehand by unassigning both disks and starting the array, then check if both emulated disks mount and the contents look correct.

If all looks good, stop the array, reassign both disks, and start to rebuild. If the emulated disks don't mount, don't rebuild; we can try other options.

Also recommend upgrading to 7.2.3, I believe this issue is related to a bug that existed in all previous 7.2 releases where the preclear signature was not detected, and that was fixed in 7.2.3 (but upgrading now won't resolve the current issue; it would just not happen again in the same circumstances).

I tested the emulated disks and all looked good, started rebuilding from parity.

Once it's complete I'll upgrade to 7.2.3 and reboot to make sure the issue doesn't reoccur.

Thank you!

You're welcome, also, please confirm if new disks were added to the array without being cleared, i.e., you did a new config to add them and then resynced parity instead of adding them to the array as new disks and letting Unraid clear them.

7 hours ago, JorgeB said:

You're welcome, also, please confirm if new disks were added to the array without being cleared, i.e., you did a new config to add them and then resynced parity instead of adding them to the array as new disks and letting Unraid clear them.

These were added without being cleared, but that was because I didn’t have parity set up yet.

I already had disks 1-7 in the array and used rsync commands to mirror the data from my previous Windows machine, which was using Storage Spaces to combine what is now my disk9 and both my parity disks into a simple pool.

Once I confirmed all the data was in the array, I moved the 3 disks from the Windows machine to the Unraid server, I also added 1 spare disk which wasn’t being used to the server at the same time (which became disk8). I deleted the partitions from the 4 new disks using Unassigned Devices and added disk8 and disk9 to the array as new disks without clearing them and let Unraid format them. Then I added the two parity disks and let parity sync complete.

The shutdown and boot yesterday was the first time I had done a power cycle after all these changes. Is there any risk of this issue occurring with any of the other drives?

16 hours ago, iojeka said:

These were added without being cleared, but that was because I didn’t have parity set up yet.

OK, thanks for confirming. It should never have happened with cleared disks, but just wanted to be sure.

17 hours ago, iojeka said:

Is there any risk of this issue occurring with any of the other drives?

Not if you clear them. Also, with 7.2.3 it should never happen even if you did the same.

Just to wrap this one up. Rebuild completed, set up local syslog, updated to 7.2.3, and rebooted. All looks good now!

Thanks for the quick and useful support JorgeB :)