Hi,

 

I stopped my array this morning and I'm now getting this error message:

All existing data on this device will be OVERWRITTEN when array is Started

Disk 1 is seen as a replacement disk by unraid, which is not correct. No disk was changed.

Additionally this disk shows a high number of UDMA CRC errors. I'm not sure if these errors were already there yesterday, but Unraid did send a notification about these, after shutting down the array.

While UDMA CRC error count often means that the SATA cable is defect this won't help me with the error message that the data will be overwritten, right?

Edit: I just checked, and the UDMA CRC error count was the same in January, therefore I don't think that this is an issue.

 

What should be my course of action?

 

If the drive is really defect, I don't want to replace it, because I've got enough space left on the other drives, should I proceed like that?
1. Unassign Disk 1

2. Copy the emulated data from Disk 1 to somewhere else

3. Create a new config without Disk 1, and rebuild Parity.

 

If the drive is not defect (which is what I currently assume), I should be able to just start the array and the same data would be rebuild on Disk 1, without any data loss, right?

 

gandalf-diagnostics-20230912-0947.zip

Edited September 12, 20232 yr by MarcelCliff

Disk1 is being detected with a slightly larger capacity, did you change controllers?

47 minutes ago, JorgeB said:

Disk1 is being detected with a slightly larger capacity, did you change controllers?

 

No, I did not touch the hardware.

The last thing I did before stopping the array was moving ~7TB to Disk 1 which was newly formatted as zfs.

That's strange, disks don't usually change capacity on their own, but since it did you'll need to rebuild, check that the emulated disk is mounting first.

2 hours ago, JorgeB said:

check that the emulated disk is mounting first.

I'm not sure what you mean by that.

 

If I understand you correctly I should do the following:

1. Start the array with Disk 1 disabled

2. Check if files on (emulated) Disk 1 is readable

3. Stop the Array

4. Assign the same disk to the first Disk's slot

5. Start the Array

6. Wait for the rebuild to complete.

 

Is this correct?

2 minutes ago, MarcelCliff said:

1. Start the array with Disk 1 disabled

2. Check if files on (emulated) Disk 1 is readable

Stop at this point, post new diagnostics, await further instructions

1 hour ago, trurl said:

Stop at this point, post new diagnostics, await further instructions

I checked some files on the emulated disk and they seem okay.

Here is the new diagnostics file.gandalf-diagnostics-20230912-1732.zipgandalf-diagnostics-20230912-1732.zip

Emulated disk is mounting, if contents look correct you can rebuild.

3 hours ago, JorgeB said:

Emulated disk is mounting, if contents look correct you can rebuild.

 

Do you think that rebuilding is better than moving the files off of the emulated disk 1?

 

Not sure if I can trust Disk 1 or the controller if the reported size suddenly changed.

 

Edit: Rebuild is running now.

Edited September 12, 20232 yr by MarcelCliff

6 hours ago, MarcelCliff said:

moving the files off of the emulated disk 1

MOVE is copy from source to destination then delete from source.

 

To read the files on the emulated disk, its contents must be calculated from parity by reading all other disks. Then, to move the files to other disks in the array, the other disks have to be written, updating parity. And, the emulated disk must be written, to delete the moved files, again updating parity.

 

So, rebuild, at the very least, has the advantage of not modifying parity while your array is compromised.

 

If you feel there is something on the emulated disk that needs to be backed up before attempting rebuild, COPYING those files somewhere off the array would be a better idea.

I've just had this exact same issue - I'm going through moving files off of one disk at a time to the rest of the array, then swapping to ZFS

 

The last disk I changed to ZFS, moved all the files off the next disk onto it, when I stopped the array I got the error saying it was the wrong disk. And checking the disk logs the reported size changed after the ZFS format, but the array config obviously didn't update.

 

Not good!

Posted a new thread as I'm currently mid-panic and hoping for some help 😱