Hi

Basically, all my data disks in the unraid array are failing.

They've had a good life (over 65k hours for the younger) but it's time to become e-waste I think.

I decided to replace all the disks in the array (Parity and data) with two 12TB disks (one for parity, the other as data).

I'm using the Basic licence so 6 disks maximum (I'm using a cache disk as well), so I have 1 extra slot.

Downtime on the array is not an issue.

What would be the best way to do the switch?

I would like to avoid too much stress on the data disks that are about to fail.

Do you have a backup of the data currently on your array?

If not, I would suggest adding one of the 12TB drives as an Unassigned Device and back up everything.

After you have everything safely backed up, you can then reconfigure your array as you would like.

That's a neat idea.

Could I copy everything from the array to a 12TB disk and then use this disk as a data disk in a new config with the other 12TB disk as parity?

I'm not 100% sure what I'm going to write is correct, so I'm going to tag a few experts to confirm I'm not just blowing smoke: @JorgeB @itimpi @ChatNoir

This should be possible. Once you have all of the data copied from the existing array to one of the new drives, you would first recreate your array (using the New Config utility in the Tools tab). Add ONLY the new data drive at this time, and remove all of the current drives from your array (including your current Parity drive). You will not need to clear or format the new drive (and obviously should not, as it now has all of your data which you copied on it).

At this point, you should be able to start your array and see the one new drive in it. All of your data should be accessible.

Now shut down your system, remove the drives you no longer need, and add the second 12TB drive you will use for parity. When you start the server back up, you should now be able to add the drive as Parity from the Main tab (array need be stopped). When you restart the array, parity rebuild will start.

That should work. Note that disk3 is not failing, at least based on SMART, CRC errors are typically due to a bad SATA cable. The other ones might be, though in some cases pending sectors show as false positives. If they are real, some files may fail to copy due to I/O errors, but at least you will know which files you will need to recover.

Thanks all, I should be receiving the new disks today. I'll try that.

Actually I went over the docs (again again) and might have found the solution (although it's in the Legacy documentation):

https://docs.unraid.net/legacy/FAQ/replacing-multiple-data-drives/#faster-method

Faster Method

This method disables the parity drive while transferring data from the smaller drives to the larger drive, and then rebuilds parity at the end. Not having parity allows faster writing to the larger drive.

1. Stop array and power down.

2. Add new larger drive and power up.

3. Stop array and use Tools - New Config.

4. Assign new larger drive to new slot, unassign parity drive, and start array.

5. Copy all files from the smaller drives to the larger drive. For example, if the smaller drives are Disk 1 and Disk 2, and the larger drive is Disk 3, then you could use these commands in console or telnet:

   rsync -avX /mnt/disk1/ /mnt/disk3
   rsync -avX /mnt/disk2/ /mnt/disk3
   

6. Stop array and power down.

7. Remove smaller drives and power up.

8. Use New Config to unassign removed drives and assign parity. You can also change the slot assignment of the new drive or any other drives at this point.

9. Start array and let parity rebuild.

Looks a lot like the procedure @ConnerVT was suggesting.

I'll disable everything besides the array during the procedure to avoid conflicts (SMB, Docker, mover and parity schedules).

On 8/4/2025 at 10:08 AM, JorgeB said:

That should work. Note that disk3 is not failing, at least based on SMART, CRC errors are typically due to a bad SATA cable. The other ones might be, though in some cases pending sectors show as false positives. If they are real, some files may fail to copy due to I/O errors, but at least you will know which files you will need to recover.

Thanks, while powered off, I'll change the SATA cables to be on the safe side.

The disks have been showing warnings for a few months now.

The procedure from the Unraid docs assume a healthy array. I suggested a potentially safer alternative - Getting all of the array data backed up (utilizing a simple "copy everything" strategy) first. This limits the amount of work your aging array drives need do. Added benefit is if there are any issues, the old array drives still contain your data if needed.

Recreating the array with just the new data drive eliminates Unraid needing to clear/zero/format the drive. Then adding the Parity drive just requires rebuilding parity, which you can then be confident is correct, as it was just done with two new, healthy drives.

Finally finished the operations.

I stopped every service running on the server (Docker, SMB, mover and parity check schedules...).

I mounted a new 12TB disk with Unassigned Disk, copied everything in /mnt/user0 into that mount point using the File Manager in the unraid webgui.

It took a while.

Then did a check of all the data with diff:

diff -rq --no-dereference /mnt/user0 /mnt/disks/diskID

It took another while but no difference were found.

Then powered down the system, removed all the HDD, installed the two new 12TB drives and powered up.

Used New Config, configured the disk with the data as Disk 1, the other as Parity.

Parity operation is still running but everything works like a charm.

Thank you for the help.

I love unraid :)

Edited August 7, 2025Aug 7 by Guillaurent