I recently moved my server into a new case and am adding some new drives my setup is currently as follows

8tb x 2 as parity

6tb x 3 data

8tb x 3 data

I bought four new 10tb on black Friday sales (spread across distributors) and want to replace my 2 parity with 10tb, and repurpose the 8tb current parity drives into the array, and add the other two 10tb into the array.

I want to remove the 6tb completely, as they are SMR drives and while I use the array for my Emby server and I know thats fine, they are also 5400 rpm and just make the parity checks so so slow.. (Is monthly checks overkill??)

My plan was to do the following after I asked on Reddit and got great help there, I just need to confirm a few things here with more a focused group. As I have come across a few things last night and today.

1. Preclear new 10tb, remove 1 8tb parity, rebuild parity

2. Clear previous parity 8tb, add into array

3. Use Unbalance plugin like in Space Invaders Video , Run unbalance as in video, Empty a 6tb onto past parity 8tb. Run user scripts as in video, clear 6tb, remove 6tb, restart array with "Parity is valid" checked

4. I want to physically remove the 6tb now, as replace it with a new 10tb, as I don't want both my parity disks on the exact same backplane in my server case.

5. Preclear 2nd new 10tb, remove 2nd 8tb, rebuild parity

6. Repeat steps 2 and 3 with other parity 8tb

7. Physically remove this 6tb from server case, put new 10tb in its place

8. Preclear other two 10tb, add into array, Run unbalance, User scripts, etc, empty final 6th onto a 10tb.

9. Remove final 6tb from case

Now I checked out the User scripts forum post Space Invader mentions, and it has a mod note saying that this method is no longer recommended as it takes a long time, so now I am a little confused and stuck. As the link it provides (Remove Drives Then Rebuild Parity" Method ) is no longer there.

So is is best to just remove a 6tb, reassign a 8tb into its place and rebuild data? And do that 3 times over? It seems that is what is showing as the thing to do in the Unraid Docs.... It seems semi risky and very intensive to rebuild the data rather than move the data..

Help? Guidance? Suggestions?

Thanks!

A point to note is that Unraid recognises disks by their serial number - not by how they are connected so (as long as you are not using a RAID disk controller). You can reply them to different physical connectors without affecting your Unraid configuration.

A much faster process to achieve what you describe would be:

• stop array and replace both iTB parity disks with the 10TB drives you plan to use to replace them.

• Start array in Maintenance mode to rebuild both parity disks in parallel. Keep the old parity disks intact at this point to give you a recovery scenario if one of the data drives fails before parity is rebuilt.

• Stop the array and unassign 2 of the the 6TB drives you want to replace.

• Start the array and confirm that Unraid mounts the ‘missing’ drives by emulating them using the remaining drives + the new parity drives. Check the contents look OK.

• Stop the array and assign the old 8TB parity drives in their place.

• Start the array in Maintenance mode to rebuild their contents onto the 8TB drives. No need to clear these drives first as the rebuild overwrites every sector anyway. Keep the old 6TB drives intact for now

• Stop the array and Unassign the remaining 6TB drive.

• Start the array to check it is being correctly emulated.

• Stop the array.

• Assign another 10TB drive in place of the just removed 6TB drive.

• Start the array to rebuild the contents of the old 6TB drive onto the 10TB drive

• You can then follow the standard process for adding the last 10TB drive as a new data drives.

BTW: the documentation has not been removed - it has just been reorganised. Array operations are covered here

Interesting. I was worried about loss of parity if I did both parity at once...

So Unraid will remember the 6tb if something terrible happens? I was worried that if I did it that way, and something failed, I would not just be able to remount and reassign the disk and have Unraid recognize it and go about its day like nothing happened...

7 hours ago, I_TheRenegade_I said:

So Unraid will remember the 6tb if something terrible happens? I was worried that if I did it that way, and something failed, I would not just be able to remount and reassign the disk and have Unraid recognize it and go about its day like nothing happened...

Not quite that simple, but if you encounter a failure at any point then ask here describing how things failed and we can guide you through the recovery steps. The important thing is to keep the old drives intact until the step to replace them has completed successfully.

OK, thanks a bunch! I'll go ahead with that and hopefully you don't see me back in here in a few days freaking out!

Thanks a bunch!