Hi all! I am experiencing a funky issue with my 4 12TB N300 drives (details here), they cannot run SMART self tests and I do not trust them. My array is setup with 2 parity, 2 data disks. I plan to migrate to 2 16TB drives (IronWolf Pro, WD Red Pro), at least for the time being. I'm pre-clearing these disks now. My current data usage is 11TB so everything will fit in just 16TB.

 

Below is my plan for the migration. I am trying to keep it relatively simple and while parity is not preserved, I'm not deleting anything from the old drives yet and the old drives still seem to be working so that's my "backup plan" tbh. Although I do have critical data backed up in multiple offsite locations.

 

Hoping to get a sanity check on the following plan. I refer to the 4 old parity/data disks as OP1-2/OD1-2, and the new ones as NP/ND.

 

1. Replace Disk 1 (OD1) with ND and rebuild from parity
2. Disable the mover and stop docker
3. Copy everything from OD2 to ND using unBalance plugin
4. Shut down and remove all old disks
5. New config
   1. Preserve Pool assignments (NVME cache pool)
   2. Assign NP as Parity, keep ND as Disk 1
6. Start Array, parity will rebuild
7. Try to get money from Toshiba

 

WDYT?

You cannot replace disk1 with a 16TB drive while parity is 12TB, would need to upgrade parity first, remove parity2, then replace disk1.

Thanks JorgeB. That oversight certainly does foil my plan. Back to the drawing board, does this look right?

1. Follow procedure for upgrading parity, replacing OP1 with NP
2. Remove OP2 (Stop array, unassign OP2, start array)
3. Follow procedure for upgrading a data disk, replacing OD1 with ND
4. Follow the "Clear Drive Then Remove Drive" Method for removing OD2

You seem to have left out step 3.5?    Copy OD2 to ND.   Your procedure as documented would lose the contents of OD2.

Ah, yes. I missed it copying the steps line by line from my editor trying to use the ordered list feature here on the forum. So then, the full procedure looks like this. Sounds like it'll work.

 

1. Follow procedure for upgrading parity, replacing OP1 with NP
2. Remove OP2 (Stop array, unassign OP2, start array)
3. Follow procedure for upgrading a data disk, replacing OD1 with ND
4. Move everything from OD2 to ND using unBalance plugin
5. Follow the "Clear Drive Then Remove Drive" Method for OD2

 

Edited February 4, 20233 yr by neal_is_king

I am officially under way now, 200+ hours of pre-clearing on these new disks and step 1 should be finished overnight. I had a question from the "Clear Drive Then Remove Drive" documentation, wondering if anyone knows:

 

Quote

Make sure that the drive you are removing has been removed from any inclusions or exclusions for all shares, including in the global share settings.

 

By the time I get to this step, I'll have Disk 1 (the new 16TB data disk with everything on it) and Disk 2 (the old 12TB disk that I am shrinking out of the array). Just want to make sure, but for each share that is not set to "Cache: Only", I want to change the Included Disks option from "All" to only "Disk 1", and I can safely leave Excluded Disks the way it is for all of these shares (which is "None")? And the reason for this is in case a file is written to the array after I've already cleared everything off of Disk 2, to prevent unraid from placing that new file on Disk 2?

 

Just to close the loop on this, I got it all sorted out and the last parity check finished without any errors. Thanks for your help JorgeB and itimpi