Hi all,

I’ve hit a double failure and want to confirm the safest way to perform a parity-swap and sequential rebuilds. I’ve read the Unraid docs thoroughly but would appreciate experienced eyes on this plan before I commit.

---

System

• Unraid OS: 7.1.4

• Server: “Bearcave”

• Motherboard: MSI Z590-A PRO (4 SATA ports)

• HBA: ASM1166 PCIe SATA card

• PSU: Be Quiet! 1000 W

• Cache: Samsung 850 EVO 500 GB (SATA) + Dual NVMe WD SN570 1 TB (BTRFS RAID1)

---

Disk Situation

Role	Model	Capacity	Status
Old Parity	WD100EMAZ	10 TB	Healthy
New Parity	Seagate Exos X22	22 TB	Installed
Data Disk 1	WD100EZAZ	10 TB	Dead
Data Disk 2	ST8000VN004	8 TB	Failed
Data Disk 3	ST8000NE0021 (IronWolf Pro)	8 TB	Healthy

New drive arriving tomorrow: Seagate Exos X16 10 TB (to replace the dead WD100EZAZ).

---

Current State

• 22 TB Exos assigned as new Parity (array stopped).

• Old 10 TB parity temporarily assigned to Disk 1 but array never started, no format performed.

• Docker + VMs disabled. UPS configured.

• Both failed data drives physically removed until replacements arrive.

---

Goals

1. Run a parity-swap: copy parity from the old 10 TB → new 22 TB.

2. Once that copy finishes, rebuild the failed 8 TB data disk onto the old 10 TB parity drive.

3. After that completes, replace the dead 10 TB data disk with the new Exos X16 10 TB and rebuild again from the freshly-built 22 TB parity.

4. Finish with a non-correcting parity check before re-enabling Docker / VMs.

---

Questions

1. Since the old 10 TB parity was briefly assigned to Disk 1 (array never started), it’s still valid for parity-swap use, correct?

2. Any metadata or config changes just from assigning a disk without starting the array?

3. Will Unraid automatically expand parity when moving from 10 TB → 22 TB during the swap, or is manual intervention required?

4. Best order to handle two failed data drives—perform parity-swap first, rebuild one disk, then the other?

5. Any issues or limitations in 7.1.4 related to the ASM1166 controller during parity-swap / rebuilds?

6. Should the old 10 TB parity remain untouched (no preclear) so Unraid can use its existing parity data for the swap?

7. Is it useful or redundant to run a non-correcting parity check after the parity-swap copy but before the first rebuild?

8. Any extra diagnostics, SMART checks, or log monitoring you’d recommend before starting?

---

Diagnostics and SMART reports are ready if needed. I just want to ensure this dual-rebuild chain is the right and safest approach before touching Start.

Thanks for any expert insight—trying to keep this resurrection to one attempt.

You cannot recover from two failed disks with single parity. You can post the diags to see if both disks have actually failed.

24 minutes ago, JorgeB said:

You cannot recover from two failed disks with single parity. You can post the diags to see if both disks have actually failed.

How do I get to the diags ?

So what should I do I was going to just add the 22tb and copy parity over from the old 10tb parity then add the new 10tb and re run parity.

But your saying that’s nit going to work what should I do next.

I have no idea I had 1 8tb that’s failed I can hear it running but it clicking and not spinning up all the other had a green tick and smart check was fine in all of them.

I find it weird another 10 tb just gone dead it’s like a brick

31 minutes ago, JorgeB said:

You cannot recover from two failed disks with single parity. You can post the diags to see if both disks have actually failed.

13 minutes ago, OsoPolar said:

How do I get to the diags ?

TOOLS >>>> Diagnostics

For complete instructions, click on the word diagnostics in any post-- including the one in this sentence!

On 10/8/2025 at 10:22 PM, Frank1940 said:

TOOLS >>>> Diagnostics

For complete instructions, click on the word diagnostics in any post-- including the one in this sentence!

Hey all, quick update after some cable and power gymnastics.

I’ve managed to recover my old 10 TB parity drive — turns out it was a shucked unit being blocked by the 3.3 V pin issue. After switching PSUs, it’s now fully visible and passing SMART.

Current state:

• Parity: 22 TB Seagate Exos (new)

• Disk 1: 10 TB WD100EMAZ (old parity)

• Disk 2: 10 TB WD100EZAZ (healthy)

• Disk 3: 8 TB Seagate (healthy)

• Array has never been started since drives were rearranged

• One new spare 10 TB drive still unused

Plan going forward:

1. Run an extended SMART test on the 22 TB to confirm it’s clean.

2. Proceed with a parity-swap rebuild — copy existing parity from the old 10 TB onto the new 22 TB, then rebuild Disk 1’s contents onto the old 10 TB drive.

3. After completion, run a non-correcting parity check to verify zero errors.

4. Optionally pre-clear and add the spare 10 TB as a new data disk once the array is healthy.

Just want to confirm this sequence is still the safest approach before starting the swap. Diagnostics from this current state are attached (bearcave-diagnostics-20251010-1127.zip).

Thanks again for the guidance so far — it’s been a saga, but the drives are finally all visible.

bearcave-diagnostics-20251010-1127.zip

1 hour ago, OsoPolar said:

Optionally pre-clear and add the spare 10 TB as a new data disk once the array is healthy.

I would do the pre-clear but not add it to the array until I needed the space. (Perhaps, you need it now but that info not apparent...) That way if one of your other data disks should fail, you could use it as a immediate replacement. Save to get another 22TB drive for that time when the 10TB will be used.

Edited October 10, 2025Oct 10 by Frank1940

So do you mean get another 22tb for parity 2 and keep the 10tb spare for another failure.
If pick up a 22tb on a BF Deal can i just add that?

Also on my current plan does that look ok ?
I have never added a new drive or moved any parity about ?

15 minutes ago, OsoPolar said:

So do you mean get another 22tb for parity 2 and keep the 10tb spare for another failure.

My take is that the more drives in a system equals more things to fail. I make if a point to never add a drive to the array just because I physically have it. There has to be a truly valid reason to add it!

Adding a second 22TB parity is a matter of your 'comfort' factor in accepting risk. If you will sleep better at night knowing you have it installed, then do it. If you think sky-diving is the best way to get a 'high', then you can probably avoid doing it. (When you have eight to ten data disks in your array, a second parity drive becomes more statistically justifiable from a reliability standpoint.)

16 minutes ago, OsoPolar said:

f pick up a 22tb on a BF Deal can i just add that?

Look for that BF Deal. That's what I tend to do. (I have two 12TB drives pre-cleared on my shelf waiting for a failure or to be required to provide additional storage space. I am in no big hurry to install them! If you look at my forum signature, you will find that I have dual parity on one server and single parity on the other one. I am willing to accept the risk on that one server...)

Thanks thats why i only ever went with one as i only had 3 drives at least now i have a spare 10tb and i can add bigger drives upto 22tb in future but i have enough for my use thank you

i think if i only use 3 or 4 drives the chance of 2 failing same time is slim

Just for clarity Frank
Does my current plan look ok ?
I have never added a new drive or moved any parity about ?

18 minutes ago, OsoPolar said:

i think if i only use 3 or 4 drives the chance of 2 failing same time is slim

Look here for the statistical view:

https://forums.unraid.net/topic/50504-dual-or-single-parity-its-your-choice/

Note that the 2% failure-rate file is probably more typical of what most Unraid users are seeing today. (I have a WD 3TB drive that the hour-on counter (Attrib # 9) has 'rolled over' and now shows 2 years and 2 months! I would guess it has been in use for, at least, ten years now.)

EDIT

BackBlaze data for a recent quarter:

https://www.backblaze.com/blog/backblaze-drive-stats-for-q2-2025/

Edited October 10, 2025Oct 10 by Frank1940

22 hours ago, Frank1940 said:

Look here for the statistical view:

https://forums.unraid.net/topic/50504-dual-or-single-parity-its-your-choice/

Note that the 2% failure-rate file is probably more typical of what most Unraid users are seeing today. (I have a WD 3TB drive that the hour-on counter (Attrib # 9) has 'rolled over' and now shows 2 years and 2 months! I would guess it has been in use for, at least, ten years now.)

EDIT

BackBlaze data for a recent quarter:

https://www.backblaze.com/blog/backblaze-drive-stats-for-q2-2025/

Is this the correct plan of action

🧾 Phase 1: Finish the PreClear test

1. Let the 22 TB Exos finish its PreClear

   • Don’t touch the array or reboot while it’s running.

   • When done, it should read: “Completed without error.”

2. Once it’s finished, set its spin-down delay back to your normal value (e.g. 15 min).

---

⚙️ Phase 2: Prepare for the parity-swap

1. Power down the server (optional, but good hygiene) and re-seat SATA + power cables.

2. Boot Unraid, do not start the array yet.

3. Confirm drive assignments on the Main page match this:

   • Parity: 22 TB ST22000NM001E

   • Disk 1: 10 TB WD100EMAZ (old parity)

   • Disk 2: 10 TB WD100EZAZ (data)

   • Disk 3: 8 TB ST8000… (data)

   • Nothing else assigned.

4. Verify that Unraid reports:

   Parity disk valid – replacement detected – ready for copy
   (or similar wording about a parity-swap.)

---

🧮 Phase 3: Run the parity-swap rebuild

1. Click Start on the array.

2. Unraid begins Stage 1: copy parity from the old 10 TB to the new 22 TB.

3. When that finishes, it automatically starts Stage 2: rebuild Disk 1’s data onto the old 10 TB drive.

4. Let both stages complete without interruption.

   • Expect roughly 30–45 hours total.

   • Monitor temps (<45 °C) and speed occasionally.

---

✅ Phase 4: Verify everything

1. When Unraid says “Array operation completed successfully,” stop and exhale.

2. Go to Main → Array Operation → Parity Check.

3. Run a non-correcting parity check — this reads everything and ensures the new parity data is perfect but doesn’t change it.

   • It should finish with 0 errors.

4. Once that’s clean, your array is fully rebuilt and protected again.

---

🪄 Optional final touches

• Add the spare 10 TB as an Unassigned Device and preclear it so you’ve got a ready-to-go standby.

• Label each drive (slot, serial, install date).

• Back up the flash drive (Main → Flash → Flash Backup).

---

Do all that in order and you’ll end up Monday or Tuesday with:
✔ a verified 22 TB parity,
✔ a rebuilt 10 TB Disk 1,
✔ zero parity errors,
✔ one precleared spare drive ready for the next inevitable drama.

9 hours ago, OsoPolar said:

When that finishes, it automatically starts Stage 2: rebuild Disk 1’s data onto the old 10 TB drive.

This is not automatic; you need to start the rebuild. Don't do anything else before starting, like rebooting or reassigning devices, or you will need to start over.

11 hours ago, JorgeB said:

This is not automatic; you need to start the rebuild. Don't do anything else before starting, like rebooting or reassigning devices, or you will need to start over.

HI
Sorry to be a noob about this but i have never had to do this before.
See image attached this the brand-new parity you can see is marked by 1 22 TB
The old parity is seated in Disk 1 now where the broken disk was.

I have not had the array on or rebooted.
So how do i get the old parity to the new parity correctly ?
Then how does the former broken disk 1 get rebuilt onto the new disk 1(which was the old parity)

What do i actually do in Unsaid to get it running ?

If the assignments are as you mention, you should see the option to begin the parity copy show next to the array start button

1 minute ago, JorgeB said:

If the assignments are as you mention, you should see the option to begin the parity copy show next to the array start button

HI Jorge

Yes i see the copy see image 1but i need to check if when i start this does this fire up the array ?
But if it does fire up the array does will it start docker and VM or do i need to set these to off ? as in image 2 and 3 ?

To avoid data writing disks is this the best option ?

Image 1

1 minute ago, OsoPolar said:

Yes i see the copy see image 1but i need to check if when i start this does this fire up the array ?

Nope, for the parity copy portion of the parity swap feature, the array remains offline, it does come online for the 2nd part, when it starts the rebuild, unless you start in maintenance mode, see the linked docs for more info.

On 10/12/2025 at 11:29 AM, JorgeB said:

Nope, for the parity copy portion of the parity swap feature, the array remains offline, it does come online for the 2nd part, when it starts the rebuild, unless you start in maintenance mode, see the linked docs for more info.

HI
Everything finished Partity transfer no errors and rebuild no errors bit i a seeing this now its doing a parity check

Can i now use Uraid as normal ? Or do i have to wait another day ?

So the unclean shutdown was after the rebuild finished? You can use the server, but array performance will be slower.

Can i thank everybody for there great help back up and running and will get 2 drives precleared ready for next time