Hi everyone,

I have a question about making my array disks slightly smaller on purpose.

My current setup uses two identical Intenso 2 TB external drives (one as Disk1 and one as Parity1). ChatGPT suggested the following procedure to "shrink" both disks by a few MB so that in the future I can replace them with other 2 TB drives from different manufacturers, even if those happen to be a few sectors smaller. The reason I have to do this is that UNRAID will not accept replacement disks that are a little smaller that the current disk: https://docs.unraid.net/legacy/FAQ/replacing-a-data-drive/#important-notes. And of course the replacement disks that I just bought actually are a few sectors smaller than my existing Intensio drives.

Before I try this, I’d like to double-check with the community whether this approach is safe and recommended.

Here is the proposed procedure:

Procedure

Step 1. Stop the Array

• In the Unraid GUI: Stop the array.

• Make sure no writes are happening during this time.

Step 2. Verify Backup

• Because we will modify partition tables, please ensure data is backed up or that you accept a minimal risk.

• Since both Parity and Disk1 are affected, a complete parity rebuild will be required anyway.

Step 3. Determine Current Disk Size

Run for each disk:

lsblk -o NAME,SIZE,PHY-SeC /dev/sdX
sgdisk -p /dev/sdX

(replace sdX with the actual device, e.g. /dev/sdc)
This shows the exact sector count and current partition layout.

Step 4. Create a New (Slightly Smaller) Partition

Delete the old partition and recreate it a little smaller than the maximum available size. Example:

sgdisk --delete=1 /dev/sdX
sgdisk --new=1:2048:3906000000 /dev/sdX

• --delete=1 removes the old partition.

• --new=1:2048:3906000000 creates a new partition starting at sector 2048 and ending before the disk’s last sector.
  The exact 3906000000 number must be adapted to your drive, typically leaving 100k–200k sectors of headroom.

This does not wipe the content, only the partition table.
However, Unraid will detect it as a new disk → array must be rebuilt.

Step 5. Re-add to Array

• Assign the resized disk(s) again in the GUI.

• Unraid will mark parity as invalid.

• Start a parity rebuild.

• Disk1 remains readable during this process.

Step 6. New Standard

From now on, both Parity and Disk1 are a little smaller.
Any future replacement disk that is equal or larger than this reduced size will be accepted, even if it is a few MB smaller than your original Intenso drives.

That’s what ChatGPT recommended.
Do you think this is a safe and reasonable approach? Or would you suggest a different method to guarantee future disk compatibility when replacing drives?

2T devices should all have the same sectors, from all manufacturers, though with SSDs they can be different since some can be 500G (and multiples) and others 512G (and multiples), but if they are the same, they should all have exactly the same sectors available.

If you still want to change that, you would need to recreate the partitions with a smaller size and do a new config.

2T devices should all have the same sectors, from all manufacturers

That is not true. I have an Intensio 2TB Model with 3907029168 sectors and a WD 2TB model with 3906963456 sectors. See the following box for proof.

Proof that same size hard drives can have different amount of sectors

root@BadServer:/dev# fdisk -l sdb
Disk sdb: 1.82 TiB, 2000365289472 bytes, 3906963456 sectors
Disk model: Elements 2621   
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 4674DFB3-C748-4594-9B3E-FA3D37932914
root@BadServer:/dev# fdisk -l sdf
Disk sdf: 1.82 TiB, 2000398934016 bytes, 3907029168 sectors
Disk model: External USB 3.0
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x920b5527

One of my Intensio drives broke down (Disk1). The other one of my Intensio drives (exact same model) is still running as Parity1.

So I bought a new 2TB disk, to replace the broken Disk1. Unfortunately that disk has 65712 sectors less than my current intensio dives. But UNRAID will only let you replace a disk with a disk that has an amount of sectors >= the amount of sectors of you current disk.

So I think my options are:

1. Buy many more 2TB disk from different manufacturers until I find one that has by chance equal or more sectors than my broken Intensio drive.

2. Whenever I need to replace a disk I upgrade to the next higher tear (2TB -> 4TB -> 6TB -> 8TB -> ...). Eventually I will either run out of money or there won't a bigger disk to buy.

I don't like these options as they are both not sustainable. Therefore I want to implement the following strategy:

1. I define a number of sectors as my standard, say 3906000000. Every 2TB hard drive will easily have more than this amount of sectors. Also I don't need the full 2TB, so it does not hurt to shrink the amount of sectors that are available to me.

2. Whenever I want to add a new disk to my array, I will first create a partition on that disk that is exactly 3906000000 sectors in size.

3. This way I will always be able to replace any 2TB disks with any other 2TB disk, because my disk will always have the exact same size, even if manufacturing varies.

I just got a little problem with this strategy. I am already operating a 2TB drive that is bigger than my future standard. So I would have to shrink that disk while there is productive data on it. Since I am not an expert, I asked ChatGPT how to do that. But since ChatGPT is known to hallucinate, I ask this community whether the approach that ChatGPT suggested is safe. If you have a better strategy for this problem, that does not involve me dribbling the partitions of a disk that is in productive use, then I would appreciate it if you share it.

To summarize:

1. Hard drives of different manufactures do have different amounts of sectors

2. UNRAID will only allow you to replace a disk if it has the same or a greater amount of sectors

3. Because of bad luck, the replacement disk that I bought does not have an equal or greater amount of sectors than my current disk

4. What is the best strategy to deal with that?

5. If the strategy that I explained above actually is the best strategy for this problem: how do I implement the strategy safely, given that my hard drives are already in use?

Thanks!

16 minutes ago, SoftwareKater said:

I have an Intensio 2TB Model with 3907029168

I assume this/was an external disk? If true it can happen with some, e.g., WD.

Intenso, AFIK, doesn't make HDDs, so they are likely using another brand, but that's not typical if it's an internal disk of a common brand.

If you still want to do that, the way to do it is to manually create a smaller partition on the disks, then add them to a new array. The disks will need to be reformatted; Unraid will keep the existing partitions.

Unraid 7.1.0+ required.

JorgeB, thank you for your time and answers.

If you still want to do that

I don't want to do that. In fact I am scared that I will lose my data, if I would do that. But the way I see it, I have no choice...

Do you have a better idea? What would you do in my situation?

Like mentioned, you should never run into that issue with typical disks. I've only seen that in the last 10 years when users are using shucked disks, and I would not recommend doing it if you don't have good backups. If you have adequate backups, there's basically no risk in doing it. After the new partitions are created and the data is restored, it's the same for Unraid.