Unable to replace failed 4TB drive: new disk is reported as ‘device is too small’ on TrueNAS SCALE 25.10 (Goldfish)

nico_lucas · November 24, 2025, 6:18pm

Hi everyone,

I’m running TrueNAS SCALE 25.10.0.1 – Goldeye, and I’m having trouble replacing a failing 4TB SMR disk in a 4TB mirror pool.

I bought a brand-new 4TB WD Red Plus (CMR) to replace the old failing WD40EFAX (SMR).

I offlined the old disk, physically removed it from the NAS, and inserted the new CMR drive.

Now the system shows the following:

root@truenas[~]# lsblk -o NAME,SIZE,MODEL
NAME SIZE MODEL
sda 3.6T WDC WD40EFAX-68JH4N1 (the good disk, SMR)
├─sda1 2G
└─sda2 3.6T

sdc 3.6T WDC WD40EFPX-68C6CN0 (the new 4TB WD Red Plus CMR)
└─sdc1 216.6G

When I try to replace the disk, I get the following error:

EZFS_BADDEV: cannot replace […] device is too small

I tried “force” but not work, scuse me for my bad English …

can anyone help me?

winnielinnie · November 24, 2025, 6:23pm

Why is there a 216GB partition on the new disk?

You might have to prepare it by wiping it, using the TrueNAS GUI, and then try using it again.

nico_lucas · November 24, 2025, 6:27pm

It’s probably a residue from TrueNAS while trying to add the disk. I just wiped the new disk just to show you the clean situation:
root@truenas[~]# lsblk -o NAME,SIZE,MODEL
NAME SIZE MODEL
sda 3.6T WDC WD40EFAX-68JH4N1
├─sda1 2G
└─sda2 3.6T
sdc 3.6T WDC WD40EFPX-68C6CN0

thank you if you can help me

winnielinnie · November 24, 2025, 6:35pm

That output seems incomplete. I don’t think you only have 2 disks in the server. There’s no boot device in your output.

What is the output for both:

smartctl -a /dev/sda | grep "User Capacity"

smartctl -a /dev/sdc | grep "User Capacity"

nico_lucas · November 24, 2025, 6:39pm

root@truenas[~]# lsblk -o NAME,SIZE,MODEL

NAME SIZE MODEL

sda 3.6T WDC WD40EFAX-68JH4N1

├─sda1 2G

└─sda2 3.6T

sdb 232.9G Samsung SSD 850 EVO 250GB

├─sdb1 260M

├─sdb2 216.6G

└─sdb3 16G

sdc 3.6T WDC WD40EFPX-68C6CN0

sdd 1.8T ST2000DM008-2FR102

├─sdd1 2G

└─sdd2 1.8T

sde 1.8T ST2000DM008-2FR102

├─sde1 2G

└─sde2 1.8T

sdf 232.9G Samsung SSD 860 EVO 250GB

└─sdf1 232.9G

zd0 100G

root@truenas[~]#

root@truenas[~]# smartctl -a /dev/sda | grep “User Capacity”

User Capacity: 4,000,787,030,016 bytes [4.00 TB]

root@truenas[~]# smartctl -a /dev/sdc | grep “User Capacity”

User Capacity: 4,000,787,030,016 bytes [4.00 TB]

root@truenas[~]#

winnielinnie · November 24, 2025, 6:41pm

This might be a bug in CE 25.10.0.1.

Not only do you have a 2GB buffer in the existing disk, but the replacement disk is large enough even without it.

If you try again, does it result in a 216GB partition on the new disk again?

nico_lucas · November 24, 2025, 6:48pm

ok thanks, I tried to replace again and same error:

“[EZFS_BADDEV] cannot replace 8556839486299796572 with /dev/disk/by-partuuid/22e87612-89d5-4bdb-b728-e0ef6e99078d: device is too small”

and the 216gb partition reappeared:

root@truenas[~]# lsblk -o NAME,SIZE,MODEL

NAME SIZE MODEL

sda 3.6T WDC WD40EFAX-68JH4N1

├─sda1 2G

└─sda2 3.6T

sdb 232.9G Samsung SSD 850 EVO 250GB

├─sdb1 260M

├─sdb2 216.6G

└─sdb3 16G

sdc 3.6T WDC WD40EFPX-68C6CN0

└─sdc1 216.6G

sdd 1.8T ST2000DM008-2FR102

├─sdd1 2G

└─sdd2 1.8T

sde 1.8T ST2000DM008-2FR102

├─sde1 2G

└─sde2 1.8T

sdf 232.9G Samsung SSD 860 EVO 250GB

└─sdf1 232.9G

zd0 100G

root@truenas[~]#

In the meantime, is there no other way to solve it by securing the pool? I’m left with only one disk and I’m in a dangerous situation

winnielinnie · November 24, 2025, 6:53pm

That’s peculiar. The 216.6GB partition being created on sdc (replacement disk) is the same as the OS partition on sdb (your boot drive).

@HoneyBadger is this a known bug?

Every time he tries to replace his missing disk in the 4TB mirror vdev, TrueNAS creates a 216.6GB partition, which just so happens to be the size of his OS partition on his boot disk.

HoneyBadger · November 24, 2025, 6:59pm

That is indeed something that bears investigation as it seems a bit too odd to be pure happenstance.

Questions.

Have you at any point rebooted the system? (If the answer is “no” - don’t do it yet.)
What’s the storage controller in use here (motherboard model if a direct motherboard connection, SAS/HBA controller if it’s an add-on card)

It almost seems like the middleware is querying the wrong disk to figure out what partition layout it should be imitating.

Collecting a debug (System → Advanced → Save Debug) and submitting a bug report from the “Smiley Face” icon in the top is a good thing to do here.

But in regards to fixing this, we have a couple options:

If you haven’t rebooted, we attempt that to see if it’s a case of a non-hotswap controller getting puzzled by this plug-pull.
If you have, or that doesn’t solve it, we manually bake the partitions from the CLI.

Either way, this feels like something I want a fix-it ticket for.

nico_lucas · November 24, 2025, 8:12pm

Hi, thanks for the help you’re giving me.

no, I haven’t tried the reboot yet
I’m waiting to upgrade to a server motherboard. However, due to my limited budget, I can’t do so at the moment.

My mb is an Asus H81 Gamer, and the offending pool 4TB is connected to the MB SATA ports directly.

I have a PCI card for two additional SATA ports, which also has an older separate 2TB pool connected to it.

I have created the ticket bug: NAS-138678

ok, for the reboot, I’ll try now and update you, for manual bake the partitions I’ll ask you for a help because I’ve never done it, but I always like to learn new experiences on TN.

nico_lucas · November 24, 2025, 8:30pm

@HoneyBadger good notice: after reboot system, the new disk as entered into status “faulted” in red colors , and after click on “replace” , the process of wipe is started and the disk are now recognized and the resilvering process is beginning. I’m waiting for the resilvering to finish to give you the result.

HoneyBadger · November 24, 2025, 9:07pm

Seems like there might be something in your motherboard that you’d need to toggle in order to enable hotplug properly.

Looking at the manual for that board it seems to point to a PCH Storage Configuration menu in your BIOS, which you should validate is set to AHCI mode (if it isn’t, don’t change it!) and then enabling SATA Hot Plug support.

nico_lucas · November 25, 2025, 10:21am

HI, the resilver is now completed, the situation is:

root@truenas[~]# zpool status -v
pool: NAS 4TB
state: ONLINE
scan: resilvered 2.69T in 13:47:27 with 0 errors on Tue Nov 25 11:10:09 2025
config:

    NAME                                      STATE     READ WRITE CKSUM
    NAS 4TB                                   ONLINE       0     0     0
      mirror-0                                ONLINE       0     0     0
        sdb2                                  ONLINE       0     0     0
        bab65f19-aee3-405e-82d0-a5d76afe614b  ONLINE       0     0     0
    cache
      6ed29a39-6506-42fc-9de9-981b16dc6160    ONLINE       0     0     0

errors: No known data errors

It looks fine, what do you think?

nico_lucas · November 25, 2025, 10:25am

root@truenas[~]# lsblk -o NAME,SIZE,MODEL

NAME SIZE MODEL

sda 3.6T WDC WD40EFPX-68C6CN0

└─sda1 3.6T

sdb 3.6T WDC WD40EFAX-68JH4N1

├─sdb1 2G

└─sdb2 3.6T

sdc 1.8T ST2000DM008-2FR102

├─sdc1 2G

└─sdc2 1.8T

sdd 1.8T ST2000DM008-2FR102

├─sdd1 2G

└─sdd2 1.8T

sde 232.9G Samsung SSD 850 EVO 250GB

├─sde1 260M

├─sde2 216.6G

└─sde3 16G

sdf 232.9G Samsung SSD 860 EVO 250GB

└─sdf1 232.9G

zd0 100G

root@truenas[~]#