3 minutes ago, JorgeB said:

Should be, there were 71T free on the last screenshot; they would need to be removed one at a time, and it will take some time to rewrite the data to the other pool devices.

From your earlier FAQ link:

On 4/24/2016 at 4:13 AM, JorgeB said:

Remove a device:

-to remove a device from a pool type (assuming there's enough free space):

btrfs dev del /dev/sdX1 /mnt/disks/yourpoolpath 

     Replace X with correct identifier, note the 1 in the end

Do we have to put the moker-vet pool into UD to do this?

Or could we work directly with the moker-vet pool like this?

btrfs dev del /dev/sdX1 /mnt/mover-vet

@JorgeB

Or can we do this?

https://docs.unraid.net/unraid-os/using-unraid-to/manage-storage/cache-pools/#using-the-webgui

2 minutes ago, trurl said:

@JorgeB

Or can we do this?

https://docs.unraid.net/unraid-os/using-unraid-to/manage-storage/cache-pools/#using-the-webgui

Nevermind, I see that is only for RAID1

So what about

17 minutes ago, trurl said:

could we work directly with the moker-vet pool like this?

btrfs dev del /dev/sdX1 /mnt/mover-vet

Docs have similar command line and even mention removing multiple. But not clear whether the docs mean only raid1 allows device removal.

https://docs.unraid.net/unraid-os/using-unraid-to/manage-storage/cache-pools/#using-the-command-line-advanced

50 minutes ago, trurl said:

Or could we work directly with the moker-vet pool like this?

btrfs dev del /dev/sdX1 /mnt/mover-vet

You can do this, GUI doesn't support removing devices from raid5 pools, only raid1.

Once the first device is removed, you can start the 2nd and then the 3rd. When all 3 have been removed, the pool must be reimported:

on main click on the first device for that pool and then "remove pool"

back on main, create a new pool with the same name and the new number of slots

assign the remaining pool devices, leave the filesystem set to auto

start the array to import the pool

1 hour ago, trurl said:

@Moker-Vet

Do you have backups of any of this data?

I think we have a plan if you are ready to proceed.

i am ready to proceed, altough i dont have a backup as of right now. its to much data to do a backup for a homeserver. altough i would prefer to limit the downtime as much as possible.

10 minutes ago, Moker-Vet said:

limit the downtime as much as possible

I expect this to take several steps, each taking many hours.

if i understand it correctly, the actuall transfer of data should not give actual downtime, but the changing of disks might give a few minutes but that i can do on times that i can have downtime. atleast if i have understood it correctly.

Post new diagnostics just so we can confirm the drive 'letters' before we issue command lines that use them.

3 hours ago, Moker-Vet said:

if i understand it correctly, the actuall transfer of data should not give actual downtime, but the changing of disks might give a few minutes but that i can do on times that i can have downtime. atleast if i have understood it correctly.

@JorgeB

If we do this

On 2/3/2026 at 10:41 AM, trurl said:

btrfs dev del /dev/sdX1 /mnt/mover-vet

then this

On 2/3/2026 at 11:38 AM, JorgeB said:

Once the first device is removed, you can start the 2nd and then the 3rd. When all 3 have been removed, the pool must be reimported

Will pool data be accessible throughout? Or will there be significant period with no access to pool data?

4 hours ago, Moker-Vet said:

prefer to limit the downtime

@Moker-Vet

What about time with decreased performance? Such as might be expected from running a parity check, for example.

Hi Trurl,

decreased performance is not an issue as the performance in my case comes from the cache pool with 2x nvme drives. the hdd drives simply store stuff and i havent found issues with the read access. and i can take some measures to decrease the overall load on them.

moker-vet-diagnostics-20260205-0811.zip

7 hours ago, trurl said:

Will pool data be accessible throughout? Or will there be significant period with no access to pool data?

The pool will always be accessible; you should avoid other large i/o to not slow down much, though. It will just be inaccessible for the brief period you need to stop the array in the end to reimport the pool.

I just tried the View Summary at the top of this thread and it seemed to do a pretty good job.

moker-vet pool is already part of user shares. And it's btrfs raid5. Not sure how well it is configured for that. carfax has

So about 20G less than "raw disk space". I don't fully understand the other part though. What are these regions about and does that affect function, capacity, performance?

I assume btrfs raid5 is going to have some performance benefits over the Unraid array. Is it worth the extra complication?

This whole setup happened sort of by accident:

On 1/30/2026 at 9:06 AM, Moker-Vet said:

created the moker-vet pool out of neccasairity because my 20tb hdd where shrinked to 10tb when added to the array. my intention was to add them to the array

And there is this

On 1/30/2026 at 9:06 AM, Moker-Vet said:

- the array is using my 4tb drives.

-the moker-vet pool is using my 4tb ssd's, and 14tb+ hdd's

both storage pools are used for the same purpose -> storing video files. i use mover between them to prevent them from getting full.

So additional trouble because of this complicated configuration.

Putting all those disks in the array would make things much simpler. Of course, you would end up with 22TB parity checks.

So, do we really want to do this? These are all HDDs. Do we want to work through putting all moker-vet drives into the array? Maybe leaving some of the small ones out in the end just on principle? (Each additional drive is an additional point of failure.)

What i think that would be best for my case (initially my game plan): my storage hdd in the array, and the ssd in a pool (zfs or btrfs), with 1 exception the 2 nvme performance drive i chose for to keep this sperate as this configuration is much higher on performance then the ssd would ever need to be.

doing this grants me:

5x 4tb ssd (to be decided in what file system, and i think i would go for raid 5 but open for ideas)

2x 1tb nvme ZFS raid 1

xxx HDD on the XFS array with 2x 22tb array disks as parity.

what comes to mind is that i have seen several smart errors but they never caused any real issue as of now. is this something concerning?

17 minutes ago, Moker-Vet said:

several smart errors

Which drives specifically? Last 4 characters of Serial # please

from the moker-vet pool:

1. 7NP2

2. YH5P

3. ZJ9P

4. AF94g

5. 2F94g

And in the array:

1. T9YW

2. T8VB

3. S4G9

1 hour ago, trurl said:

These are all HDDs.

Correction, the 2x1TB and 5x4TB are SSDs so won't be going to the array.

T9YW - parity

187 Reported_Uncorrect -O--CK 090 090 000 - 10

Should be OK.

T8VB - disk12

187 Reported_Uncorrect -O--CK 083 083 000 - 17

Should be OK.

S4G9 - disk13

5 Reallocated_Sector_Ct PO--CK 100 100 010 - 25

187 Reported_Uncorrect -O--CK 001 001 000 - 149

Might be OK.

You should do extended SMART self-test on each of those array drives.

All of the moker-vet drives just have a few UDMA_CRC_Error_Count

Click on each SMART warning (👎) on the DASHBOARD page to Acknowledge, and it will warn again if they increase.

1 minute ago, trurl said:

do extended SMART self-test on each of those array drives

You can do them simultaneously, will take a few hours but they are only 4TB.

2 minutes ago, trurl said:

You can do them simultaneously, will take a few hours but they are only 4TB.

And that shouldn't affect removing drives from moker-vet.

The first one is the one called moker-vet, sdv, 22TB, which we will reuse as parity.

@JorgeB

Does it renumber the pool drives when we remove one of them? So moker-vet2 would become moker-vet, moker-vet3 would become moker-vet2, etc.?

18 minutes ago, trurl said:

Does it renumber the pool drives when we remove one of them? So moker-vet2 would become moker-vet, moker-vet3 would become moker-vet2, etc.?

They will only change once you stop the array and reimport the pool. This can be done after each disk is removed, if you need to reuse them already, or just once all 3 have been removed.

On 2/3/2026 at 11:38 AM, JorgeB said:

Once the first device is removed, you can start the 2nd and then the 3rd.

Do we wait for each to complete before starting the next? How will we know it has completed? Just wait for the command prompt to return?