Hello Team,

I have my Cache running on RAID1, 2x1TB NVME disks and needed to updated to 2x2TB NVME (due to being 100% full...). I stopped the array, powered down the server and replaced ONE of the RAID1 disks. I powered up the server and in the missing Cache disk, I selected the new 2TB NVME disk. Starting the array shows the new disk fine in the Main-tab:

BUT strangely, the NEW disk does NOT show up in the Cache Settings:

So, neither the RAID1 rebuild has not started automatically nor I am able to manually trigger it by selecting "Convert to raid1 mode".

What am I missing here about the procedure to replace ONE (1) RAID1 drive and rebuild the array?

Thank you.

Edited September 26, 2025Sep 26 by G3orgios

Please post the diagnostics.

Diagnostics as requested:
eos-diagnostics-20250926-1238.zip

Thank you.

The device failed to be added to the pool:

Sep 26 10:45:44 Eos kernel: BTRFS info (device nvme0n1p1): dev_replace from <missing disk> (devid 2) to /dev/nvme1n1p1 started
Sep 26 10:45:44 Eos kernel: BTRFS warning (device nvme0n1p1): failed setting block group ro: -28

Could be a filesystem issue, but try again, first reimport the pool with just nvme0n1:

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 1 slot

assign the working pool device (nvme0n1), leave the filesystem set to auto

start the array to import the pool, it should change the pool to two slots and show the missing device

if yes, stop the array, assign the new device to the missing slot, start array, post new diags.

Following the instructions,

• Pool removed

• Pool added, same name, 1-slot, auto-FS

• Indeed, pool showed up as 2-slots

• added NEW (empty) disk on slot-2, start-array

SAME status, exactly as my initial post and images. Here are the fresh diagnostics

eos-diagnostics-20250926-1421.zip

Thank you.

Sep 26 14:14:35 Eos kernel: BTRFS warning (device nvme0n1p1): failed setting block group ro: -28

Same error, recommend backing up and reformatting the pool.

@JorgeB The existing disk (nvme0n1) was 100% full and I thought this might be an issue (I had all docker containers not starting up sometimes). Thus I removed some files to make it "operational". After this I RETRIED your procedure (remove pool, re-add as 1-slot, set the NEW disk as stand-alone and formatted BTRFS, re-insert it as slot-2) but again the same result with the NEW NVME disk (ready in BTRFS this time) not showing up within the Cache Settings (same as my initial post).

Then I took another path, I setup the NEW NVME disk as stand-alone and copied all files from the existing disk (nvme0n1) to the NEW one. Then i removed the pool, added a new pool with same name ("cache") and 1-slot, using the NEW disk (nvme0n1p1). This worked and it seems that the dockers and apps operate properly.

A couple of questions please:

• Is there something I may have missed by doing so instead of the proper RAID1 repair?

• How do you recommend adding the 2nd 2TB NVME to this "cache" pool (1-slot currently) and activating both disks in RAID1?

Thank you.

24 minutes ago, G3orgios said:

Is there something I may have missed by doing so instead of the proper RAID1 repair?

Repair in this case may not have been possible; hence why I suggested backup/reformat.

24 minutes ago, G3orgios said:

How do you recommend adding the 2nd 2TB NVME to this "cache" pool (1-slot currently) and activating both disks in RAID1?

Stop the array, change pool slots to two, add the new device; it should automatically add it as a mirror.

Apologies, I missed to post an update here.
Indeed, adding the 2nd 2TB NVMe worked and the 2nd disk has finally synced (RAID1). Still, strange to have all this in the first place, where the newly added disk should have simply replicated the 1st (operational) disk and used to rebuilt the pool. Hopefully the system will be become more "clever" in the future!

Thank you.