Here is a pic of my Main screen:

Then I turned my array off and unassigned both of my Cache-Nvme pool devices and now it won't even let me start the array. Now it says this error:

I've tried doing several things mentioned online including a rescue command which seemed to partially work but I don't fully understand it. I still cannot get it to work. I want to keep all my data if possible and am entirely unsure if this is a bug or if my drives are failing (I doubt both failed at the same time since they were mirrored). I want to prevent any data loss if possible so I have avoided reformatting. Here are my diagnostics:

oracle-diagnostics-20260213-1228.zip

Edit: I was able to do a New Config and start the array like this so it seems like my data is intact. Docker still fails to start though. I do not know how to fix my other drive. After stopping the array I used this command but again am not too sure if i did this properly btrfs rescue zero-log /dev/nvme0n1p1

Edited February 13Feb 13 by dinosaurthug

27 minutes ago, dinosaurthug said:

do a New Config and start the array

Post new diagnostics

Here are new diagnostics, friend. Thank you!

oracle-diagnostics-20260213-1302.zip

You will need to use the CLI to correct the pool, with the array running type:

btrfs balance start -f -dconvert=single -mconvert=dup /mnt/cache-nvme

when that finishes, type

btrfs dev remove missing /mnt/cache-nvme

Then stop the array and reimport the pool with the single device:

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

start the array to import the pool

If it fails to import, post new diags, you may need to wipe/disocnnect the other device; if it imports, you can stop the array and add the old deivice to create a new mirror.

8 minutes ago, JorgeB said:

You will need to use the CLI to correct the pool, with the array running type:

btrfs balance start -f -dconvert=single -mconvert=dup /mnt/cache-nvme

when that finishes, type

btrfs dev remove missing /mnt/cache-nvme

Then stop the array and reimport the pool with the single device:

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

start the array to import the pool

If it fails to import, post new diags, you may need to wipe/disocnnect the other device; if it imports, you can stop the array and add the old deivice to create a new mirror.

Doing the first command gave me this error. I did not try the second command since this gave an error. I am trying to be extra cautious. How should I proceed? Thank you very much.

root@oracle:~#

btrfs balance start -f -dconvert=single -mconvert=dup /mnt/cache-nvme

WARNING:

Conversion with missing device(s) can be dangerous.

Please use 'btrfs replace' or 'btrfs device remove' instead.

Safety timeout skipped due to --force

ERROR: error during balancing '/mnt/cache-nvme': Read-only file system

There may be more info in syslog - try dmesg | tail

root@oracle:~#

Edited February 13Feb 13 by dinosaurthug

Here are more updated diagnostics. I am lost. Would appreciate any help! Will donate if you successfully help!

oracle-diagnostics-20260213-2134.zip

filesystem has a bad log tree; we can try to fix that, but before doing so, run memtest, since btrfs was detecting data corruption, and this is typically bad RAM:

Feb 13 21:27:56 oracle kernel: BTRFS info (device nvme0n1p1): bdev <missing disk> errs: wr 0, rd 0, flush 0, corrupt 80, gen 0

Feb 13 21:27:56 oracle kernel: BTRFS info (device nvme0n1p1): bdev /dev/nvme0n1p1 errs: wr 0, rd 0, flush 0, corrupt 85, gen 0