Hello, I have had an issue with my docker containers lately. The service is failing to start. I've tried scrubbing the cache pool, balancing it and deleting the vdisk file. Still no luck. I think maybe one of my SSDs is bad. I bought a new 1TB drive to replace the 500GBs. How can I upgrade the cache pool and fix the issues? Any help would be appreciated. Thanks!

Diagnostics attached!

tower-diagnostics-20251007-1640.zip

Docker image is corrupt, you will need to recreate it, but I also that the pool is missing a device, is this expected?

It is not expected. I have unplugged and replugged the cache drives. I will report the diagnostics after the system reboots.

Attached

tower-diagnostics-20251008-1448.zip

Pool is still only using a device, correction will need to be done manully:

             Data      Data     Metadata Metadata  System                              
Id Path      single    RAID1    single   RAID1     single   Unallocated Total     Slack
-- --------- --------- -------- -------- --------- -------- ----------- --------- -----
 1 /dev/sdb1 141.00GiB 43.00GiB  1.00GiB   2.00GiB 32.00MiB   278.73GiB 465.76GiB     -
 2 missing           - 43.00GiB        -   2.00GiB        -   402.13GiB 447.13GiB     -
-- --------- --------- -------- -------- --------- -------- ----------- --------- -----
   Total     141.00GiB 43.00GiB  1.00GiB   2.00GiB 32.00MiB   680.86GiB 912.89GiB 0.00B
   Used      139.40GiB 38.15GiB 19.44MiB 242.66MiB 48.00KiB 

With the array running type:

btrfs balance start -dconvert=single,soft -mconvert=single,soft /mnt/cache

When that's done type

btrfs dev remove missing /mnt/cache

Then post new diags

OK, should I do this before or after I install the new 1TB cache drive?

Before

I opened a terminal and ran the command. I got an error

"root@Tower:~# btrfs balance start -dconvert=single,soft -mconvert=single,soft /mnt/cache

ERROR: error during balancing '/mnt/cache': Invalid argument

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

tower-diagnostics-20251010-1159.zip

Sorry, forgot the force option, use:

btrfs balance start -f -dconvert=single,soft -mconvert=single,soft /mnt/cache

7 minutes ago, JorgeB said:

Sorry, forgot the force option, use:

btrfs balance start -f -dconvert=single,soft -mconvert=single,soft /mnt/cache

I ran the command in the terminal and got another error. Not sure if I am doing something wrong or not

"root@Tower:~# btrfs balance start -f -dconvert=single,soft -mconvert=single,soft /mnt/cache

ERROR: error during balancing '/mnt/cache': Structure needs cleaning

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

That suggests a file system issue; post new diags since there should be more info there.

OK thanks. I have attached the diagnostics.

tower-diagnostics-20251010-1434.zip

Yep:

Oct 10 13:51:30 Tower kernel: BTRFS critical (device sdb1): corrupted leaf, root=7 block=2775787454464 owner mismatch, have 10 expect 7

In this case, recommend backing up the pool, reformatting, and then restoring the data.

1 hour ago, JorgeB said:

Yep:

Oct 10 13:51:30 Tower kernel: BTRFS critical (device sdb1): corrupted leaf, root=7 block=2775787454464 owner mismatch, have 10 expect 7

In this case, recommend backing up the pool, reformatting, and then restoring the data.

OK, I am trying to move it from the cache pool to the array and it says that there is an invalid target. Im not sure what I am doing wrong. Thanks

How, using the mover?

Using the built in tool in unraid shares. This is the error that I get.

I also tried to use CA backup/ restore, but it is telling me that there are no docker containers to backup.

On 10/10/2025 at 10:46 PM, PrisonMike said:

Using the built in tool in unraid shares. This is the error that I get.

You cannot copy from a disk share to a user share, only disk to disk or share to share.

OK, can you give me some guidance on how to move/ copy the data to the array? Thanks!

You can use the mover:

https://docs.unraid.net/unraid-os/using-unraid-to/manage-storage/cache-pools/#backing-up-your-cache-pool-to-the-array

OK, looks like I got everything. Plex appdata was corrupt unfortunately. So were some other appdata folders.

I changed my cache pool to ZFS raid1. So in the future, would I just be able to swap out a drive like the array and not lose data?

If it's a device problem, yes, note that it won't protect against file system corruption or other events, so you should always have a backup of anything important.

Got it. It seems like this issue happens about every 12-18 months. Is there anything I can do to prevent it?

If you mean the Docker image getting corrupt, it's usually the result of an underlying issue, e.g., a dropping device or another problem.