Hiya,

i woke up today to see that my cache was full (from downloads overnight). I think the scheduled mover didn't work as i was seeding from the cache...)

I stopped docker and tried mover which didn't move anything. I saw in the Docker settings tab that it still thought docker was running though the option was toggled to off.

I restarted the machine hoping this would then let me move data to the array and when i logged back in i get this error next to the cache drives (2x ssd - 1tb/3.8tb)

system and appdata shares are gone. appdata backup is available but not sure how recent this is as mover hasnt run. How can i get this running again please?

Pastebin

text error warn system array login Feb 4 10:20:36...

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

Btrfs is giving an out of space error, so it may not work, but try

btrfs rescue zero-log /dev/nvme1n1p1

Then restart the array and post new diags.

that does seem to have worked! the drive has now mounted and showing half full. thankyou very much!

what has that command done as the combined cache was previously showing as 100% and now down to around 50%?

do you suggest i run mover now before restarting docker?

Pastebin

Feb 4 11:00:17 spivnet emhttpd: /dev/sda1: LABEL_FATBOOT...

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

1 hour ago, stavros83 said:

what has that command done as the combined cache was previously showing as 100% and now down to around 50%?

It only clears the log, it doesn't touch the data. Please attach here the complete diagnostics.

spivnet-diagnostics-20260204-1304.zip

Pool is stilhaving space issues:

                  Data      Metadata System                                 
Id Path           RAID0     RAID1    RAID1     Unallocated Total     Slack  
-- -------------- --------- -------- --------- ----------- --------- -------
 1 /dev/nvme1n1p1 927.48GiB  4.00GiB  32.00MiB     1.02MiB 931.51GiB       -
 2 /dev/sdd1        1.66TiB  4.00GiB  32.00MiB     1.83TiB   3.49TiB 3.50KiB
-- -------------- --------- -------- --------- ----------- --------- -------
   Total            2.57TiB  4.00GiB  32.00MiB     1.83TiB   4.40TiB 3.50KiB
   Used             2.56TiB  3.98GiB 224.00KiB  

There's no space on one of the devices to create a new metadata chunk, and since it's RAID1, it requires space in both, so it's going read-only.

You can try to disable ALL writes to the pool, like stopping Docker and VM services, then restart to see if it doesn't go read-only. If it still does, you can try mounting it read-only to backup and recreate.

OK i thought id changed the whole cache to Raid0? Ive moved what data i can with mover but throughout its still shown as 2.82tb used and none available.

Docker is currently off and there are no VMs. Ill restart and see. is there a terminal command to recreate?

Metadata is always raid1, unless you use the CLI.

Command to recreate what?

"you can try mounting it read-only to backup and recreate"

just what you said. sorry, as you can tell i am very new to this!

the restart seems to have worked. the cache is now showing the free space.

Post new diagnostics

spivnet-diagnostics-20260204-1533.zipyeah i spoke too soon. mover wouldnt move any further data and when i tried turning docker on none of the containers started (error 403). Ive turned docker off again and the cache has gone back to showing no free space

You cannot make ANY writes to the pool before you try to move some data; just enabling Docker service will make some.

Try restarting the array with all the services disabled.

I disabled everything and restarted. I spivnet-diagnostics-20260204-1622.ziptried mover but it only moved a couple hundred kb. cache still says 0tb available

prior to restarting docker and everything going wrong, the mover had stopped. it still showed more than expected data on the cache but was showing free space

Your cache drives are different sizes, so raid1 mirror capacity would only be as large as the smaller drive, 1TB.

raid0 also wouldn't work to get the total capacity.

You need to use Single profile.

Copy everything off of it that you can and reformat using Single profile.

8 minutes ago, stavros83 said:

the mover had stopped

Mover probably can't help with this because Move is copy from source to destination, then delete from source. And it might not be able to delete from cache.

sorry to ask a stupid question, whats the best way to copy this off of the cache. id previously used mover to move everything to the array so i could swap the cache drive. it took a long time for quite a small amount of data compared to what is showing on the cache now.

9 minutes ago, stavros83 said:

copy this off of the cache

Use the built-in File Manager in the Unraid webUI

Pool was not showing read-only in the diags, is there anything you can try deleting before starting to move data?

If not, or it it's the same, you can mount the pool read-only and then use your favorite tool to copy the data elsewhere, like the array. To mount the pool read-only, you can use:

mkdir /temp

mount -o ro /dev/sdX1 /temp

56 minutes ago, trurl said:

so raid1 mirror capacity would only be as large as the smaller drive, 1TB.

With current kernels, Btrfs supports different size devices; once one is full, it will write just to the other, problem is the metadata; in raid1 it still requires the 2 devices, and that is what is causing the issue, so I still don't recommend using raid0 with different size devices.

sorry, ive started copying across now. Ill reformat the cache and remove the smaller ssd and then copy things back. Ill report back. thanks both for all your help so far!

Hi, i have now got everything back where i think it should be, potentially some of the hard links for some files might be broken. The only thing that Ive noticed that's different apart from plex refusing to start is that all of my shares are showing as unprotected? spivnet-diagnostics-20260205-0857.zip

Can't tell anything about filesystems without the array started in Normal mode.

Do that and post new diagnostics.

spivnet-diagnostics-20260205-2157.zip

I deleted and reinstalled the plex container and restarted. that seems to have sorted that out but the shares still have exclam marks.

Since your cache has no redundancy, any files on cache are not protected.

4TB cache should be plenty for caching, and cached files will be moved to the array anyway.

But docker/VM related shares should be on cache or other pool so they will perform better and so array disks can spin down.

You have an unassigned 1TB nvme. If you could put another 1TB nvme with it, you could make a good mirrored pool for those shares and they would be protected.