Pool import full/empty

TrueNAS General
1

My boot disk failed - it wasn’t redundant. (cost choices were made that weren’t the best)
New more reliable disk installed, comes up, lets me import pools, the smaller NVME SSD imports perfectly, my bigger data HDD pool imports without error. But it doesn’t work right.

It’s 2 8GB (or I guess GiB nowadays) HDDs in mirror mode. All of the previously set up datadets show up in the Datasets part of the TrueNAS web UI. But in the Usage section of TrueNAS’s Storage Dashboard it shows 100% used - 7.14 TiB used, 0B available.
If I go to the Shell (under System) use the Linux ‘df’ command I get:
Filesystem 1K-blocks Used Available Use% Mounted on
EpsilonMachine 128 128 0 100% /mnt/EpsilonMachine

but going to /mnt/EpsilonMachine in the shell is empty.
I’m new to TrueNAS - using it after a prior SSD failure - any ideas/assistance would be greatly appreciated

2

Keep a current download of your system configuration file. You can do a fresh install of TrueNAS on a new boot device, upload your configuration and be back to normal. Mirrored boot isn’t, really required.

The version of TrueNAS and a good description of your hardware and pool layouts could help. Guessing boot disk, NVME for Apps and a mirror VDEV of 2, 8TB hard disks?

zfs list -r -t filesystem,volume -o space <poolname>
that should show you what is using space. Replace ‘poolname’ with the name of your mirror pool

3

TrueNAS 25.04.2.6 community
It’s a Terramaster F2-423
Intel N5095 16 GB RAM
the two Main bays are the data pool with 2x8TB Seagate Ironwolfs
one NVME is 1 TB apps
OS was on the internal USB port (no longer booted, I admit, poor choice), now on a 128 GB NVME.

in the shell the command suggested gives: zsh: command not found: zfs

4

Are you in the GUI Shell in TrueNAS? It’s System then Shell. It just ran fine on 25.10 here

5

OK the zfs command worked with sudo
and it shows all my datasets:

xx_admin@truenas[/mnt/Holodeck]$ sudo zfs list -r -t filesystem,volume -o space EpsilonMachine
NAME AVAIL USED USEDSNAP USEDDS USEDREFRESERV USEDCHILD
EpsilonMachine 0B 7.14T 0B 104K 0B 7.14T
EpsilonMachine/.system 0B 1.89G 0B 1.56G 0B 337M
EpsilonMachine/.system/configs-ae32xxxxxx 0B 19.1M 0B 19.1M 0B 0B
EpsilonMachine/.system/cores 0B 96K 0B 96K 0B 0B
EpsilonMachine/.system/netdata-ae32xxxxx 0B 317M 0B 317M 0B 0B
EpsilonMachine/.system/nfs 0B 132K 0B 132K 0B 0B
EpsilonMachine/.system/samba4 0B 248K 0B 248K 0B 0B
EpsilonMachine/xxxxxx 0B 14.2G 0B 14.2G 0B 96K
EpsilonMachine/xxxxxxx 0B 96K 0B 96K 0B 0B
EpsilonMachine/Shared 0B 2.44T 0B 191G 0B 2.26T
EpsilonMachine/Shared/xxxxx 0B 3.33G 0B 3.33G 0B 0B
EpsilonMachine/Shared/xxxxx 0B 2.25T 0B 10.6G 0B 2.24T
EpsilonMachine/Shared/xxxx 0B 120K 0B 120K 0B 0B
EpsilonMachine/Shared/xxxxl 0B 2.24T 0B 2.24T 0B 0B
EpsilonMachine/urbackup 0B 4.68T 0B 4.68T 0B 0B
EpsilonMachine/xxxxx 0B 104K 0B 104K 0B 0B

but running ‘df’ doesn’t show the datasets for that pool, but does for my other one

xx_admin@truenas[/mnt/Holodeck]$ df
Filesystem                                  1K-blocks    Used Available Use% Mounted on
udev                            7938432       0   7938432   0% /dev
tmpfs                           1622236    2736   1619500   1% /run
boot-pool/ROOT/25.04.2.6      117134976  180736 116954240   1% /
...[normal looking linux fs]...
boot-pool/grub                116962944    8704 116954240   1% /boot/grub
EpsilonMachine                      128     128         0 100% /mnt/EpsilonMachine
Holodeck                      944522752     128 944522624   1% /mnt/Holodeck
Holodeck/xxxxxx               944522752     128 944522624   1% /mnt/Holodeck/xxxxx
...[all the Holodeck Datasets]...
Holodeck/.system/netdata-aexx 944558208   35584 944522624   1% /var/db/system/netdata
tmpfs                           1622232       0   1622232   0% /run/user/950

and going to /mnt/EpsilonMachine and running ‘ls’ gives an empty directory, while it shows the all the dataset sub-directories for /mnt/Holodeck

6

The SMB shared folders are owned by another user? try to su USERNAME and then try df.

Try posting the command results using preformatted text mode. Hit Ctrl+e or (</>) on toolbar and then paste. It should make it a lot more readable and keep the formatting you see in the Shell window.

It just looks like you filled up your server. urbackup is 4.68T and Shared section is big too.

7

I figured out the preformatted code on the second block, I tried screenshots but it wouldn’t let me embed them here.

Yes the drive is very full - urbackup will suck up any available space with incremental backups if given the chance. I was still trying to tune that down.

’sudo df’ (ie as root) gives

Filesystem                     1K-blocks    Used Available Use% Mounted on
EpsilonMachine                       128     128         0 100% /mnt/EpsilonMachine

(with the other lines trimmed out) - everything else shows way more than 128 1K-blocks, which tells me something isn’t quite right.

8

You should have set a quota to prevent that. Due to Copy-on-Write, ZFS cannot work without free space.
Try deleting snapshots to get back some space.

If that fails (ZFS needs free space to delete and free space!), you’ll need to stripe an extra drive with your mirror to have free space. A USB drive if need be. Then delete something to recover some space and remove the extra drive (from GUI!).