SSD space usage

[Fifou2202]

Hello.

I installed a Debian with Umbrel on a 1TB SSD.
When I log into the umbrel shell and run the "df" command it tells me that I am at 35% disk space used.
On the other hand in the dashboard of the SSD in proxmox I am at 81%.

How is it possible. I just have this installed on this SSD.

Thanks.

 

[Dunuin]

You should tell us more about your storage (for example output of cat /etc/pve/storage.cfg, pvesm status and lsblk). And that Debian with 1TB SSD is a VM? How do you get that SSD into the VM or are you using a virtual disk?

 

[LnxBil]

Do you have the SSD mounted on the host and on the guest?

 

[Fifou2202]

The command cat :

root@pve:~# cat /etc/pve/storage.cfg
dir: local
path /var/lib/vz
content backup,vztmpl,iso

lvmthin: local-lvm
thinpool data
vgname pve
content images,rootdir

dir: SSD_USB
path /mnt/pve/SSD_USB
content backup,snippets,images,vztmpl,rootdir,iso
is_mountpoint 1
nodes pve

 

[Fifou2202]

The command pvesm status :

Name Type Status Total Used Available %
SSD_USB dir active 960302804 741198340 170250004 77.18%
local dir active 28465204 6699648 20294276 23.54%
local-lvm lvmthin active 65708032 4914960 60793071 7.48%

 

[Fifou2202]

The command lsblk :

NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 111.8G 0 disk
├─sda1 8:1 0 1007K 0 part
├─sda2 8:2 0 512M 0 part
└─sda3 8:3 0 111.3G 0 part
├─pve-swap 253:0 0 5G 0 lvm [SWAP]
├─pve-root 253:1 0 27.8G 0 lvm /
├─pve-data_tmeta 253:2 0 1G 0 lvm
│ └─pve-data-tpool 253:4 0 62.7G 0 lvm
│ ├─pve-data 253:5 0 62.7G 1 lvm
│ └─pve-vm--101--disk--0 253:6 0 10G 0 lvm
└─pve-data_tdata 253:3 0 62.7G 0 lvm
└─pve-data-tpool 253:4 0 62.7G 0 lvm
├─pve-data 253:5 0 62.7G 1 lvm
└─pve-vm--101--disk--0 253:6 0 10G 0 lvm
sdb 8:16 0 931.5G 0 disk
└─sdb1 8:17 0 931.5G 0 part /mnt/pve/SSD_USB

 

[Fifou2202]

I have a 120 GB SATA SSD for Proxmox and a Debian container. And a second 1TB SSD in USB.

 

[Dunuin]

Ok, so looks like you are talking about your "SSD_USB" that is just used as a directory storage with virtual qcow2 disks on it?
So 77% or 81% is the usage of the filesystem on that physical SSD. Then you probably got a 960GB virtual disk stored on that and this is 35% used.
I'm no qcow2 expert, but I guess you always get overhead because it is copy-on-write and can use snapshots. So I wouldn't wonder when writing 300GB of data inside the virtual disk would result in a 650GB file on the physical disks filesystem. Or maye you also put other stuff on that SSD like backups, other virtual disks and so on.

Whats the output of find /mnt/pve/SSD_USB/ -type f -exec du -Sh {} + | sort -rh | head -n 20 which will show the 20 biggest files on that SSD?

 

[Fifou2202]

root@pve:~# find /mnt/pve/SSD_USB/ -type f -exec du -Sh {} + | sort -rh | head -n 20
708G /mnt/pve/SSD_USB/images/100/vm-100-disk-0.qcow2

 

[Fifou2202]

No other stuff in this storage

 

[Dunuin]

What filesystem is the SSD formated with? Did you create any snapshots? Is is the complete discard/TRIM chain from guest OS to physical SSD working?

 

[Fifou2202]

SSD filesystem is ext4. No snapshots. How can I check this: Does the full strip/TRIM chain from guest OS to physical SSD work?

 

[Dunuin]

You need to check in the guest OS if it mounts the filesystem with options like "discard" or got a cron/systemd timer that runs fstrim -a from time to time. You also need to use a protocol like SCSI that supports the TRIM command. And "discard" must be enabled in PVE webUI for the virtual disk. And your hardware like the disk controller needs to support the TRIM command (some HW raid controllers might block it).