The size of among of data inside VM doesn't mach the total of virtual disk.

[Gilberto Ferreira]

Hello there

I have a VM here with Ubuntu Focal.
The root disk lay down in a LVM group which has the site about 1TB.
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/vgcloud-root 982G 654G 279G 71% /

But when I get the information about disk size using qemu-img, surprisely I get other results.

qemu-img info /vmsgluster/images/109/vm-109-disk-1.qcow2
image: /vmsgluster/images/109/vm-109-disk-1.qcow2
file format: qcow2
virtual size: 0.977 TiB (1073741824000 bytes)
disk size: 299 GiB
cluster_size: 65536
Format specific information:
compat: 1.1
compression type: zlib
lazy refcounts: false
refcount bits: 16
corrupt: false
extended l2: false

So I am a bit confuse.
The disk size is related to the Avail data inside the VM?

And worst, when I checked the storage in Proxmox I get this:
Filesystem Size Used Avail Use% Mounted on
gluster2:VMSGLUSTER 5.9T 327G 5.6T 6% /vmsgluster

If I list the storage I get:
pvesm list VMS
Volid Format Type Size VMID
VMS:100/vm-100-disk-0.qcow2 qcow2 images 107374182400 100
VMS:101/vm-101-disk-0.qcow2 qcow2 images 42949672960 101
VMS:102/vm-102-disk-0.qcow2 qcow2 images 34359738368 102
VMS:102/vm-102-state-snap-13042023.raw raw images 4003732480 102
VMS:103/vm-103-disk-0.qcow2 qcow2 images 107374182400 103
VMS:103/vm-103-state-teste.raw raw images 1776717824 103
VMS:104/vm-104-disk-0.qcow2 qcow2 images 322122547200 104
VMS:105/vm-105-disk-0.qcow2 qcow2 images 21474836480 105
VMS:105/vm-105-disk-1.qcow2 qcow2 images 274877906944 105
VMS:106/vm-106-disk-0.qcow2 qcow2 images 21474836480 106
VMS:106/vm-106-disk-1.qcow2 qcow2 images 1073741824000 106
VMS:107/vm-107-disk-0.qcow2 qcow2 images 21474836480 107
VMS:107/vm-107-disk-1.qcow2 qcow2 images 64424509440 107
VMS:108/vm-108-disk-0.qcow2 qcow2 images 21474836480 108
VMS:109/vm-109-disk-1.qcow2 qcow2 images 1073741824000 109
VMS:9998/vm-9998-disk-0.qcow2 qcow2 images 21474836480 9998
VMS:9999/vm-9999-disk-0.qcow2 qcow2 images 21474836480 9999

pve02:/vmsgluster/images# du -chs *
127G 100
771G 101
68M 102
60G 103
224G 104
74G 105
29G 106
47G 107
119G 108
300G 109 ---------------> this is the VM which has the disk about 1TB (virtual)
197M 9998
321M 9999
1.8T total


Sorry for this long thread!
This is not a issue but I found myself a bit confuse about this.

 

[Richard]

Hello there

I have a VM here with Ubuntu Focal.
The root disk lay down in a LVM group which has the site about 1TB.
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/vgcloud-root 982G 654G 279G 71% /

Assuming this is from inside Ubuntu I guess the majority of the (virtual, but seen inside Ubuntu as physical) disk is used as lvm-group.
"Used" means occupied by files as the filesystem on it reports.

But when I get the information about disk size using qemu-img, surprisely I get other results.

qemu-img info /vmsgluster/images/109/vm-109-disk-1.qcow2
image: /vmsgluster/images/109/vm-109-disk-1.qcow2
file format: qcow2
virtual size: 0.977 TiB (1073741824000 bytes)

Quite correct: ~ 1TB defiened virtual disk size

disk size: 299 GiB

But only 299GB space of physical disk (seen the gluster-fs as just one physical disk in this context) is really used for the qcow2 file.
Note: the value is not necessary identical to the "used" space in the VMs subsystem since qcow2 format has it's own logic how to compress data.

cluster_size: 65536
Format specific information:
compat: 1.1
compression type: zlib
lazy refcounts: false
refcount bits: 16
corrupt: false
extended l2: false

So I am a bit confuse.
The disk size is related to the Avail data inside the VM?

And worst, when I checked the storage in Proxmox I get this:
Filesystem Size Used Avail Use% Mounted on
gluster2:VMSGLUSTER 5.9T 327G 5.6T 6% /vmsgluster

If I list the storage I get:
pvesm list VMS
Volid Format Type Size VMID
VMS:100/vm-100-disk-0.qcow2 qcow2 images 107374182400 100
VMS:101/vm-101-disk-0.qcow2 qcow2 images 42949672960 101
VMS:102/vm-102-disk-0.qcow2 qcow2 images 34359738368 102
VMS:102/vm-102-state-snap-13042023.raw raw images 4003732480 102
VMS:103/vm-103-disk-0.qcow2 qcow2 images 107374182400 103
VMS:103/vm-103-state-teste.raw raw images 1776717824 103
VMS:104/vm-104-disk-0.qcow2 qcow2 images 322122547200 104
VMS:105/vm-105-disk-0.qcow2 qcow2 images 21474836480 105
VMS:105/vm-105-disk-1.qcow2 qcow2 images 274877906944 105
VMS:106/vm-106-disk-0.qcow2 qcow2 images 21474836480 106
VMS:106/vm-106-disk-1.qcow2 qcow2 images 1073741824000 106
VMS:107/vm-107-disk-0.qcow2 qcow2 images 21474836480 107
VMS:107/vm-107-disk-1.qcow2 qcow2 images 64424509440 107
VMS:108/vm-108-disk-0.qcow2 qcow2 images 21474836480 108
VMS:109/vm-109-disk-1.qcow2 qcow2 images 1073741824000 109

This shows the logical size which is ~ 1TB

VMS:9998/vm-9998-disk-0.qcow2 qcow2 images 21474836480 9998
VMS:9999/vm-9999-disk-0.qcow2 qcow2 images 21474836480 9999

pve02:/vmsgluster/images# du -chs *
127G 100
771G 101
68M 102
60G 103
224G 104
74G 105
29G 106
47G 107
119G 108
300G 109 ---------------> this is the VM which has the disk about 1TB (virtual)

Quite correct it shows the physically occupied space (299GB and 300GB can be considered as the same value)

 

[Gilberto Ferreira]

since qcow2 format has it's own logic how to compress data.

So this is all about qcow2 or raw compress?
And I am bring raw here because the disk was restored from a backup were the virtual disk was raw (from a LVM-thin in the old server).
I changed it to qcow2 in hope to make clear to understand what's going on.
But that's ok. This is not dangerous in any way.
Thanks