Noob question about OVF import and PVE disk allocations (/var/tmp/) and such

[mbarnickle]

I am relatively new to Proxmox though I've worked with VMWare for decades. Before I go all in on Proxmox I'm doing some hands on testing in my home lab. I have a Beelink GTR9-Pro with 128GB RAM and a 2TB NVMe drive for storage. It should be enough for me to learn the lay of the land.

I've installed Proxmox, deployed a few containers, and successfully converted one vanilla Windows VM from VMware (stand alone, not messing with ESXi yet) to a working VM. Good enough, slowly learning as I go. But I've hit a problem around storage that I cannot figure out.

The Windows VM I converted was small, with a 10Gb HDD. I won't drag you through those details because they are only relevant to say "OVF import worked and that VM works." When I try to import a larger working VM things go awry. I've already used OVF tool to convert the image. This resulted in two vmdk files of roughly 60GB and 250GB. When I try to import the 60GB vmdk using the proxmox GUI I get the following error:

starting file import from: /var/tmp/pveupload-af39f071e5ef790606b290a45b9d32a4
target node: pve
target file: /var/lib/vz/import/ForgeVM-disk1.vmdk
file size is: 59147923456
command: cp -- /var/tmp/pveupload-af39f071e5ef790606b290a45b9d32a4 /var/lib/vz/import/ForgeVM-disk1.vmdk
TASK ERROR: import failed: cp: error copying '/var/tmp/pveupload-af39f071e5ef790606b290a45b9d32a4' to '/var/lib/vz/import/ForgeVM-disk1.vmdk': No space left on device

Before we even get into my request for you folks to lay some knowledge on me, help me understand the details I should post. You know, so you can point out "I see the issue, your head doesn't actually belong in that orifice!" and such.

Seeing my issue above, what Proxmox commands generate the correct details for you to even know what my issue is (other than ignorance)?

 

[woma]

Maybe you did that already ...

Check available space of /var/lib/vz with df -h
As my test system has only 28GB available space a 60GB import task would fail, too.

oot@node3:~# df -h
Filesystem        Size  Used Avail Use% Mounted on
...
rpool/var-lib-vz   28G  101M   28G   1% /var/lib/vz
...

My storage is too small, so I would have to allow "import" content (datacenter>storage and Edit button) for another storage that is big enough and use that for import of that 60GB VM.

If your storage is big enough but has no free space, I would check directories with find /var/lib/vz -type d -exec du -sh {} \;
Backups, ISO files, VM disks etc. could be stored there and use up space.
Maybe prior failed imports did not clean up their files. If so I would expect big amount of data in "import" directory.

Output looks like that and gives occupied space per directory, in my case everything is "empty" after fresh install.

root@node3:~# find /var/lib/vz -type d -exec du -sh {} \;
3.5K    /var/lib/vz
512     /var/lib/vz/images
512     /var/lib/vz/import
512     /var/lib/vz/dump
1.5K    /var/lib/vz/template
512     /var/lib/vz/template/cache
512     /var/lib/vz/template/iso

 

[mbarnickle]

Ok, let's see where this goes. FYI, I added a 4TB NVMe drive to this unit for VM storage. Once I get my head around Proxmox most of my running VM's will be fairly large. In any event, here are the results of the commands you mentioned.

df-h results:

root@pve:~# df -h
Filesystem            Size  Used Avail Use% Mounted on
udev                   51G     0   51G   0% /dev
tmpfs                  11G  1.7M   11G   1% /run
/dev/mapper/pve-root   94G  4.2G   85G   5% /
tmpfs                  55G   43M   55G   1% /dev/shm
efivarfs              128K   35K   89K  29% /sys/firmware/efi/efivars
tmpfs                 5.0M     0  5.0M   0% /run/lock
tmpfs                 1.0M     0  1.0M   0% /run/credentials/systemd-journald.service
tmpfs                  55G     0   55G   0% /tmp
/dev/nvme0n1p2       1022M  9.1M 1013M   1% /boot/efi
/dev/fuse             128M   16K  128M   1% /etc/pve
tmpfs                 1.0M     0  1.0M   0% /run/credentials/getty@tty1.service
tmpfs                  11G  4.0K   11G   1% /run/user/0

find /var/lib/vz -type d -exec du -sh {} \; results:

root@pve:~# find /var/lib/vz -type d -exec du -sh {} \;
28K     /var/lib/vz
4.0K    /var/lib/vz/dump
4.0K    /var/lib/vz/import
4.0K    /var/lib/vz/images
12K     /var/lib/vz/template
4.0K    /var/lib/vz/template/cache
4.0K    /var/lib/vz/template/iso

At this point I'm sure I have enough physical space to import a Windows 11 VM with two drives (100GB and 300GB). I have it exported to an OVF format. And I'm familiar with the process of creating a "new dumb" VM and then a tiny dance of replacing the disks etc. But the import, even of the smaller system disk, fails with lack of space.

 

[mbarnickle]

FYI, I tried to import an OVF just now and get the same error as before.

()
starting file import from: /var/tmp/pveupload-196fef808bfb7fd260b2a7a017d63f2d
target node: pve
target file: /var/lib/vz/import/VM-disk1.vmdk
file size is: 59147923456
command: cp -- /var/tmp/pveupload-196fef808bfb7fd260b2a7a017d63f2d /var/lib/vz/import/VM-disk1.vmdk
TASK ERROR: import failed: cp: error copying '/var/tmp/pveupload-196fef808bfb7fd260b2a7a017d63f2d' to '/var/lib/vz/import/VM-disk1.vmdk': No space left on device

I feel like I'm getting close to asking my real question in some intelligent manner. In any event, what now Proxmox community? How can I make the full capacity of my local disk available for imports? Those imported disks will be immediately moved to local-lvm for actual use. Honestly, I don't understand the need for "local" to be a temporary import mechanism. Why not just import the disk to it's long-term target? But I'm hardly knowledgeable enough to question Proxmox architecture yet. Anyhoo, how to get things worked out for my upcoming 500GB+ disks to be imported?