Errors when moving VMs to another project

1

I have a VM in the default project, which runs fine. If I stop it and move it to another project called “Proxmox”, the VM no longer starts. It shows this error message:

Unable to locate matching firmware: [{Code:/nix/store/cxm4xcn3cgq9np2qs76ls6a4dh2rh6vq-incus-ovmf/OVMF_CODE.4MB.fd Vars:/nix/store/cxm4xcn3cgq9np2qs76ls6a4dh2rh6vq-incus-ovmf/OVMF_VARS.4MB.ms.fd} {Code:/nix/store/cxm4xcn3cgq9np2qs76ls6a4dh2rh6vq-incus-ovmf/OVMF_CODE.fd Vars:/nix/store/cxm4xcn3cgq9np2qs76ls6a4dh2rh6vq-incus-ovmf/OVMF_VARS.ms.fd}]

The strange thing is that the firmwares are available in the directory:

$ ls /nix/store/cxm4xcn3cgq9np2qs76ls6a4dh2rh6vq-incus-ovmf
OVMF_CODE.4MB.fd OVMF_CODE.fd OVMF_VARS.4MB.fd OVMF_VARS.4MB.ms.fd OVMF_VARS.fd OVMF_VARS.ms.fd seabios.bin

Now, If I try to move the VM back to default, I get the following error:

Error deleting storage volume: Failed to remove '/var/lib/incus/storage-pools/default/virtual-machines/Proxmox_Proxmox3': remove /var/lib/incus/storage-pools/default/virtual-machines/Proxmox_Proxmox3: device or resource busy

After that, the VM is visible in both projects, “default” and “Proxmox”, but only is missing the block device in “Proxmox”. I can delete the instance in the “Proxmox” project. The version in the “default” project runs fine again.

I can repeat the whole thing getting exactly the same result.

$ incus version
Client version: 6.21
Server version: 6.21

$ incus config show Proxmox3 --project default
architecture: x86_64
config:
limits.cpu: “4”
limits.memory: 6GiB
volatile.cloud-init.instance-id: f5c27a6b-6c5c-4efe-ad75-3fac7d353513
volatile.eth0.host_name: tapf91ac964
volatile.eth0.hwaddr: 10:66:6a:d5:11:d4
volatile.last_state.power: RUNNING
volatile.last_state.ready: “false”
volatile.uuid: 6df2d229-c8ed-49c5-b8a7-163052ad9274
volatile.uuid.generation: 6df2d229-c8ed-49c5-b8a7-163052ad9274
volatile.vm.definition: pc-q35-10.1
volatile.vm.rtc_adjustment: “0”
volatile.vm.rtc_offset: “-1”
volatile.vsock_id: “2055157134”
devices:
eth0:
nictype: bridged
parent: vmbr0
type: nic
vlan: “10”
vlan.tagged: 2-4094
root:
path: /
pool: default
size: 100GiB
type: disk
sda:
source: /dev/sda
type: disk
ephemeral: false
profiles:

default
stateful: false
description: “”

Any idea, what is the problem?

2

Well, for the time being I found a workaround: Exporting the instance and reimporting it into the correct project worked. Still no idea, where the error came from…

3

The firmware error you can usually reset it by toggling the value of security.secureboot as that will force the firmware files to be looked at again and symlinks be re-generated.

4

Thanks, that sounds better than exporting and reimporting. :+1: