Windows Server 2022 on snap lxd 5.1 (Ubuntu 20.04)

,

I’m trying to get a lxd vm running but when i start it , it always shows Status: ERROR.

lxc info --show-log vm doesn’t show any log message, quemu is still running in background.

Also removing /var/snap/lxd/common/lxd/logs/vm doesn seem to make any difference.
There are some Ubuntu containers running fine.
I’ve also have a running kvm vm with libvirt, which i like to switch later too.

Host System

CPU: Opteron 3280 (8 core)
RAM: 16 GB
OS: Ubuntu 20.04

VM Config

architecture: x86_64
config:
  limits.memory: 4GiB
  security.secureboot: "false"
  volatile.cloud-init.instance-id: <redacted>
  volatile.eth0.hwaddr: <redacted>
  volatile.uuid: <redacted>
  volatile.vsock_id: "17"
devices:
  iso:
    boot.priority: "10"
    source: /opt/libvirt/images/en-us_windows_server_2022_updated_april_2022_x64_dvd_d428acee_virtio.iso
    type: disk
  root:
    path: /
    pool: default
    size: 60GiB
    type: disk
ephemeral: false
profiles:
- default
stateful: false
description: ""

default profile

config: {}
description: Default LXD profile
devices:
  eth0:
    name: eth0
    nictype: bridged
    parent: kvm
    type: nic
  root:
    path: /
    pool: default
    type: disk
name: default
used_by:
- /1.0/instances/vm
# some other containers

Can be the network config the issue? Do i need a different profile?

Sounds like a QEMU crash of some kind.
Can you check dmesg for anything suspicious and look at /var/snap/lxd/common/lxd/logs/NAME/ for both qemu.log and other .log files?

I did a new try now and it works now. I’ve upgraded host from ubuntu 20.04 to 22.04 and lxd is now on 5.13. maybe something was fixed. Also secure boot works.

limits.memory=4GB caused a crash when loading the windows installer. should i use GB or GiB ?

Maybe the newer windows image or newer virtio drivers made adifference too.