[solved] Install error looking for libvirt/boot

[sarnold]

After adding the virsh session connect args it coughs the above install error with both new and existing VM templates and I have no idea where "boot" is coming from. The storage setup matches the README:

$ virsh --connect qemu:///system vol-list vm-pool
 Name                     Path
------------------------------------------------------------------------------
 source-rocky10.img       /var/lib/libvirt/images/vms/source-rocky10.img
 source-ubuntu26.04.img   /var/lib/libvirt/images/vms/source-ubuntu26.04.img
 vm-Rocky10VM.qcow2       /var/lib/libvirt/images/vms/vm-Rocky10VM.qcow2
 vm-Ubuntu26VM.qcow2      /var/lib/libvirt/images/vms/vm-Ubuntu26VM.qcow2

Install error:

$ bin/launch-vm.sh -d ubuntu26.04 -n Ubuntu26VM -c 4 -m 4096 -s 25
Downloading cloud image to: /tmp/cloudimg-T2ZxKJ/source-ubuntu26.04.img
--2026-04-10 12:15:13--  https://cloud-images.ubuntu.com/resolute/current/resolute-server-cloudimg-amd64.img
Resolving cloud-images.ubuntu.com... 185.125.190.40, 185.125.190.37, 2620:2d:4000:1::17, ...
Connecting to cloud-images.ubuntu.com|185.125.190.40|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 868919296 (829M) [application/octet-stream]
Saving to: ‘/tmp/cloudimg-T2ZxKJ/source-ubuntu26.04.img’

/tmp/cloudimg-T2ZxK 100%[===================>] 828.67M   433KB/s    in 32m 58s 

2026-04-10 12:48:11 (429 KB/s) - ‘/tmp/cloudimg-T2ZxKJ/source-ubuntu26.04.img’ saved [868919296/868919296]

Creating volume 'source-ubuntu26.04.img' in pool 'vm-pool'...
Vol source-ubuntu26.04.img created

Vol vm-Ubuntu26VM.qcow2 cloned from source-ubuntu26.04.img

Resizing volume 'vm-Ubuntu26VM.qcow2' in pool 'vm-pool' to 25G...
Size of volume 'vm-Ubuntu26VM.qcow2' successfully changed to 25G

Starting install...
ERROR    Could not start storage pool: cannot open directory '/var/lib/libvirt/boot': No such file or directory
Domain installation does not appear to have been successful.
If it was, you can restart your domain by running:
  virsh --connect qemu:///system start Ubuntu26VM
otherwise, please restart your installation.

[sarnold]

Solution is just create the directory with normal root perms:

# mkdir /var/lib/libvirt/boot
# ls -l /var/lib/libvirt/
total 16
drwxr-xr-x 2 root root 4096 Apr 11 15:48 boot
drwxr-xr-x 2 root root 4096 Apr 11 11:49 dnsmasq
drwx--x--x 3 root root 4096 Apr 10 19:00 images
drwxr-xr-x 8 qemu qemu 4096 Apr  2 22:27 qemu

BUT I still have no idea why it's necessary at all. After running the launch-vm.sh script, it shows up in virt-manager storage pool details with the pool name boot-scratch

[sarnold]

I also added system URI with --connect to the virsh and virt-install commands and everything worked except the vm-pool directory needed qemu group read perms:

# ls -l /var/lib/libvirt/images/
total 4
drwxr-x--x 2 root qemu 4096 Apr 11 17:34 vms

[bkram]

Nice you managed to solve it

[sarnold]

But wait, there's more... I gave up on the everything user-session and instead I enabled uri_default = "qemu:///system" after copying the system libvirt.conf to the XDG location: ~/.config/libvirt/libvirt.conf (also reverted most of my changes). I got VM scripts working except the cloud-init and finally figured out I was struggling with this: virt-manager/virt-manager#975 .

I ended up swapping passwd bits for ssh key and switched the "cloud" user to my own user and added - default plus default key handling, where the default user is mostly the short distro name, ie, rocky, ubuntu, debian, etc.

User changes:

users:
  - default
  - name: myuser
    sudo: ALL=(ALL) NOPASSWD:ALL
    shell: /bin/bash
    groups: [sudo, wheel, users]
    lock_passwd: False
    ssh_authorized_keys:
      - ssh-ed25519 ...MxAMQ...

ssh_genkeytypes: ['ed25519', 'rsa']
ssh_authorized_keys:
  - ssh-ed25519 ...MxAMQ...