Fresh installed EndeavourOS and, after all the updates, I wanted to create a VM, but Boxes crashes as soon as I select a ISO file (I downloaded the latest Debian, Mint and I also tried the same ISO used to install the host).
For what I found in journalctl I think I may have multiple issues:
nov 10 07:08:00 wintermute virtqemud[83116]: libvirt version: 10.9.0
nov 10 07:08:00 wintermute virtqemud[83116]: hostname: wintermute
nov 10 07:08:00 wintermute virtqemud[83116]: operation failed: Unable to find 'efi' firmware that is compatible with the current configuration
nov 10 07:12:12 wintermute systemd[1180]: Started dbus-:1.2-org.gnome.Boxes@6.service.
nov 10 07:12:13 wintermute gnome-boxes[83571]: Theme directory 32/emblems of theme Eleven-Dark has no size field
nov 10 07:12:13 wintermute gnome-boxes[83571]: Theme directory 32@2x/emblems of theme Eleven-Dark has no size field
nov 10 07:12:13 wintermute gnome-boxes[83571]: Theme directory 48/emblems of theme Eleven-Dark has no size field
nov 10 07:12:13 wintermute gnome-boxes[83571]: Theme directory 48@2x/emblems of theme Eleven-Dark has no size field
nov 10 07:12:13 wintermute gnome-boxes[83571]: Using GtkSettings:gtk-application-prefer-dark-theme together with HdyStyleManager is unsupported. Please use HdyStyleManager:color-scheme instead.
nov 10 07:12:13 wintermute virtqemud[83597]: libvirt version: 10.9.0
nov 10 07:12:13 wintermute virtqemud[83597]: hostname: wintermute
nov 10 07:12:13 wintermute virtqemud[83597]: operation failed: Unable to find 'efi' firmware that is compatible with the current configuration
nov 10 07:12:13 wintermute virtstoraged[83626]: libvirt version: 10.9.0
nov 10 07:12:13 wintermute virtstoraged[83626]: hostname: wintermute
nov 10 07:12:13 wintermute gnome-boxes[83571]: Error setting up default broker: Failed to set autostart flag on storage pool: internal error: pool has no config file
nov 10 07:12:13 wintermute virtstoraged[83626]: internal error: pool has no config file
nov 10 07:12:13 wintermute virtqemud[83597]: internal error: pool has no config file
nov 10 07:12:13 wintermute gnome-boxes[83571]: (src/gnome-boxes.p/app.c:3106):boxes_app_setup_default_source_co: runtime check failed: (default_connection != null)
nov 10 07:12:25 wintermute gnome-boxes[83571]: gvir_connection_get_node_info: assertion 'GVIR_IS_CONNECTION(conn)' failed
nov 10 07:12:25 wintermute kernel: gnome-boxes[83571]: segfault at 20 ip 000055f5f68940a0 sp 00007ffd52a9b7c0 error 4 in gnome-boxes[5d0a0,55f5f684f000+7f000] likely on CPU 24 (core 8, socket 0)
nov 10 07:12:25 wintermute systemd-coredump[83674]: Process 83571 (gnome-boxes) of user 1000 terminated abnormally with signal 11/SEGV, processing...
nov 10 07:12:25 wintermute systemd[1]: Started Process Core Dump (PID 83674/UID 0).
nov 10 07:12:27 wintermute systemd-coredump[83675]: Process 83571 (gnome-boxes) of user 1000 dumped core.
nov 10 07:12:27 wintermute systemd[1]: systemd-coredump@3-83674-0.service: Deactivated successfully.
nov 10 07:12:27 wintermute systemd[1]: systemd-coredump@3-83674-0.service: Consumed 1.606s CPU time, 1.4G memory peak.
nov 10 07:12:27 wintermute systemd[1180]: dbus-:1.2-org.gnome.Boxes@6.service: Main process exited, code=dumped, status=11/SEGV
nov 10 07:12:27 wintermute systemd[1180]: dbus-:1.2-org.gnome.Boxes@6.service: Failed with result 'core-dump'.
nov 10 07:12:27 wintermute systemd[1180]: dbus-:1.2-org.gnome.Boxes@6.service: Unit process 83597 (virtqemud) remains running after unit stopped.
nov 10 07:12:27 wintermute systemd[1180]: dbus-:1.2-org.gnome.Boxes@6.service: Unit process 83626 (virtstoraged) remains running after unit stopped.
nov 10 07:12:27 wintermute systemd[1180]: dbus-:1.2-org.gnome.Boxes@6.service: Consumed 2.609s CPU time, 152.4M memory peak.
Not sure what to do, I was able to run gnome boxes in Mint 22 on the same hardware when I was choosing the new distro, so it must be a software issue.
Qemu, the underlying software that Gnome Boxes uses, was updated very recently. It is likely the culprit here. Maybe Gnome Boxes didn’t put out an updated to match? Maybe the Qemu packages have a bug? Could be a few reasons.
I just checked Virt-Manager, as it also uses Qemu. No issues.
That said, please run the following in a terminal and post it in a reply.
Thanks for the reply, in the meanwhile I reinstalled boxes and I was able to create a VM (plain Debian 12 with Cinnamon). It’s kinda working, but when I reboot/shutdown and restartthe guest it will crash after the debian boot menu multiple times.
Everything seems fine, except Wayland… Have you tested this in X11?
Even if you don’t want to use X11 for whatever reason, it’s good to test it to figure out if an issue is app-related or display compositor-related.
If it does work in X11, you could also briefly and quickly test it in Hyprland/Sway then uninstall it afterward. This will let you know if it’s a combination of KDE and Wayland or just Wayland that’s causing the issue.
That said, Boxes has not been updated for a while now (Sep. 19), whereas, Qemu was updated recently (Nov. 7). It could just be this.
It’s been running on Wayland only so far, but not for personal preference: it’s just the EndeavourOS default. Would be running KDE on X11 better or worst somehow?
I tried and it works the same, including a way bigger issue that I noticed few hours ago: when I suspend the host it takes a minute to actually suspend, and when I wake it up all peripherals are dead (monitors, keyboard and mouse) and it needs a hard reset.
This only happens if I run a VM, either on wayland or x11.
If I don’t launch Boxes at all I can suspend (and it takes half a second) and wake up the host with no issues how many times I want, otherwise I must reboot or shutdown.
I think that I need to try to downgrade Qemu at this point, since I had none of these problems on Mint 22.
Downgrading is only a temporary solution, especially sine downgrading can cause issue with other related packages. Maybe you should try virt-manager instead?
More tedious to set up than Boxes, but gives you more control, and like I said, I don’t have the issue you are having right now, even though it hasn’t been updated since April.
I indeed started using virt-manager/viewer after your last reply and it seems quite better.
With boxes I also had the VM frozen multiple times while opening some projects in VSCode, but no such issue in virt-viewer so far.
After work I’ll try to suspend and wake up the host a couple of times. I hope the problem is just gnome boxes.
Just tried suspend in eos-cinnamon using virt-manager and it locked up tight. Never tried Gnome Boxes. I’m not a user of suspend so it’s not a show stopper for me.
Well, I could run VMs AND suspend&resume the host with Mint 22, until 2 days ago when I installed EndeavourOS. Having been busy with the reinstallation and the other Qemu/Boxes problems, now I’m not so sure that suspension has ever worked on EOS.
But it’s quite important to me. I may adapt to use hibernation instead, but I think it requires a swap partition and I have a single btrfs.
tanto no gnome boxes e virt-manager, resolvi o problema instalando esses pacotes no debian com comando a seguir: sudo apt-get install libvirt-daemon-system libvirt-daemon-system-systemd
