Hello Evgeni,

as this bug report is quite old, I first re-checked this with the versions in
testing from a few days ago:

qemu-system-x86                      1:5.1+dfsg-4+b2
tuned                                2.10.0-1
linux-image-5.9.0-4-cloud-amd64      5.9.11-1

Without tuned, cirros boots, and gets stuck after a while trying to find a
cloud-init source:

  checking http://169.254.169.254/2009-04-04/instance-id
  failed 1/20: up 9.66. request failed
  [...

But that's fine. With tuned, the symptoms are a bit different now: QEMU does
not show "warning: host doesn't support requested feature: CPUID" any more; but
we already established that this was just a red herring.

I see the "GRUB loading, please wait..." for several seconds, then the screen
gets cleared, and nothing happens any more.  QEMU is spinning at 100% CPU. So
aside from the warning, it's still the same.

I then dist-upgraded this VM to current sid:

qemu-system-x86                      1:5.2+dfsg-2
tuned                                2.14.0-1
linux-image-5.9.0-5-cloud-amd64      5.9.15-1

The behaviour is still the same.

This is in our cockpit test VMs, which are based on the official testing cloud
images, but they install some additional packages.

I re-tested this straight from the current official cloud images:

  curl -o sid.qcow2 
https://cloud.debian.org/images/cloud/sid/daily/20201220-490/debian-sid-nocloud-amd64-daily-20201220-490.qcow2
  qemu-system-x86_64 -cpu host -nographic -m 2048 -device virtio-rng-pci -drive 
file=sid.qcow2,if=virtio -snapshot

The `-cpu host` option is necessary to provide /dev/kvm inside the VM (the
default "QEMU Virtual CPU" does not).

Log in as "root" (no password), then run

  apt update
  apt install -y --no-install-recommends qemu-system-x86 tuned

  curl -L -O https://download.cirros-cloud.net/0.3.5/cirros-0.3.5-i386-disk.img
  qemu-system-x86_64 -enable-kvm -nographic cirros-0.3.5-i386-disk.img -snapshot

The boot hangs as above. Again, when not installing or stopping tuned, it
works.

Thanks,

Martin

Reply via email to