https://bugzilla.suse.com/show_bug.cgi?id=1156497 https://bugzilla.suse.com/show_bug.cgi?id=1156497#c23 --- Comment #23 from Dan Čermák <dcermak@suse.com> --- (In reply to Oliver Kurz from comment #22)

Interesting. Could you share a little bit more details where the changes have been applied?

The changes were applied here: https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/10805 This just changes the amount of RAM and CPU cores available to the vagrant boxes in the SUT.

...

So my guess is that probably the openQA worker VM has maybe less cores

hm, do you mean the qemu VM that is started as the SUT?

yes

In the latest test run of the initially mentioned test scenario in the autoinst-log.txt file one can find the qemu command line:

```

/usr/bin/qemu-system-x86_64 -only-migratable -chardev ringbuf,id=serial0,logfile=serial0,logappend=on -serial chardev:serial0 -soundhw hda -global isa-fdc.driveA= -m 1536 -cpu host -netdev user,id=qanet0 -device virtio-net,netdev=qanet0,mac=52:54:00:12:34:56 -boot order=c -device usb-ehci -device usb-tablet -smp 1 -enable-kvm … ```

see the parameters "-m 1536" meaning 1.5GB of RAM and "-smp 1" meaning a single CPU core. Both parameters can be configured with test parameters, e.g. `QEMURAM=2048` and `QEMUCPUS=2`, see https://github.com/os-autoinst/os-autoinst/blob/master/doc/backend_vars. asciidoc for details

So it is as I guessed: if you launch the box with more cores and more RAM than the SUT has, then the CPU locks up when kvm is nested inside kvm. If I'd guess further, then it would be something along the lines of the CPU scheduler not really being build for the situation that the inner KVM VM has more cores than the outer one. -- You are receiving this mail because: You are on the CC list for the bug.