Created attachment 1025320 [details] fedora 20 VM pstack Description of problem: Sometimes my VMs hang with strange CPU usage pattern after wake up from suspend. I'm running fedora 20 and fedora 21 VMs. Sometimes even virtual machine manager GUI becomes unresponsive tryingto work with such VM. Strangely though after some time the machine recovers by itself (sometimes). In this occasion both machines returned to normal after calling `pstack` on them. Attaching logs from both. Version-Release number of selected component (if applicable): qemu-kvm-2.1.3-6.fc21.x86_64 How reproducible: hard, usually machine should be suspended for several hours
Created attachment 1025321 [details] fedora 21 VM pstack
What's running inside the VMs? Does whatever is running in the VMs use NTP or kvmclock?
Created attachment 1025322 [details] fedora20 qemu log
Created attachment 1025323 [details] fedora 21 qemu log
Created attachment 1025324 [details] CPU usage pattern of fedora21 when hung attaching screenshot of CPU usage as reported by virtual machine manager
@Richard, on the fedora20 machine I see /usr/sbin/chronyd running and on the fedora 21 machine I don't see neither ntp, nor chronyd running. On both machines I see: > cat /sys/devices/system/clocksource/clocksource0/current_clocksource > kvm-clock
How often does the hang happen? For the f21 VM, please provide: - sudo virsh dumpxml $vmname - /var/log/libvirt/qemu/$vmname.log
Created attachment 1028014 [details] fedora 21 xml dump Attaching the xml dump. QEMU log already attached.
Created attachment 1029780 [details] fedora 20 - vm-manager hang on reset I just got another hang. Trying to 'reset' the VM through virtual machine manager caused it to hang and I see some errors in qemu log (attached). `virsh destroy` fixed the issue.
(In reply to Aleksandar Kostadinov from comment #9) > Created attachment 1029780 [details] > fedora 20 - vm-manager hang on reset > > I just got another hang. Trying to 'reset' the VM through virtual machine > manager caused it to hang and I see some errors in qemu log (attached). > `virsh destroy` fixed the issue. This is likely a separate issue, please file a new bug report
Both your VMs have this: <pm> <suspend-to-mem enabled='no'/> <suspend-to-disk enabled='no'/> </pm> Which means s3/s4 support should be disabled. This is the default in virt-manager and gnome-boxes. How are you triggering suspend in the VM? Can you do this for your VMs, reboot them, and see if the suspend issues go away: sudo virt-xml $vmname --edit --pm suspend_to_mem=on,suspend_to_disk=on --confirm FWIW there's another bug that was similar to this: https://bugzilla.redhat.com/show_bug.cgi?id=1223541
I do not trigger suspend in the VMs. I just close the laptop lid and it is suspending the host OS. I'll try out the option though to see if it resolves any of the issues. wrt filing a new bug - what should it be all about? One has to reproduce the VM freeze to get vm-manager freeze.
Sorry I got my bugs confused, thought this was about suspending inside VMs.
Created attachment 1057904 [details] pstack fedora22 on fedora21 host I've got another hang, attaching pstack.
One other strange thing. Now my vm is fedora22 and host is fedora22. I tried regular machine shutdown but it somehow neded up in `pmsuspend` mode. I had to `virsh dompmwakeup vmname` and then force off to recover.
*** This bug has been marked as a duplicate of bug 1233568 ***
*** This bug has been marked as a duplicate of bug 1380893 ***