Description of problem: After shutting down a guest, libvirtd is no longer running. /var/log/messages says: Aug 27 11:44:58 worm libvirtd: 11:44:58.794: error : qemudDomainGetMemoryBalloon:3229 : operation failed: could not query memory balloon allocation Version-Release number of selected component (if applicable): libvirt-0.7.0-6.fc12.x86_64 qemu-0.10.91-0.8.rc1.fc12.x86_64 kernel-2.6.31-0.174.rc7.git2.fc12.x86_64 qemu-system-x86-0.10.91-0.8.rc1.fc12.x86_64 How reproducible: 100% Steps to Reproduce: 1.Start guest. 2.Shut down guest. 3.Attempt to start guest. Actual results: libvirtd is no longer running. Expected results: Still running after guest is shut down.
Hmm, I don't seem to be seeing this now...
...and of course it just happened again, while shutting down an F-11 guest.
I can't reproduce this Was it a segfault? Can you get a stack trace?
Can you provide the guest configuration file which exhibits this crash. Also is this being seen with a guest whose disks are on NFS as per your other bugs ?
It turns out libvirtd is still running but virt-manager's connection to it is broken. No, the disks are not on NFS. $ ls -Z /home/twaugh/VM/RHEL54.img -rw-rw-r--. root root system_u:object_r:virt_image_t:s0 /home/twaugh/VM/RHEL54.img
Created attachment 361476 [details] RHEL54.xml
I was hitting this issue as well, but not anymore after updating libvirt to latest rawhide version (libvirt-0.7.1-1.fc12.x86_64). Tim, can you give that a shot? Guest was win2003, no fancy setup, just a single ide disk. shutdown (not destroy) caused the issue. virsh list would show 'nostate' for the VM, and all 'info' calls would fail with "could not query memory balloon'. There is another bug filed to handle virt-manager's poor behavior in coping with this situation (bz 522168).
twaugh: has latest libvirt fixed it for you too?
I'm not seeing it break with libvirt-0.7.1-4.fc12.x86_64.
Thanks Tim
*** Bug 500968 has been marked as a duplicate of this bug. ***