Hide Forgot
Description of problem: [RFE] libvirt enhanced notifications (diskspace on hdd with VM image) I recently ran into situation when libvirt was stopping one VM for unknown reason. There was no message neither in /var/log/messages nor in libvirt log file. It was finally caused by fact that the VM I was starting had sparse file image and there was no diskspace, so qemu-kvm was not able to write data into VM image. I believe it is needed to inform customer why VM was internally stopped (not by user) or more generally, it would be great to have improved virtualization notifications about similar events, both in libvirt and also in virt-manager. Version-Release number of selected component (if applicable): N/A (libvirt-0.10.2-18.el6_4.15.x86_64 libvirt-0.10.2-29.el6.x86_64) How reproducible: N/A (100%) Steps to Reproduce: N/A Actual results: Missing notifications in some situations. Expected results: Notifications in place. Additional info:
What do you mean "for unknown reason"? Did you try "virsh domstate --reason VM"? Libvirt should tell you why the domain was paused and it also sends an event to registered clients when this happens. Thus, I guess the only part you are missing is support for state reasons and events in virt-manager, right?
Thanks for pointing me to `virsh domstate --reason` command [root@mx ima]# virsh domstate --reason r6x2 paused (I/O error) which clearly points to disk / image problem. Command-line users are able to see the reason. The outstanding item is to add those notifications into virt-manager. Thanks for quick reply.