Hide Forgot
Description of problem: After blockpull a disk from backing file and then do blockpull again for the same guest and image, the guest will be destroyed. Version-Release number of selected component (if applicable): qemu-kvm-0.12.1.2-2.211.el6.x86_64 kernel-2.6.32-223.el6.x86_64 libvirt-0.9.8-1.el6.x86_64 How reproducible: 100% Steps to Reproduce: 1. Prepare a qed image with installing guest on it 2. Create new qed image with backing file pointing on above disk # qemu-img create -f qed -o backing_file=/var/lib/libvirt/images/guest.qed,backing_fmt=qed /var/lib/libvirt/images/new.qed # qemu-img info /var/lib/libvirt/images/new.qed image: /var/lib/libvirt/images/new.qed file format: qed virtual size: 3.0G (3221225472 bytes) disk size: 260K cluster_size: 65536 backing file: /var/lib/libvirt/images/guest.qed (actual path: /var/lib/libvirt/images/guest.qed) 3. Start a guest with new qed image # virsh dumpxml guest ... <disk type='file' device='disk'> <driver name='qemu' type='qed' cache='none'/> <source file='/var/lib/libvirt/images/new.qed'/> <target dev='hda' bus='ide'/> <address type='drive' controller='0' bus='0' unit='0'/> </disk> ... # virsh start guest 4. Do blockpull for this new qed image and guest # virsh blockpull guest /var/lib/libvirt/images/new.qed 5. Wait for the blockpull finished # virsh blockjob guest /var/lib/libvirt/images/new.qed --info until nothing return 6. After it finished, do blockpull to the same image and guest again # virsh blockpull guest /var/lib/libvirt/images/new.qed Wait for several seconds, then check the guest status # virsh list Actual results: The guest is destroyed. Expected results: The guest still running. Additional info: I test with libvirt virsh command in above steps, and I also test with qemu human monitor command block_stream and find that the problem still exist, so I report the bug in qemu-kvm
(In reply to comment #0) > 5. Wait for the blockpull finished > # virsh blockjob guest /var/lib/libvirt/images/new.qed --info > until nothing return > > 6. After it finished, do blockpull to the same image and guest again > # virsh blockpull guest /var/lib/libvirt/images/new.qed Aren't you destroying the first disk you created? > > Wait for several seconds, then check the guest status > # virsh list > > > Actual results: > The guest is destroyed.
(In reply to comment #2) > (In reply to comment #0) > > 5. Wait for the blockpull finished > > # virsh blockjob guest /var/lib/libvirt/images/new.qed --info > > until nothing return > > > > 6. After it finished, do qemu-img create -f qed -o backing_file=/var/lib/libvirt/images/guest.qed,backing_fmt=qed /var/lib/libvirt/images/new.qed to the same image and guest again > > # virsh blockpull guest /var/lib/libvirt/images/new.qed > > Aren't you destroying the first disk you created? > Even if I remove the first disk guest.qed, the second time executing blockpull will still cause guest destroied. > > > > Wait for several seconds, then check the guest status > > # virsh list > > > > > > Actual results: > > The guest is destroyed.
(In reply to comment #3) > (In reply to comment #2) > > (In reply to comment #0) > > > 5. Wait for the blockpull finished > > > # virsh blockjob guest /var/lib/libvirt/images/new.qed --info > > > until nothing return > > > > > > 6. After it finished, do qemu-img create -f qed -o > backing_file=/var/lib/libvirt/images/guest.qed,backing_fmt=qed > /var/lib/libvirt/images/new.qed > to the same image and guest again Please ignore above comments. Sorry > > > # virsh blockpull guest /var/lib/libvirt/images/new.qed > > > > Aren't you destroying the first disk you created? > >