I moved all my VMWare servers to qemu-kvm on CentOS 5.6 64-bit systems using qcow2 images. Before any operation on a VM I'm used to take a snapshot first. So I was looking into a way to do that with qemu-kvm on a less important VM. Here's my story. Take a snapshot: qemu-img snapshot -c somename.snapshot myvm.qcow2 Let's check this out: qemu-img snapshot -l myvm.qcow2 Looking good. Maybe an extra snapshot: qemu-img snapshot -c somename.snapshot myvm.qcow2 Great, two snapshots. Not possible with VMWare server 1. Now, let's delete a snapshot: qemu-img snapshot -d myvm.qcow2 Nope, didn't work. Or was it: qemu-img snapshot -d somename.snapshot Didn't work either. The VM is still running though. I've been reading you shouldn't take a snapshot on a live VM? I shutdown the VM, no luck. Now the VM doesn't boot anymore. After one error after another I tried this: qemu-img snapshot -l myvm.qcow2 qemu-img: Could not open 'myvm.qcow2' I rebooted the host server, but no matter what I try, qemu-img is unable to open the image. This is the error when I try to boot the vm: File "/usr/lib64/python2.4/site-packages/libvirt.py", line 333, in create if ret==-1: raise libvirtError ('virDomainCreate() failed', dom=self) libvirtError: internal error Process exited while reading console log output: raw_pread: read beyond end of file How do I fix this before taking snapshots on production servers?
Looks like I shouldn't been taken a snapshot from a running guest. I was not able to repair the image.
