On Sun, 2011-04-10 at 03:47 -0500, Johnny Hughes wrote: > On 04/09/2011 12:04 PM, compdoc wrote: > >> A similar incident was reported during the QA. Look at the .xml file. > >> If it says type='raw', change it to type='qcow2' and restart libvirtd. > >> Would that fix the problem ? > >> > >> Akemi > > > > Thank you. After reading your message, I googled the error and found a > > webpage that describes a slightly different procedure than yours, but which > > does the same thing: > > > > http://ubuntuforums.org/showthread.php?t=1638708 > > > > Everything is working now. > > > > :) > > I am going to add this to the Release Notes for 5.6 on the Wiki now. Turns out that wasn't the only problem I faced in my migration. With 2 KVM servers, both sharing a volume mounted via NFS for VMs, I migrated all VMs to the second node, upgraded the first, them moved them all back to KVM1. Instant disk corruption on all VMs. Boom. I have a second pair of KVM servers. I tested one VM with my normal migrate-them-out-of-the-way procedure, and it, too, suffered MASSIVE filesystem corruption. This was even after I'd made the qcow2 mods and restarted libvirtd. The only way I was able to not have to rebuild the remaining non-corrupted VMs was to shut them down on one node then bring them back up again. Turns out live migration doesn't work in this upgrade. (Though I'll test regular live migration tomorrow, given that all 4 KVM servers have now been upgraded.) -I