On Mon, Dec 02, 2013 at 10:32:15AM +0100, Nicolas Ecarnot wrote:
Le 29/11/2013 13:36, Dan Kenigsberg a écrit :
Do you mean the xml that one can see in the vdsm.log?
Actually, I've meant everything: from Engine's parameters to the vmCreate command, down to the error from libvirt. The domxml blow suggests that the error might be even further below, since I do not see an address collision here.
<disk device="cdrom" snapshot="no" type="file"> <address bus="1" controller="0" target="0" type="drive" unit="0"/> <target bus="ide" dev="hdc"/> </disk> <disk device="disk" snapshot="no" type="block"> <address bus="0" controller="0" target="0" type="drive" unit="0"/> <target bus="ide" dev="hda"/> </disk> <disk device="disk" snapshot="no" type="block"> <address bus="1" controller="0" target="0" type="drive" unit="1"/> <target bus="ide" dev="hdb"/> </disk> <disk device="disk" snapshot="no" type="block"> <address bus="0" controller="0" target="0" type="drive" unit="1"/> <target bus="ide" dev="hdd"/> </disk>
Could you provide the vmCreate line from vdsm.log, and also the part of libvirtd.log since that domxml gets in until the error is spewed out?
Wow, this is weird! After some shutdown/reboot (I have other things to do with this VM), I CAN NOT reproduce the issue. I swear I have not changed the disk setup, neither added a component or what else. What I get now is the VM up and running with 3 IDE disks, a state I couldn't reach so far. I don't know what to conclude?
shutdown/reboot of what? Vm? Host? Engine? Maybe you have still access to the old vdsm.logs? Otherwise the conclusion is that someone else would bump into this bug.. I have a suspicion that Engine is somehow confused by the device addresses reported by Vdsm. But I'd need to see the logs to stop guessing.