Bug #9909
closed
Libvirt Compute Resources have same UUID
Description
My compute resources have the same UUID so I can't build or see any information about the second compute resource. When trying to build a system, I selected the second compute resource, but it got created on the first one.
I noticed this at first when I tried to associate vms to a compute resource and the vms moved incorrectly, but it all seems to stem from the same uuids. I believe this is due to the compute resource trying to use the mac address of the network configured in libvirt, which I'm not using because I'm trying to use the system's br0, which was never imported into libvirt.
Updated by Dominic Cleal about 9 years ago
- Category set to Compute resources - libvirt
I don't quite get this, to be honest.
The first issue sound very strange, as Foreman simply uses the connection URL in your compute resource settings (e.g. qemu+ssh://foo.example.com) for where to create a VM. If those are incorrect then it could cause a VM to be created on the wrong hypervisor, but otherwise Foreman should never have any reason to switch them or think they're in any way related.
The second issue could happen if you have a virtual machine on both hypervisors with identical MAC addresses. The UUID shouldn't matter, as it's only the MAC that's compared. Check "virsh dumpxml foo | grep mac" to check what libvirt thinks the VM MACs are, and then compare it to what's stored on the Foreman host before associating.
Updated by Leah Fisher about 9 years ago
- Status changed from New to Closed
Sorry, I finally saw that the url was in fact the same, I don't know how I missed it with everything I was looking at.