[Linux-cluster] vm.sh with and without virsh

[jr]

Hello everybody,
I'm having trouble using the current vm.sh that uses virsh instead of xm 
commands.
When i'm using the new one, the dom0s are unable to start any of my 
virtual machines.
This is what i get:

[root at PhySrv07 ~]# rg_test test /etc/cluster/cluster.conf start vm 
LVSDirector01
Running in test mode.
Starting LVSDirector01...
Hypervisor: xen
Management tool: virsh
Hypervisor URI: xen:///
Migration URI format: xenmigr://target_host/
Virtual machine LVSDirector01 is libvir: Xen error : Domain not found: 
xenUnifiedDomainLookupByName
error: failed to get domain 'LVSDirector01'

<debug>  virsh -c xen:/// start LVSDirector01
libvir: Xen error : Domain not found: xenUnifiedDomainLookupByName
error: failed to get domain 'LVSDirector01'

Failed to start LVSDirector01

using the old non-virsh /usr/share/cluster/vm.sh agent i can 
successfully start the virtual machines:


[root at PhySrv07 ~]# rg_test test /etc/cluster/cluster.conf start vm 
LVSDirector01
Running in test mode.
Starting LVSDirector01...
# xm command line: LVSDirector01 on_shutdown="destroy" 
on_reboot="destroy" on_crash="destroy" RGMANAGER_meta_refcnt="0" 
depend_mode="hard" exclusive="0" hardrecovery="0" max_restarts="0" 
--path="/cluster/XenDomains" restart_expire_time="0"
Using config file "/cluster/XenDomains/LVSDirector01".
Started domain LVSDirector01
Start of LVSDirector01 complete

I get a feeling that the way I have my domUs prepared is not suited for 
virsh or I am missing some deps I should have met.
What is required for vm.sh successfully using virsh? (I'd really rather 
have that for the synchronous/asynchronous nature of virsh vs xm).
Thanks and best regards,
Johannes