Using opennebula 1.4 with Shared-NFS on KVM hypervisor . VM_DIR is shared between all the hosts to allow live migration. If for any reason NFS connection is lost and user tries to create a new VM the following opennebula commands hangs and sits there for over. The VM state is in <b>boot</b> for ever.<br>
<br><b>sh -c ssh hostone 'mkdir -p /mnt/vmdir/4727/images && cat > /mnt/vmdir/4727/images/deployment.0 && virsh --connect qemu:///system create /mnt/vmdir/4727/images/deployment.0' ; echo ExitCode: $? 1>&2<br>
</b><br>Because of that all other VM monitoring commands to that host queue up <br><br><b>sh -c ssh -n hostone virsh --readonly --connect qemu:///system dominfo one-4716 ; echo ExitCode: $? 1>&2</b><br><br>Even libvirt virsh list command hangs.<br>
<br>I think Opennebula should check the storage connectivity before proceeding with VM creation. <br><br>Ranga<br>