The `controld` resource agent properly returns an error if DLM is waiting for fencing Previously, the monitor action of the `ocf:pacemaker:controld` resource agent, used to manage the DLM control daemon in a Pacemaker cluster, did not report an error if DLM was waiting for fencing. Therefore, a Pacemaker cluster using such a resource could end up with DLM (and any resources relying on it, such as CLVM) hanging indefinitely if all nodes lost and regained quorum. Now, the agent properly returns an error in that case, and Pacemaker properly recovers the resources.