If your cluster node experiences a CPU, RAM, or motherboard failure, it is an unrecoverable failure and the cluster node will go offline. The cluster fencing mechanism will try to fence the problematic cluster node anyway, making sure it is not accessing the cluster's shared storage, which could lead to cluster data corruption. If your cluster node experiences disk failure, it should not affect the cluster node operation due to RAID disk redundancy, which enables normal cluster operation.
The cluster from the following example is configured to provide a webserver cluster service that includes a cluster IP address and an Apache webserver instance. The webserver cluster service is running on the node-1 cluster node. In the following screenshot, you can see the current state of the cluster:
You can proceed to simulate hardware failure on the part of the node-1 cluster node. This can be done by cutting off the power to the cluster node.