Cluster performs fail-overs - detected a problem (cphad) / (fwd)
Both FWD process (monitored by 'FWD' pnote) and CPHAMCSET process (monitored by 'CPHAD' pnote) are running in User Mode.
These processes send life-signs to the kernel on a regular basis (configurable as 'timeout' of the pnote).
When the kernel does not receive the life-signs from these processes for longer than defined in timeout of the corresponding pnote, kernel assumes that monitored process failed. As a result, by design, a 'problem' state is reported on the corresponding pnote, and then, by design, kernel initiates the failover between cluster members.