The problem VCS agents would report errors similar as below
VCS WARNING V-16-1-10023 Agent
Running the command, would give a list of the defunct process
#ps -ef |grep -i defunct
When VCS agents stop sending 'I am alive' messages, then High Availability Deamon (HAD) will attempt to kill the 'hung' agent processes and restart them.
During this activity it may be possible that some of the VCS Agent processes do not get killed and show up as 'defunct' processes.
1) Stop the cluster using 'hastop -local -force'
2) Now, manually kill these 'defunct' processes.
3) Run the command "ps -ef |grep -i defunct" to confirm the defunct process are deleted.
If issue still persists, then a system reboot will be needed to clean them the defunct processes.