Hi,
a) When vcs is hanging
>> If VCS is hanging on any one of the nodes, other node will not receive the updates via LLT & hence other node would detect the failure & initiate the failover. If partial VCS is hanging which means some component of resource or particular resource type is hanging, then also, the monitor process should time out & fault the resource. Depending on parameters like "critical resources" in the path or in service group, VCS will take the decision to failover service group to other node. If in worst case VCS doesn't attempt to failover the service group, you can manually initiate the failover by issuing below command
# hagrp -switch <service_group> -to <target_node>
b) At maintanance activity in node A and need to switch teh service grp to node B?
>> Issue the above mentioned command & failover the group to other node. Once successfully failed over, you can do the maintenance work
G
