When the primary companion fails over on HACMP 5.2, the entire node fails over, not only the primary companion. During this node failover, the IP address of the servicing host (the primary node) is swapped with another standby address. In some networking environments, this may cause all the processes on the initial IP address to freeze and eventually timeout. Because of this, when you use Sybase Failover with HACMP on AIX:
Do not allow clients to log in directly to the primary node
Limit the primary node to running only one high availability application at a time
