During normal companion mode, do not run both Adaptive Servers in a symmetric configuration at the full capacity of their system resources. For example, each machine can run at 60% of the possible configuration for user connections, data cache, remote server connections, and so on. This allows the secondary companion to manage both the failed-over Adaptive Server and its own Adaptive Server with a reasonable level of performance. If both Adaptive Servers maximize their system resources, failover succeeds, but performance may be poor.
