Limitations on AdaptiveHA

AdaptiveHA enables VCS to make dynamic decisions about failing over an application to the biggest available system. Due to the overcommitting behavior of the virtualization technologies such as KVM, RHEV, and VMware, a user can create multiple guest system in a host system with system resources (Memory, CPU or SWAP) values higher than the actual values available in the host. Due to this behavior VCS engine considers the guest system with the overcommitted resources as highly available for failing over an application.

For example, on a host machine with in-built physical memory of 4 GB, if you create two guest systems each with 4 GB and 8 GB of memory respectively, VCS engine selects the guest system with 8GB of overcommitted memory value for failing over an application irrespective of the available physical memory in the host.