Avoiding a partitioned cluster

Should the cluster become partitioned, each partition might force on the volume group and continue to run. In this case, two unequal copies of the data will be active at the same time. This situation can cause data divergence and does not allow a clean recovery. Were this to happen with a concurrent volume group, the consequences would be even worse, as the two sides of the cluster would have made uncoordinated updates.

PowerHA® SystemMirror® automatically uses monitoring of all available storage and network paths to avoid a partitioned cluster