Automatic Split-Brain resolution
In case there is no communication path between the nodes, Cluster Resource Manager can activate
resources on both nodes at the same time. This situation is called Split-Brain. After both nodes appear
on the network again, the data changes that occurred on the nodes must be discarded on one of them.
System Platform High Availability Failover uses a DRBD feature to recognize which node became active
as the last one. This node is rebooted immediately and its data changes are discarded. After successful
reboot, it will synchronize the changes that occurred on the survival node since the disconnection before
split-brain.
™
Installing and Configuring Avaya Aura
System Platform
April 2010
99