To demonstrate what would happen if the primary server were to fail, you can kill off its process, which should initiate the failover process.
The SEPs information appears.
The Windows Task Manager appears.
The PID column appears.
run runner.int
This reruns the previous program; if you examine the system directories for each of the remaining servers, you'll notice that the timestamp for file.dat has been updated, confirming that the file is still being replicated within the group.