User`s guide

host reinitiates such requests, the surviving controller services them. The time
out of the host’s requests is the only indication to the host that a fault of some
kind has occurred.
The surviving controller continues to monitor the status of the failed controller
until it is restarted or replaced by the user. Once the failed controller is replaced
and its replacement reestablishes communication with the surviving controller,
the failback operation begins, returning subsystem control to the replacement
controller for its SCSI targets.
2.4.2 Transparent Controller Failover Resulting from Operator Action
You can initiate a failover operation by running the C_SWAP local program from
the CLI interface, for the purposes of swapping a controller or cache module. The
C_SWAP program forces the controller on which it is executed to assume control
over the entire storage subsystem.
2–22 Controller Technical Description