User`s guide

2. The firmware executive (EXEC) initializes and a call is made to the diagnostic
and execution monitor (DAEMON) to complete the subsystem self-test
diagnostics. When executing these diagnostics, DAEMON uses the services
provided by the EXEC to handle any interrupts and errors generated during
the tests. Errors detected during these tests are displayed on the OCP LEDs.
The diagnostics run by DAEMON include the following:
• A test of the host port—The controller still becomes operational if this
port is not functioning.
• A test of the device ports—The controller still becomes operational if
one or more of its device ports are not functioning. However, if all of the
SCSI–2 ports are unresponsive, the controller does not initialize.
• A test of the memory devices on the cache module—Only 1 MB
of the memory must test valid for the cache module to be operational.
However, if the cache module fails, the controller still functions by using
a portion of shared memory for minimal read caching operations.
The remainder of cache is tested after initialization, in the background
during normal operation.
3. The EXEC finishes the initialization and starts the functional firmware on
the controller.
If a hard failure occurs anytime during the initialization sequence, the controller
OCP LEDs display an error indication.
CAUTION
The program card must remain in its slot for continuous controller
subsystem operations. If the program card is removed during normal
operations, the controller automatically resets.
5.1.1 Dual-Redundant Controller Configuration Initialization Sequence
The two controllers in a dual-redundant configuration exchange messages during
their individual initialization sequences. The first message occurs just after the
first controller starts initializing. This message informs the other controller that
an initialization is taking place, so that the other controller does not assume that
the initializing controller is malfunctioning and does not attempt to disable it.
5.1.2 Controller Subsystem Initialization
When the controller is part of a StorageWorks subsystem, controller initialization
takes place during power-up. In the event of a reset due to a partial or complete
power failure, equipment failure, error condition, or a host clear, a subset of the
initialization sequence runs in the subsystem.
A complete StorageWorks subsystem initialization includes the following steps:
1. When turning on the power, all shelves in the subsystem are reset and
entities in the shelves, including SBB devices, controller, and cache modules,
begin their initialization sequences.
2. Each entity then completes its initialization and self-tests as applicable.
Controller Operations 5–3