HP A7990A - StorageWorks SAN Director 4/16 Blade Switch, Hardware Reference Manual

Page: 18 / 144

18 Overview
• Hot-swappable World Wide Name (WWN) cards
The high-availability software architecture of the DC Director provides a common framework for all
applications that reside on the system, allowing global and local states to be maintained through any
component failure. High-availability elements consist of the High Availability Manager, the heartbeat, the
fault/health framework, the replicated database, initialization, and software upgrade.
The High Availability Manager controls access to the standby CP blade, facilitates software upgrades,
prevents extraneous switchover activity, closes and flushes streams, provides flow control and message
buffering, and supports a centralized active and standby state.
Reliability
The DC Director uses the following error detection and correction mechanisms to ensure reliability of data:
• Data is protected by the Error Detection and Correction mechanism, which checks for encoder errors
and fault isolation (EDFI), such as cyclic redundancy checking (CRC), parity checking, checksum, and
illegal address checking
• Power-on self-test (POST)
• Dual CP Blades that enable hot, non-disruptive fast firmware upgrades
• Each CP Blade contains one serial port and two Ethernet ports, for management and for service.
Offline CP Blade diagnostics and remote diagnostics simplify troubleshooting. The standby CP Blade
monitors diagnostics to ensure it is operational, in case a failover is required.
• Bus monitoring and control of blades and other FRUs.
Serviceability
The DC Director provides the following features to enhance and ensure serviceability:
• Modular design with hot-swappable components
• Flash memory that stores two firmware images per CP Blade
• USB port on CP blades for all tasks that formerly required an FTP/SCP server, including software and
firmware upgrades.
• Non-volatile random-access memory (NVRAM), containing the serial number, serial number, revision
information, and part number information.
• Background health-check daemon
• Memory scrubber, self test, and bus ping to determine if a bus is not functioning
• RASlog messages
• SMI-S compliant
• Watchdog timers
• Status LEDs
• Predictive diagnostics analysis through Fabric Watch
• SNMP (including version 3) integration with higher-layer managers
Software features
The Fabric OS allows any Fibre Channel-compliant device to attach to the switches as long as it conforms
to the device login, name service, and related FC standards. Each operating environment requires that a
FC Host Bus Adapter (HBA) be available with a standards-compliant driver for correct interface to the
fabric.
Fabric OS consists of a set of embedded applications running on top of an embedded Linux operating
system kernel. These applications include:
• Name server
• Alias server
• Zone server
• Simple Network Management Protocol (SNMP) agent