This chapter introduces you to the Silicon Graphics fibre channel mass storage options available with your Origin2000 or Onyx2 host system.
This chapter includes the following main sections:
Fibre channel is the general name of an integrated set of standards being developed by the American National Standards Institute (ANSI).[1] The fibre channel standard defines a high-speed data transfer interface that can be used to connect workstations, mainframes, supercomputers, storage devices, and displays. The fibre channel standard addresses the need for fast transfers (up to 1 Gbit per second) of large amounts of information. Currently, fibre channel's main use is as an interface to storage.
Conceived as a generic, efficient physical transport system that can support multiple protocols, the standard also relieves system manufacturers of the burden of supporting the variety of channels and networks currently in place, because it provides one standard for networking, storage, and data transfer. Note that the Silicon Graphics initial implementation is for communication with mass storage systems only.
The fibre channel provides a general transport vehicle for upper level protocols (ULPs), including the intelligent peripheral interface (IPI) and small computer system interface (SCSI) command sets; high-performance parallel interface (HIPPI) data framing; internet protocol (IP); and IEEE 802.2. Proprietary and other command sets can also use and share the fibre channel, although such use has not been defined as part of the fibre channel standard and is not supported by Silicon Graphics host systems. The Silicon Graphics implementation currently supports only the SCSI fibre channel protocol.
In recent years, technical developments have precipitated a greater need for extremely fast data links. Performance improvements have spawned increasingly data-intensive and high-speed networking applications, such as multimedia and scientific visualization. However, the existing network interconnects between computers and I/O devices have been unable to run at the speeds needed. Fibre channel development has been aimed at practical, inexpensive, yet expandable means of quickly transferring data between workstations, mainframes, supercomputers, desktop computers, storage devices, displays, and other peripherals.
The two basic types of data communication between processors and peripherals have been channels and networks:
A channel provides a direct or switched point-to-point connection between the communicating devices. A channel is typically hardware intensive and transports data at high speed with low overhead.
A network is an aggregation of distributed nodes (workstations, file servers, or peripherals) with its own protocol that supports interaction among these nodes. A network has relatively high overhead because it is software-intensive; consequently, it is slower than a channel. Networks can handle a more extensive range of tasks than channels because they operate in an environment of unanticipated connections, whereas channels link only a few devices with predefined addresses.
Table 1-1 summarizes the differences between networks and channels.
Table 1-1. Networks and Channels
Network | Channel |
|---|---|
Unanticipated connections; must be able to connect any pair of devices; addresses and routes are usually not predefined | Defined domain, usually a closed-system, small set of devices in close proximity; every device address is predefined |
Software intensive; high protocol overhead | Hardware intensive; low overhead |
Interconnects processors at moderate to high speeds using a form of switching over relatively large distances | Interconnects processors and high-performance peripherals at high speeds, relatively short point-to-point links |
Uses various protocols for error detection, error correction, and data retransmission | Simple error correction or data retransmission on receipt of a busy signal; message is passed back to central processor if more complex error recovery is required |
The fibre channel (FC) standard attempts to combine the simplicity and speed of channel communications with the flexibility and interconnectivity of protocol-based network communications. Despite the name fibre channel, the FC architecture represents neither a channel nor, for that matter, a real network topology.
Fibre channel ports can be connected in three topologies: as point-to-point links, in an arbitrated loop, or to a fabric, or switch. The Silicon Graphics fibre channel option topology is the arbitrated loop (FC-AL). Chapter 2, “Fibre Channel Architecture,” explains this topology in detail. Some of the benefits of FC-AL are
Low connectivity cost.
FC-AL supports up to 125 devices per single host adapter, reducing costs as well as cabinet or subsystem space. (Silicon Graphics supports up to 110 disks per fibre channel loop.)
Copper cabling is used for device connection within each subsystem (cabling one enclosure to another in a loop).
Host-to-storage enclosure connections can be copper or (optionally) fiber optic.
Fibre channel cables use fewer connection pins and conductors.
High performance. FC-AL features make it well suited to both high-speed and transaction-processing applications:
The dual-ported fibre channel XIO interface delivers up to 200-MB-per-second peak data transfer rates.
FC-AL supports host-to-primary enclosure connection distances up to 25 meters using copper cable and up to 300 meters with optical cable options.
Silicon Graphics FC implementation relies on dual-ported fibre disk drives that provide redundant FC data paths from the enclosure backplane to the drive.
Reliability and ease of use. FC-AL devices are quick to install or reconfigure:
The high reliability is advantageous in densely populated cabinets or subsystems.
The FC-AL design enables higher data availability for today's systems and applications.
Widely available standard components are used (no jumpers are required to configure the drive).
Sophisticated error detection code provides better data integrity than traditional parity bit schemes.
Support for SCSI protocols.
Low latency for decreased overhead and more throughput.
Storage. The Silicon Graphics implementation supports a rack with 11 enclosures and up to 110 disks total on a single FC-AL loop
Data transfer. DC-balanced 8b/10b signals using odd or even disparity; variable-length data frames (maximum 2 KB) with complete error checking (32-bit CRC).
Shortwave optical interface adapters (media interface adapters) make it possible using optional fibre optic cable, to place the storage up to 300 meters away from the host.
Fibre channel is useful for a variety of applications:
Scientific graphics and video markets, using high-bandwidth rates with large I/O requests.
File server and database markets, which must support large amounts of I/Os per second (IOPS) with relatively small random I/Os.
Some database applications also require access to vast amounts of storage, which can be in the terabytes range. High availability is also a requirement for much of this market. The Silicon Graphics fibre channel implementation addresses these needs by supporting up to 100 RAID disks in an IOPS environment on a single FC-AL.
For applications requiring large capacity with lower IOPS capability, the Silicon Graphics fibre channel implementation supports up to 110 non-RAID disks on a single FC-AL.
The emerging video-on-demand market, which must provide many random data streams at a guaranteed data rate.
High-availability RAID is also a requirement for this market. Fibre channel RAID addresses these needs by supporting up to 100 RAID disks in a well-balanced access pattern on a single arbitrated loop. The arbitrated loop also provides an arbitration fairness scheme that prevents high-priority requests from starving low-priority requests.
Table 1-2 summarizes how fibre channel compares with other storage interfaces.
Table 1-2. Silicon Graphics Storage Interface Solutions
| Fast-10 Wide SCSI-2 | Fast-20 Wide (Ultra) SCSI | Silicon Graphics Gigabit Fibre Channel Implementation |
|---|---|---|---|
Bandwidth | 20 MBps | 40 MBps | 100 MBps |
Duplex | Half | Half | Full |
Distance (copper) | 25 m | 25 m | 25 m (27.3 yards) |
Distance (optical) | N/A | N/A | 300 m (328 yards) |
Connectivity | 16 | 16 | 110 +1 FC-AL |
Administrative method for hot-plugging a disk | No | Yes | Yes |
Fair arbitration | No | No | Yes |
Optical support | No | No | Yes (optional) |
Dual port | No | No | Yes (optional) |
Conductors in cable | 68 | 68 | 4 in copper or 2 in optical fiber |
Fibre channel can address high availability in several ways. At the simplest level, the dual ports of one or more FC disks or disk enclosures can each be connected to a different host so that, if a host fails, the other host can still access the disks. Optional secondary fibre channel link control cards (LCCs) installed in the fibre enclosures provide an additional physical connection to the second port of each disk in the vault. High-availability loop configurations require the use of an optional FailSafe configuration with a fibre channel hub (or switch).
Optional redundant power supplies for FC disk enclosures that can be connected to separate power sources are available. They provide protection against power supply failure and some insurance against power circuit failure.
Each enclosure uses a fan pack with a redundant fan. The pack can be removed and replaced immediately without shutting off the disk enclosure.
RAID technology provides redundant disk resources in disk-array configurations that make the storage system more highly available. A RAID array maintains parity or mirrored data that lets the disk group survive a single disk module failure without losing data.
RAID disk arrays continue operating even if one of the disk modules fail. When a disk module in the array does fail, the data is still available and the storage processor (SP) can reconstruct it from the surviving disk(s) in the array. The RAID subsystem begins reconstructing data as soon as
A hot spare (dedicated replacement disk module) is available
The failed disk module is replaced with a new disk module
If a disk module has been configured (bound) as a hot spare (see “RAID Hot Spare”), it is available as a replacement for a failed disk module immediately. Reconstruction begins as soon as the failure is detected. The SP automatically writes to the hot spare and rebuilds the group using the information stored on the surviving disks. Performance is degraded while the SP rebuilds the data and parity on the new module. However, the storage system continues to function, giving users access to all data, including data stored on the failed module.
Similarly, when a new disk module is inserted to replace a failed one, the SP automatically writes to it and rebuilds the group using the information stored on the surviving disks. The length of the rebuild period, during which the SP re-creates the second image after a failure, can be specified when RAID levels are set and disks are bound into RAID units. These processes are explained in the Origin FibreVault and Fibre Channel RAID Administrator's Guide .
The Silicon Graphics fibre channel RAID product line supports most types of RAID disk groups. The Origin FibreVault and Fibre Channel RAID Administrator's Guide defines and describes these supported configurations.
 | Note: RAID functionality is dependent on the release level of your hardware, firmware, and software. Upgrades may be required to support certain RAID functions. Contact your service provider for information regarding upgrades to your fibre channel system. See your system release notes for any changes in the RAID support levels. |
A hot spare is a dedicated replacement disk unit on which users cannot store information. The hot spare is an additional disk added to the enclosure that already has a RAID disk array installed. The capacity of a disk module that you bind as a hot spare must always be at least as great as the capacity of the largest disk module it might replace.
If any single disk in a fibre channel RAID group fails, the SP automatically begins rebuilding the failed disk module's structure on the hot spare. When the SP finishes rebuilding, the disk group functions as usual, using the hot spare instead of the failed disk. When you replace the failed disk, the SP starts copying the data from the former hot spare onto the replacement disk. When the copy is done, the disk group consists of disk modules in the original slots, and the SP automatically frees the hot spare to serve as a hot spare again.
A hot spare is most useful when you need the highest data availability. Note that the SP finishes rebuilding the disk module onto the hot spare before it begins copying data to a replacement disk, even if you replace the failed disk during the rebuild process.
The Silicon Graphics fibre channel implementation allows you to shut down and remove individual disk modules in the non-RAID FibreVault using a command-line interface. This administrative “warm swapping” of disk modules must be executed in a specific sequence to guarantee data integrity.
A single RAID disk module can be removed from an array and replaced without special administrative procedures. This procedure should be done only with full knowledge of the status of the LUN that the drive is part of.
See the Origin FibreVault and Fibre Channel RAID Administrator's Guide for more information. See “Adding or Replacing a Disk Module” in Chapter 4 for mechanical information on removing or replacing a fibre disk.
 | Caution: Pulling and replacing (hot swapping) a drive without implementing the proper administrative sequence could result in data loss if the drives are improperly removed or replaced. |
Two types of fibre channel interface boards are available to plug into host systems. The XIO and PCI interface boards each use the same standard or optional cable connections. The following sections describe their features and options:
The Silicon Graphics host system can use an XIO-based fiber controller board for fibre channel communication. Figure 1-1 shows the XIO fibre channel controller in the overall Origin2000 and Onyx2 I/O structure.
Each XIO fibre board has two connectors, and each connector can be used to control up to 110 fibre disks in a fibre rack. Figure 1-2 shows the fibre channel XIO board and connectors. The 9-pin connector farthest from the locking handle is channel 0; the connector closest to the handle is channel 1.
| Note: The XIO boards used in the OCTANE workstation use a different hook locking mechanism than the one shown in Figure 1-2. Do not install a fibre channel XIO board from an Origin2000 or Onyx2 into an OCTANE workstation. The board may not come out and partial disassembly of the OCTANE workstation may be required to remove it. The boards are functionally identical in all other respects. |
The fibre channel PCI board is a half-size PCI option board that provides a single- connection high-performance FC-AL interface between the host system and the fibre enclosure(s).
This board can be inserted in an Origin200 chassis or in the optional PCI module of an Origin2000, Onyx2 system, or OCTANE workstation.
For an Origin2000 or Onyx2 system, only a qualified Silicon Graphics service engineer should install the board. The board installs in the optional PCI module in XIO slot 2, or by using an optional XIO to PCI adapter board (p/n 030-1275-xxx).
Customers who have an Origin200 server should follow the installation instructions in the latest version of the Origin200 and Origin200 GIGAchannel Owner's Guide (p/n 007-3708-nnn).
Customers with an OCTANE workstation must follow the installation instructions in the OCTANE PCI Module Installation Guide (p/n 007-3547-nnn) or OCTANE Workstation Owner's Guide .
Figure 1-3 shows the fibre channel PCI board.
Each fibre channel XIO board has two 9-pin connectors. The fibre channel PCI interface board has a single 9-pin connector. Note that each fibre channel loop uses one fibre channel connector.
Figure 1-4 shows the fibre channel connector pin assignments.
The fibre channel option board provides bus-protocol conversion between an 8-bit (400- MB-per-second) XIO interconnect link and a 64-bit PCI interface. This conversion is provided by the fibre channel board's “bridge” interface ASIC, see Figure 1-5. The second (PCI-to-fibre-channel) conversion is provided by the PCI-64 fibre channel ASIC. The board also provides an interface between a little endian 64-bit PCI interface and a 1.062-Gb fibre channel interface.
The FC option board supports fibre channel Class 3 operations as a loop port (L_Port). The firmware supports Class 3 and FC-AL transfers only. For details on ports, Class 3, and FC-AL, see Chapter 2, “Fibre Channel Architecture.”
If the fibre channel XIO board loses power or the physical fibre channel connection is broken, the link that the board is attached to becomes inoperable. High-availability loop configurations require the use of an optional FailSafe configuration with a fibre channel hub (or switch).
Transceivers (SerDes) embedded on the board convert 10-bit parallel data using 8b/10b encoding to differential serial signals. They also provide frame synchronization, word alignment, and clock recovery for incoming serial data.
For incoming serial data, two recovered clocks at 53.125 MHz for odd or even bytes are provided as outputs on two pins, 180 degrees out of phase. Any required equalization to compensate for high-frequency losses for copper cables (by attenuating the lower frequencies to match) is supplied externally to the fibre channel option board. Serial data in both directions between the transformer and the external connector is AC coupled using a capacitor.
Figure 1-5 is a fibre channel XIO card block diagram.
Both a copper and an optical interface are supported. The copper interface has a female DB9 connector. Fibre channel arbitrated loop (FC-AL) disk drives have a copper native interface.
For distances up to 300 meters, an optional fiber-optic interface option using a media interface adapter (MIA) module is available. A hot-pluggable MIA external module with shortwave CD-ROM lasers provides the physical interface. The MIA is a full-duplex module that converts photons to electrons in one direction and converts electrons to photons in the other direction. An industry-standard duplex SC connector, shown in Figure 1-6, supplies the external fiber-optic connection on the optional MIA.
Grounding issues are very important in Origin and Onyx2 systems. Each chassis must be well grounded through its power connector. All chassis connected by copper cables must share the same transformer, be grounded through the same earthing rod, and be on the same branch circuit. If you have any doubts about the quality of the ground connection, consult with a qualified electrician. Using an optical cable between the fibre enclosure(s) and the host connection eliminates any problems related to common grounding.
| Caution: Any difference in ground potential greater than 500 millivolts (0.5 volts) between two chassis connected by copper cables can cause severe equipment damage and create hazardous conditions. |
The branch circuit wiring should be provided with an insulated grounding conductor that is identical in size, insulation material, and thickness to the earthed and unearthed branch-circuit supply conductors. The grounding conductor should be green, with or without one or more yellow stripes. This grounding or earthing conductor should be connected to earth at the service equipment or, if supplied by a separately derived system, at the supply transformer or motor-generator set. The power receptacles in the vicinity of the systems should all be of an earthing type, and the grounding or earthing conductors serving these receptacles should be connected to earth at the service equipment.
Fibre channel makes it possible to configure an array along a single interface instead of across many separate interfaces, as is done with SCSI storage. Since the drives constituting the FC array unit are organized along an interface, this arrangement is sometimes described as a longitudinal array.
Chapter 4, “Fibre Channel Component Replacement,” is an end-user description of the installation and replacement of FC-AL storage disks. Up to five fibre channel RAID modules can be installed in a rack (each requires a separate loop connection). Fibre channel RAID enclosures use the same XIO fibre channel controller boards as non-RAID enclosure and disk combinations.
This section introduces
Disk modules used in non-RAID fibre channel enclosures can be removed or added using an administrative procedure for taking an individual disk module offline. Sectors on non-RAID drives are 512 bytes, and RAID disks use 520 byte sectors. The two types of drives are not interchangeable. Single RAID disks can be removed without special administrative procedures. Always follow the guidelines listed in the Origin FibreVault and Fibre Channel RAID Administrator's Guide .
Figure 1-7 shows features of the FC-AL disk module.
Each disk module consists of one 3.5-inch FC-AL disk drive in a carrier. The disk drive carrier provides smooth, reliable mating with the chassis slot guides and midplane connectors. The latch on the carrier handle holds the disk module in place to ensure proper connection with the midplane. You must push the latch to remove a disk module from the slot. You can insert the drive module in only one direction. All Silicon Graphics fibre disk drives are dual-port capable.
Always confirm the type, capacity (in gigabytes) and speed of a fibre drive (RAID or non-RAID), before you install it. Each drive module has an identifying sticker with its part number. Table 1-3 provides descriptions and part numbers for the fibre drives.
Table 1-3. Fibre Channel Disk Drive Descriptions
Fibre Drive Type | Drive Capacity | Drive Speed | Sector Sizing | Part Number |
|---|---|---|---|---|
Non-RAID | 9.1 GB | 7200 RPM | 512-byte sectors | 9470140 |
RAID | 8.8 GB | 7200 RPM | 520-byte sectors | 9470192 |
Non-RAID | 9.1 GB | 10,000 RPM | 512-byte sectors | 9470282 |
RAID | 8.8 GB | 10,000 RPM | 520-byte sectors | 9470281 |
Non-RAID | 18 GB | 7200 RPM | 512-byte sectors | 9470255 |
RAID | 17.8 GB | 7200 RPM | 520-byte sectors | 9470257 |
| Note: Always consult with your service provider before replacing a disk with another that has a different capacity or speed rating (different part number). |
Be aware of the following information when adding or replacing disks:
RAID and non-RAID fibre drives are not compatible and may not be mixed in the same enclosure.
A lower capacity RAID drive must not be used as a replacement disk or hot spare in an array of higher capacity disks. For example, you must never install an 8.8 GB RAID drive as a hot spare or replacement in an enclosure of 17.8 GB disks.
Installing a lower speed (RPM) drive in an array of faster RPM disks causes the entire array bandwidth (read and write speed) to slow to the lower speed capacity.
The FibreVault and fibre channel RAID enclosures are high-performance, high–capacity, disk–array subsystems using an FC-AL as an interconnect interface. These modular and scalable designs with optional high-availability features make expansion easy when storage needs increase. The fibre channel RAID expansion is a FibreVault using RAID disk modules.
Each enclosure contains from one to ten disk modules, plus other components. The FibreVault is available in a rackmount or deskside tower enclosure. Figure 1-8 shows an example of the rackmountable FibreVault enclosure.
The front door, which has a built-in electromagnetic interference (EMI) shield, must be closed for the FibreVault or RAID enclosures to be EMI compliant. When the door is open, you can remove or install drive modules and change the enclosure address. Chapter 3, “Fibre Channel Storage,” has complete details of enclosure operation; Chapter 4, “Fibre Channel Component Replacement,” contains complete procedures for replacing and adding disk modules and other owner-replaceable components.
The key for the locking latch fits any FibreVault or fibre channel RAID door (see Figure 1-9). Figure 1-9 shows an example of the fibre channel RAID enclosure.
For proper cooling and normal operation, each disk slot in a fibre enclosure must contain either an FC-AL disk module or a disk module filler. Likewise, each LCC and power supply position must have either the actual component or a filler panel installed.
Fibre channel enclosure features include the following:
Two LED status indicators for each disk module slot:
Check (amber): If it blinks, a cooling fault occurred; if it stays on, another type of component fault has occurred.
Activity (green): The enclosure is powered on.
One link controller card (LCC); an optional second LCC requires special application software for high-availability system configurations.
One autoranging power supply, optional second power supply; separate power cords.
Fan assembly containing three fans.
Address range (16-bit switch); LEDs to indicate address.
Disk slot bypass feature for use when a disk module has failed.
DB9 connectors and 0.3-meter (11.8 inches) “twinax” cable for connection to another enclosure using the same FC-AL. FibreVault options in Origin racks use 0.5 meter (19.7 inch) enclosure connection cables.
Automatic loopback of signals on the last enclosure in a chain, eliminating the need for a loopback connector or terminator.
Disk modules, redundant LCCs and power supplies, and the fan assembly can be replaced with no tools while the enclosures are powered on. An optional additional power supply and LCC and the multiple fans in the fan assembly provide high- availability configurations. If two LCCs and their cabling are installed, the disk drives support dual–port FC-AL interconnects.
| Note: Dual-port (dual-LCC) configurations require IRIS FailSafe or a customer specific software application to use these features fully. The standard IRIX operating system does not currently support high-availability dual-port FC-AL configurations. |
Up to 11 non-RAID FibreVault enclosures (or three RAID enclosures) can be mounted in a fibre channel rack. The deskside non-RAID chassis supports one FibreVault enclosure. The fibre channel RAID deskside holds one RAID (DPE) enclosure and one FibreVault RAID expansion (DAE). Enclosures are connected to a host fibre controller board with 10-meter or optional 25-meter copper cable or an optional 25-, 100-, or 300-meter optical cable used with a media interface adapter (MIA) at each end. FibreVaults used in Origin and Onyx2 system racks use a 3-meter (10 foot) copper cable to connect the host's controller board to the fibre channel enclosure.
The 19-inch fibre channel rack houses up to 11 Origin FibreVault enclosures, mounted horizontally. Figure 1-10 shows the fibre channel rack.
For further details on fibre enclosures, see Chapter 3, “Fibre Channel Storage,” for more discussion and additional illustrations.
Cables are included for connecting enclosures together in a rack to form one large disk storage loop.
For further details on FibreVault, see Chapter 3, “Fibre Channel Storage.”
Two deskside fibre channel “tower” systems are available:
The deskside FibreVault tower system holds one non-RAID fibre enclosure in a vertical orientation.
The deskside fibre channel RAID tower system holds one fibre channel RAID enclosure and one FibreVault RAID expansion enclosure mounted vertically.
Figure 1-11 shows the two types of fibre channel deskside systems.
Note that the fibre channel enclosures that mount in the fibre channel rack and deskside towers are exactly the same. The mounting orientation (vertical vs. horizontal) is the only difference. All internal components remove, replace, and operate in the same manner.
The non-RAID FibreVault enclosure can be mounted in the Origin or Onyx2 rack. The FibreVaults in Origin racks do not have redundant power options. The FibreVault enclosures install in the Origin or Onyx2 rack in the same location(s) and take up approximately the same space as the Origin Vault SCSI disk and tape unit.
A maximum of three FibreVaults may be installed in the Origin rack, one in the Onyx2 rack. All enclosure installations are done by Silicon Graphics trained and approved installers.
Normal reads and writes go through the dksc driver; configuration and environmental status use the devscsi driver.
For fibre channel RAID applications, an agent (ssmagent) provides the interface to the devscsi driver for configuration and environmental status. The agent can be configured to talk to a dedicated logical unit number (LUN) on the RAID. On top of the agent is a graphical user interface (GUI) and a command-line interface (CLI), which can be used for configuration of both RAID and non-RAID fibre arrays.
Figure 1-13 diagrams the software interface. The software driver implementation follows the private loop direct attach (PLDA) profile (see Figure 1-13).
For more information on using the administrative CLI and GUI, see Origin FibreVault and Fibre Channel RAID Administrator's Guide .
Chapter 5, “Configuring a Fibre Channel System,” explains how to use some of the commands for specific purposes.
The dksc and devscsi drivers are the same as those for previous implementations of SCSI. The existing SCSI request (high-to low-level) interface is unchanged, but extensions have been added:
dksc driver: The function of this driver is essentially unchanged from its SCSI-only ancestor. It interfaces to both the SCSI and fibre channel low-level drivers and implements access to basic disk I/O functions.
devscsi driver: Like its SCSI-only ancestor, this driver provides an interface to both the SCSI and fibre channel low-level drivers and implements SCSI pass-through command functionality. The devscsi driver can send arbitrary SCSI commands to a fibre channel disk just as arbitrary SCSI commands are sent to a SCSI disk; for example, a sequence of SCSI commands for downloading new firmware to a SCSI device.
scsiha: The driver scsiha allows access to all fibre channel–specific functions not implemented by dksc and devscsi for both device-targeted and controller-targeted commands. This interface is used for commands that address a fibre channel loop as a whole, and are not addressed to any specific device or do not require a response from a device. Examples are the Loop Initialization command (LIP) or an enable PBE command to a specific device.
The low-level driver provides resources for up to 1,000 concurrent outstanding commands.
Each fibre channel interface (XIO controller board) address is chosen during loop initialization This ensures that it does not conflict with other devices (including other controllers).
Each controller trys to get ID 125. If ID 125 is already taken, the controller takes the next- highest ID number available. This is automatic, so no system hostid is required.
[1] For information on the ANSI Fibre Channel standards, contact Global Engineering Documents, 15 Inverness Way East, Englewood, CO 80112 USA (303)-397-0271 or (800)-854-7179 (U.S. & Canada).