When an adapter fails, a a replacement can be inserted and connected to the existing set of drives. Dell supports adapter swapping only when all the attached disks are migrated to a new adapter that has a clear configuration.
Array
An array of disk drives combines the storage space on the disk drives into a single segment of storage space. The RAID controller can group disk drives on one or more SCSI channels into an array. A hot spare drive does not actively participate in an array.
Array Spanning
Array spanning by a logical drive combines storage space in two arrays of disk drives into a single, contiguous storage space in a logical drive. The logical drives can span consecutively numbered arrays that each consist of the same number of disk drives. Array spanning promotes RAID level 1 to RAID level 10 and RAID level 5 to RAID level 50.
Asynchronous Operations
Operations that bear no relationship to each other in time and can overlap. The concept of asynchronous I/O operations is central to independent access arrays in throughput-intensive applications.
(Basic Input/Output System) The part of the operating system in an IBM PC-compatible system that provides the lowest level interface to peripheral devices. The BIOS is stored in ROM in every IBM or compatible PC. BIOS also refers to the Basic Input/Input Output System of other "intelligent" devices, such as RAID controllers.
Specifies that reads are buffered in cache memory, but does not override the other cache policies, such as read ahead or write.
Caching
The process of utilizing a high speed memory buffer, referred to as a "cache", in order to speed up the overall read or write performance. This cache can be accessed at a higher speed than a disk subsystem. To improve read performance, the cache usually contains the most recently accessed data, as well as data from adjacent disk sectors. To improve write performance, the cache may temporarily store data in accordance with its write back policies. See the definition of Write-Back for more information.
Channel
An electrical path for the transfer of data and control information between a disk and a disk adapter. A channel can also be referred to as a "bus", such as a SCSI bus.
Clearing
In the BIOS Configuration Utility, the option used to delete data from physical drives.
Cold Swap
The replacement or exchange of a device in a system after powering down the system. In reference to disk subsystems, a cold swap requires that you turn the power off before replacing a defective hard drive.
Consistency Check
An examination of the disk system to determine if all conditions are valid for the specified configuration (such as parity).
The amount of data per unit time moved through a channel. For disk I/O, bandwidth is expressed in megabytes per second (MB/s).
Degraded Drive
A disk drive that has become non-functional or has decreased in performance.
Direct I/O
Specifies that reads are not buffered in cache memory, but does not override the other cache policies, such as read ahead or write.
Disk
A non-volatile, randomly addressable, rewriteable mass storage device, including both rotating magnetic and optical storage devices and solid-state storage devices, or non-volatile electronic storage elements.
Disk Array
A collection of disks from one or more disk subsystems controlled by array management software. The array management software controls the disks and presents them to the array operating environment as a virtual disk.
Disk Duplexing
A variation on disk mirroring in which a second disk adapter or host adapter and redundant disk drives are present.
Disk Mirroring
Writing duplicate data to more than one (usually two) hard disks to protect against data loss in the event of device failure. Disk mirroring is a common feature of RAID systems.
Disk Spanning
The process of creating one logical drive composed of multiple arrays. Spanning is used to create complex RAID sets, such as RAID levels 10 and 50. Spanning utilizes striping to distribute data across all member disk drives.
Disk Striping
A type of disk array mapping. Consecutive stripes of data are mapped round-robin to consecutive array members. A striped array (RAID level 0) provides high I/O performance at low cost, but provides less data reliability than any member disk.
Disk Subsystem
A collection of disks and the hardware that controls them and connects them to one or more controllers.
The hardware can include an intelligent adapter, or the disks can attach directly to a system I/O bus adapter.
Double Buffering
A technique that achieves maximum data transfer bandwidth by constantly keeping two I/O requests outstanding
for adjacent data. A software component begins a double-buffered I/O stream by issuing two requests in rapid sequence.
Thereafter, each time an I/O request completes, another is immediately issued. If the disk subsystem can process requests
fast enough, double buffering allows data to be transferred at the full-volume transfer rate.
Drive Roaming
Drive roaming occurs when the hard drives are changed to different target IDs or channels on the same controller.
(A single-channel adapter can perform drive roaming.) When the drives are placed on different channels or target IDs,
the controller detects the RAID configuration from the configuration information on the drives.
Configuration data is saved in both NVRAM on the RAID controller and in the hard drives attached to the controller.
This maintains the integrity of the data on each drive, even if the drives have changed their target ID.
A drive that has ceased to function, that consistently functions improperly, or that is inaccessible.
Fast SCSI
A variant on the SCSI-2 bus. It uses the same 8-bit bus as the original SCSI-1 but runs at up to 10MBs (double the speed of SCSI-1).
Firmware
Software stored in read-only memory (ROM) or Programmable ROM (PROM). Firmware is often responsible for the behavior of a system when it is first turned on. A typical example would be a monitor program in a system that loads the full operating system from disk or from a network and then passes control to the operating system.
Format
The process of writing a specific value to all data fields on a physical drive, (hard drive), to map out unreadable or bad sectors. Because most hard drives are formatted when manufactured, formatting is usually done only if a hard disk generates many media errors.
Any system to which disks are directly attached, (not attached remotely.) Mainframes, workstations, and personal computers can all be considered host systems.
Hot Spare
An idle, powered on, stand-by drive ready for immediate use in case of disk failure. It does not contain any user data. Up to eight disk drives can be assigned as hot spares for an adapter. A hot spare can be dedicated to a single redundant array or it can be part of the global hot-spare pool for all arrays controlled by the adapter.
When a disk fails, the controllers' firmware automatically replaces and rebuilds the data from the failed drive to the hot spare. Data can be rebuilt only from logical drives with redundancy (RAID levels 1, 5, 10, or 50; not RAID 0), and the hot spare must have sufficient capacity. The system administrator can replace the failed disk drive and designate the replacement disk drive as a new hot spare.
Hot Swap
The manual replacement of a failed drive while the disk subsystem is running (performing its normal functions).
Hot Swap Disk Drive
Hot swap drives allow a system administrator to replace a failed disk drive in a system without powering down the system and suspending services. The hot swap drive is pulled from its slot in the drive cage; all power and cabling connections are integrated into the drive enclosure backplane. The replacement hot swap drive can then slide into the slot. Hot swapping only works for RAID 1, 5, and 10 configurations.
The process of writing zeros to the data fields of a logical drive and, in fault tolerant RAID levels, generating the corresponding parity to put the logical drive in a Ready state. Initializing erases previous data and generates parity so that the logical drive will pass a consistency check. Arrays will work without initializing, but they can fail a consistency check because the parity fields have not been generated.
I/O Driver
A host system software component (usually part of the operating system) that controls the operation of peripheral adapters attached to the host system. I/O drivers communicate between applications and I/O devices and in some cases participate in data transfers.
A complete or partial representation of a logical array. The storage space in a logical drive is spread across all the physical drives in the array or spanned arrays.
Each RAID controller can be configured with up to forty logical drives in any combination of sizes.
Configure at least one logical drive for each array. A logical drive can be in one of three states:
Online: All participating disk drives are online.
Degraded: (Also "Critical") a single drive in a redundant array (not RAID 0) is not online. Data loss can occur if a second disk drive fails.
Offline: Two or more drives in a redundant array (not RAID 0) or one or more drives in a RAID 0 array are not online.
I/O operations can be performed only with logical drives that are online or degraded.
The relation between multiple data addressing schemes, especially conversions between member disk block addresses and block addresses of the virtual disks presented to the operating environment by array management software.
MB
A megabyte; an abbreviation for 1,000,000 (10 to the sixth power) bytes.
Mirroring
The process of providing complete redundancy using two disk drives, by maintaining an exact copy of one disk drive's data on the second disk drive. If one disk drive fails, the contents of the other disk drive can be used to maintain the integrity of the system and to reconstruct the failed drives.
Multi-threaded
Having multiple concurrent or pseudo-concurrent execution sequences. Multi-threaded processes allow throughput-intensive applications to efficiently use resources to increase I/O performance.
Capacity expansion by adding volume or another hard drive, while the host system is accessible and/or active.
Operating Environment
An operating environment can include the host system where an array is attached, any I/O buses and adapters, the host operating system and any additional software required to operate the array. For host-based arrays, the operating environment includes I/O driver software for the member disks but does not include array management software, which is regarded as part of the array itself.
An extra bit added to a byte or word to reveal errors in storage (in RAM or disk) or transmission. Parity is used to generate a set of redundancy data from two or more parent data sets. The redundancy data can be used to reconstruct one of the parent data sets. However, parity data does not fully duplicate the parent data sets. In RAID, this method is applied to entire drives or stripes across all disk drives in an array. Parity consists of dedicated parity, in which the parity of the data on two or more drives is stored on an additional drive, and distributed parity, in which the parity data are distributed among all the drives in the system. If a single drive fails, it can be rebuilt from the parity of the applicable data on the remaining drives.
Partition
A complete or partial representation of a logical drive, usually represented to a user by an operating system as a physical disk. Also called a logical volume.
PERC 4e/Di
The DellTM PERC 4e/Di consists of an LSI 1030 chip on the motherboard to offer RAID control capabilities. PERC 4e/Di supports all dual-ended and LVD SCSI devices on Ultra320 and Wide SCSI channels with data transfer rates up to 320 MB/s (Megabytes per second).
PERC 4e/Di provides reliability, high performance, and fault-tolerant disk subsystem management. It is an ideal RAID solution for the internal storage of Dell's workgroup, departmental, and enterprise systems. PERC 4e/Di offers a cost-effective way to implement RAID in a server and provides reliability, high performance, and fault-tolerant disk subsystem management.
Physical Disk
A hard drive. A hard drive consists of one or more rigid magnetic discs rotating about a central axle, with associated read/write heads and electronics. A physical disk is used to store information, (data), in a non-volatile and randomly accessible memory space.
Physical Disk Roaming
The ability of adapters to detect when hard drives have been moved to different slots in the storage enclosure, such as after a hot swap.
Protocol
A set of formal rules describing how to transmit data, generally across a network or when communicating with storage sub-systems. Low-level protocols define the electrical and physical standards to be observed, bit- and byte-ordering, and the transmission and error detection and correction of the bit stream. High-level protocols deal with the data formatting, including the message syntax, the terminal to system dialogue, character sets, sequencing of messages, etc.
Redundant Array of Independent Disks (originally Redundant Array of Inexpensive Disks) is an array of multiple small, independent hard drives that yields performance exceeding that of a Single Large Expensive Disk (SLED). A RAID disk subsystem can improve I/O performance relative to a system using only a single drive. The RAID array appears to the controller as a single storage unit. I/O is expedited because several disks can be accessed simultaneously. Redundant RAID levels (RAID levels 1, 5, 10, and 50), provide data protection.
RAID Level Migration
RAID level migration (RLM) changes the array from one RAID level to another. It is used to move between optimal RAID levels. You can perform a RLM while the system continues to run, without having to reboot. This avoids downtime and keeps data available to users.
RAID Levels
A style of redundancy applied to an array. It can increase the performance of a logical drive though it may decrease usable capacity. Each logical array must have a RAID level assigned to it.
Read-Ahead
A memory caching capability in some adapters that allows them to read sequentially ahead of requested data and store the additional data in cache memory, anticipating that the additional data will be needed soon. Read-Ahead supplies sequential data faster, but is not as effective when accessing random data.
Ready State
A condition in which a workable hard drive is neither online nor a hot spare and is available to add to an array or to designate as a hot spare.
Rebuild
The regeneration of all data to a replacement disk from a failed disk in a logical drive with a RAID level 1, 5, 10 or 50 array. A disk rebuild normally occurs without interrupting normal operations on the affected logical drive, though some degradation of performance of the disk subsystem can occur.
Rebuild Rate
The percentage of CPU resources devoted to rebuilding.
Reconstruct
The act of remaking a logical drive after changing RAID levels or adding a physical drive to an existing array.
Redundancy
The provision of multiple interchangeable components to perform a single function to cope with failures and errors. Common forms of hardware redundancy are disk mirroring, implementations of parity disks or distributed parity.
Replacement Disk
A hard drive replacing a failed member disk in a RAID array.
Replacement Unit
A component or collection of components in a system or subsystem that is always replaced as a unit when any part of the collection fails. Typical replacement units in a disk subsystem include disks, adapter logic boards, power supplies and cables.
(Small Computer System Interface) A processor-independent standard for system-level interfacing between a system and intelligent devices, such as hard disks, floppy disks, CD-ROM, printers, and scanners. SCSI can connect up to 15 devices to a single adapter (or host adapter) on the system's bus. SCSI transfers 8, 16 or 32 bits in parallel and can operate in either asynchronous or synchronous modes. The synchronous transfer rate is up to 320 MB/s.
The original standard is now called SCSI-1 to distinguish it from SCSI-2 and SCSI-3, which include specifications of Wide SCSI (a 16-bit bus) and Fast SCSI (10 MB/s transfer.) Ultra 160M SCSI is a subset of Ultra320 SCSI and allows a maximum throughput of 160 MB/s, which is twice as fast as Wide Ultra2 SCSI. Ultra320 allows a maximum throughput of 320 MB/s.
SCSI Channel
The RAID controller controls hard drives via 320M SCSI buses (channels) over which the system transfers data in either LVD or 320M SCSI modes. Each adapter controls two SCSI channels.
SCSI Disk Status
A SCSI disk drive (physical drive) can be in one of these four states:
Online: A powered-on and operational disk.
Hot Spare: A powered-on, stand-by disk ready for use if another disk fails.
Not Responding: The disk is not present, not powered-on, or has failed.
Rebuild: A disk to which one or more logical drives is restoring data.
SCSI ID
Each SCSI device on the RAID controller SCSI bus must have a different SCSI address number (Target ID or TID) from 0 to 15. Notice that one ID is used by the SCSI controller, usually ID 7. Set the SCSI ID switch on each disk drive to the correct SCSI address. See the RAID controller documentation, chassis labels or disk enclosure documentation for the correct switch settings.
Spare
A hard drive available to back up the data of other drives.
Striping
Segmentation of logically sequential data, such as a single file, so that segments can be written to multiple physical devices in a round-robin fashion. This technique is useful if the processor can read or write data faster than a single disk can supply or accept it. While data is being transferred from the first disk, the second disk can locate the next segment. Data striping is used in some modern databases and in certain RAID devices.
Stripe Size
The amount of data written to each disk. Also called "stripe depth." You can specify stripe sizes of 2 KB, 4 KB, 8 KB, 16 KB, 32 KB, 64 KB, and 128 KB for each logical drive. A larger stripe size produces improved read performance, especially if most of the reads are sequential. For mostly random reads, select a smaller stripe size.
Stripe Width
The number of disk drives across which the data are striped.
A resistor connected to a signal wire in a bus or network for impedance matching to prevent reflections, e.g., a resistor connected across signal wires at the end of a SCSI cable.
A variant on the SCSI-2 interface. Wide SCSI uses a 16-bit bus, double the width of the original SCSI-1.
Write-Back
In Write-Back caching mode, the controller sends a data transfer completion signal to the host when the controller cache has received all the data in a disk write transaction. Data are written to the disk subsystem in accordance with policies set up by the controller. These policies include the amount of dirty/clean cache lines, the number of cache lines available, elapsed time from the last cache flush, and others.
Write-Through
In Write Through caching mode, the controller sends a data transfer completion signal to the host when the disk subsystem has received all the data in a transaction. The controller cache is not used.
