GB2341715A - Magnetic disk redundant array - Google Patents

Abstract

A plurality of magnetic disk drives (<B>301</B>, <B>302</B>, <B>303</B>) are configured to store machine readable data in a protected way such that data is recoverable in the event of a single disk failure. The array of disks is housed for application directly into an existing disk bay of a computer (<B>101</B>). The array is connectable to the computer as if it were a single conventional computer disk and the drives are controlled by an operating system on the computer as if they were a single storage volume.

Description

2341715 Data Storage The present invention relates to an array of magnetic

disks configured to store machine readable data in a protected way, such that data is recoverable in the event of disk failure.

Arrays configured to store machine readable data in a protected way are known and are often referred to as a redundant array of inexpensive disks, usually abbreviated to the acronym "RAID". Several RAID procedures are known and most of these share the approach of generating redundant data by an exclusive ORing process from which, in the event of any of the disks failing, all of the data can be reconstituted from the remaining operational disks.

When all of the disks are operational, the array is said to be working in its protected mode. In the event of one disk failure, the system may still remain operational, in that data may be read from the disks, but the data ceases to be protected and a further disk failure would result in data loss.

With a single disk failure the system is said to be working in an unprotected mode at which point an operator would be advised that disk replacement is required and that the lost data needs to be reconstituted. Thus, a disk would be physically removed, replaced and then the lost data would be reconstituted on to the new disk.

As personal computer systems and workstations become more powerful, allowing more sophisticated software applications to be executed and the degree of data storage available in such systems increases, with disks containing several gigabytes of data now becoming widely used, a greater demand has been created for the installation of protected systems using disk redundancy. Complete RAID subsystems may be purchased for external connection but a problem with such known systems is that the cost can be very prohibitive. In many situations, the cost of such a RAID system 2 tends to be higher than the cost of a personal computer system. Thus, there is a requirement for providing RAID protection at reduced cost.

Personal computer systems are usually housed in desktop units or tower units having spare bays allowing additional disks to be received. Thus, it is possible for many hard disk drives to be included within a tower housing and additional interface cards may be provided if required. Thereafter, it is possible for the RAID calculations to be effected by the resident host CPU, such that the additional extra cost is quite modest. However, a major problem with such a configuration is that a significant processor overhead is required in order to perform the RAID calculations, resulting in a severe degradation in overall system performance.

According to a first aspect of the present invention, there is provided a plurality of data storage devices configured to store machine readable data in a proteeted way such that data is recoverable in the event of a single device failure, wherein the devices are housed for application directly into an existing disk bay for a computer; the devices are connectable to a disk interface as if they were a single conventional storage volume; and said devices are controlled by an operating system installed on a computer as if they were a single storage volume.

In a preferred embodiment, the disks are interfaced to an IDE connection and three disks may be received in respective IDE connections.

Preferably, the array presents a SCSI interface to a host computer and the array may be configured to be housed in two or more five and one quarter inch drive bays.

According to a second aspect of the present invention, there is provided a method of equipping a personal computer with a plurality of data storage devices configured as a redundant array by interfacing said devices to conventional five and one quarter inch drive bays, such that protected machine readable data is recoverable in the event of a single disk failure, comprising the steps of supporting the array within an existing disk bay for a 3 computer, connecting the array to the computer as if it were a single conventional computer disk; and controlling said drives by an operating system installed on a computer as if it were a single storage volume.

The invention will now be described by way of example only, with reference to the accompanying drawings, in which:

Figure 1 shows a personal computer system; Figure 2 shows an array of disks being inserted into a computer system; Figure 3 details the array shown in Figure 2; Figure 4 shows an exploded view of the array identified in Figure 3; Figure 5 shows a rear face view of the array back plane; Figure 6 shows a circuit for implementing RAID calculations; and Figure 7 illustrates the removal of a damaged disk from the array; and Figure 8 shows an alternative embodiment for the extrusion identified in Figure 4.

A personal computer system is shown in Figure 1 in which a main system tower 101 supplies visual information to a visual display unit 102 and receives manual commands via a keyboard 103. The main system tower houses a central processing unit, memory circuits and other standard associated electronics as is well known in the art. The personal computer system may be an IBM PC type system, a Mackintosh system or any other computer type equipment used for individual use, possibly in a networked configuration. Alternatively, the main system tower 101 may constitute a network server, possibly running an appropriate server operating system, such as Windows NT server.

Tower 101 includes conventional five and one quarter inch disk bays.

Within these disk bays a plurality of devices have been mounted, including a three and a half inch floppy disk drive 105, a tape streamer 106, a CD ROM drive 107 and an array of magnetic disks 108, embodying the present invention.

4 Array 108 is detailed in Figure 2 and is shown being installed into the main system tower 101. The array 108 of magnetic disks is configured to store machine readable data in a protected way such that data is recoverable in the event of a single disk failure. The array of disks is housed for application directly into an existing disk bay of a computer, such as the main system tower 101. The array is connectable to the computer as if it were a single conventional computer disk and the array is operated by an operating system installed on the computer as if it were a single disk.

Each empty drive bay is protected by a removable plastic cover and unit 107 locates within an aperture equivalent to the width of two bays, requiring the removal of two such covers. The array includes a housing 201, locatable within the two bay aperture and towards its rear includes conventional power and data connectors; such that the housing as a whole is connected to the main system tower using a conventional SCSI connection.

Thus, the main system perceives the disk array as if it were a single disk and the operating system, executed by the main system, controls the operation of the array using equivalent commands to those required for the operation of a single storage volume.

The array 107 is detailed in Figure 3 and contains a total of three IDE drives 301, 302 and 303. An exploded view of the array is illustrated in Figure 4, which shows each of the individual IDE drives 301, 302 and 303 being supported by aluminium extrusions, in the form of a left extrusion 401 and a right extrusion 402. These extrusions hold the disk drives 301, 302 and 303 firmly in place and facilitate the removal and replacement of individual disk drives when disk failure occurs.

Disk drives 301, 302 and 303 are located in relatively close proximity and in order to maintain preferred operational temperatures, an electric fan 403 is positioned between the front of the disk drives and a front housing 404. In this respect, the main front housing includes ventilation grilles 405.

Each IDE drive 301, 302 and 303 locates within a conventional IDE socket 406, 407, 408, in addition to respective power supply sockets 409, 410, 411. Thus, from the perspective of each IDE drive, the physical drives are located into sockets substantially similar to those found on an IDE bus of a standard computer system.

RAID calculations are performed within the device itself, using conventional hardware RAID circuitry mounted on circuit board 412, having electrical connections to the back plane circuit board 413. Right extrusion 402 defines a cavity 414, configured to receive circuit board 412. The extrusions 401 and 402 are held in position by an upper plate 415 and a lower plate 416, secured by appropriate bolts 417.

The rear face of back plane 413 is illustrated in Figure 5. The back plane includes a conventional SCSI socket 501 and a power supply socket 502. The array therefore presents itself to the main system as a single disk drive, requiring a single disk drive connection via SCSI interface 501.

Back plane 413 also includes rows of holes 503 to facilitate ventilation of the disks. Thus, cooling air is brought in through ventilation holes 405, blown between the disks 301, 302 and 303 and then exits through holes 503.

The circuit implemented on board 412 is illustrated diagramatically in Figure 6. The circuit includes a central processing unit 601 which communicates with an input/output circuit 602 via a CPU bridge 603. In addition, operation of CPU 601 is controlled by a CPU mode select circuit 604. Power from the housing is directed to a three volt supply regulating circuit 605, arranged to supply power to operational circuits via supply rails.

The CPU 601 receives data relating to the operational environment from an environmental detecting circuit 606. This information may be received directly, as shown in Figure 6, or it may be directed via other control circuitry to allow combined environmental information to be returned to the CPU 601.

6 Further output circuitry includes IDE controllers 607 and 608 and a SCSI controller 609. These circuits communicate with the back plane sockets via a one hundred and eighty way connector 610.

Input/output circuit 602 supplies driving current to six LED's 701, 702, 703, 704, 705 and 706 shown in Figure 7. Each of these LED's is visible by means of respective holes 711, 712, 713, 714, 715 and 716 in the front panel 404. Each LED is a Hewlett Packard HSMF-C655 and actually includes a green LED and a red LED which may be operated independently.

LED 701 indicates the overall operational integrity of the system and primarily confirms that CPU 601 is operating correctly. Thus, when the system is fully operational, LED 701 is illuminated green. Alternatively, if faults have been detected within the controller, LED 701 is illuminated red.

LED 702 represents the environmental monitoring status and is primarily concerned with operational temperature. Environmental circuit 606 includes a temperature sensor and a fault condition is generated if this sensor detects that operational temperatures have become excessive. In addition, a tachometer is associated with fan 403 and a fault condition is generated if this detects that rotation of the fan has ceased. Malfunction of fan 403 represents a serious problem in that this could result in all three drives being permanently damaged such that no protection is offered by the RAID configuration. The system also detects the presence of appropriate voltages on voltage supply rails, as supplied by power supply unit 605 in addition to detecting appropriate terminator power on the SCSI bus.

When the supply rail voltages are correct, SCSI terminator power is correct, the fan is operational and the system is working at its optimal operational temperature, LED 702 is illuminated green. If the system encounters problems and diverges from its preferred operational characteristics, such a condition is detected and LED 702 is illuminated orange. Under these conditions, further operation of the system is permitted but warnings may be generated to the effect that a job should be closed 7 down and that the device should be investigated. If problems continue and the situation worsens, particularly if the operational temperature becomes very high, LED 702 is illuminated red. Under these conditions, power to the drives is removed and an error condition is generated such that further access to the drives is not permitted.

LED 703 indicates that the SCSI connection is fully operational by being illuminated green. Furthermore, when the SCSI bus is actually in use, LED 703 is illuminated orange.

LED's 704, 705 and 706 represents operational characteristics of the individual drives 301, 302 and 303 respectively. When the drives are operational, the LED's are illuminated green and then illuminated orange when the actual data transfer takes place. Furthermore, if a disk error is detected, to the effect that an individual disk has failed, its respective LED is illuminated red.

In response to a single disk failure, it is preferable for the system to be placed off-line and for the damaged disk to be replaced immediately so that the lost data may be reconstituted and the system returned to protected mode operation. In order to replace a disk, the front panel is removed, an operation facilitated by the front panel 404 being retained simply to the main housing by means of an interference connection. Having removed the front panel 404 it is restrained by wires 717 required for supplying electrical power to fan 403.

The disk drives include tapped holes towards their front-right comer and each of said tapped holes receives a threaded stud 719. Stud 719 allows its respective disk 301 to 303 to be removed by the application of a stud hook 720. Force is applied in the direction of arrow 721, thereby forcing the respective disk drive away from its IDE and data sockets, such as sockets 406 and 409 etc.

An alternative embodiment is illustrated in Figure 8. In this embodiment, side panels and a base panel are fabricated as a single 8 extrusion 801. The housing is then completed by the application of a top panel 802. The top panel 802 is secured to the lower extrusion 801 by means of bolts 803 and circuitry held within the extrusion is further secured by an adhesive clip 804.

9

Claims (22)

1. Claims

   I A plurality of data storage devices configured to store machine readable data in a protected way such that data is recoverable in the event of a single device failure, wherein the devices are housed for application directly into an existing disk bay for a computer; the devices are connectable to a disk interface as if they were a single conventional storage volume; and said devices are controlled by an operating system installed on a computer as if they were a single storage volume.
2. 2. Data storage devices according to claim 1, wherein said storage devices are magnetic disk drives.
3. 3. Data storage devices according to claim 2, wherein the magnetic disks are interfaced to an IDE connection.
4. 4. Data storage devices according to claim 3, wherein three disks are received in respective IDE connections.
5. 5. Data storage devices according to any of claims 1 to 3, wherein said devices present a SCSI interface to a host computer.
6. 6. Data storage device according to any of claims I to 5, configured to be housed in two or more five and one quarter inch drive bays.
7. 7. Data storage devices according to any of claims I to 6, including means for detecting when said devices are operating in non- ideal conditions.
8. 8. Data storage devices according to claim 7, including means for detecting when said devices are operating at excessive temperatures.
9. 9. Data storage devices according to claim 7 or claim 8, including means for detecting non-operation of a cooling fan.
10. 10. Data storage devices according to claim 7 or claim 8, including means for directly detecting an excessive operational temperature.
11. 11. Data storage devices according to any of claims 7 to 10, including means for removing drive power to said devices upon detecting a non-ideal operating condition.
12. 12. Data storage devices according to any of claims I to 11, including a detachable front panel and a cooling fan secured to said front panel, including ventilation openings arranged to direct a cooling air- stream between the individual devices.
13. 13. A plurality of data storage devices according to any of claims I to 12, wherein said devices are connectable in a computer housing and the devices are controlled by the operating system of said computer.
14. 14. A method of equipping a personal computer with a plurality of data storage devices configured as a redundant array by interfacing said devices to conventional five and one quarter inch drive bays, such that protected machine readable data is recoverable in the event of a single disk failure, comprising the steps of supporting the array within an existing disk bay for a computer; --i connecting the array to the computer as if it were a single conventional computer disk; and controlling said drives by an operating system installed on a computer as if it were a single storage volume.
15. 15. A method according to claim 14, wherein said data storage devices are magnetic disk drives.
16. 16. A method according to claim 15, wherein said magnetic disk drives are interfaced to an IDE connection.
17. 17. A method according to claim 16, wherein three disks are received in respective IDE connections.
18. 18. A method according to any of claims 14 to 17, wherein said devices present a SCSI interface to a host computer.
19. 19. A method according to any of claims 14 to 18, wherein said devices are housed in two or more five and one quarter inch drive bays.
20. 20. A method according to any of claims 14 to 19, wherein non ideal operating conditions for said devices are detected.
21. 21. A plurality of data storage devices substantially as herein described with reference to the accompanying Figures.
22. 22. A method of equipping a personal computer substantially as herein described with reference to the accompanying Figures.