JPH04279924A - Constitution control system for array disk device - Google Patents

Abstract

PURPOSE:To efficiently perform array re-constitution in accordance with the change of array constitution such as alternation and exchange, etc., flexibly without fixing the master-slave relation of plural disk devices, byte or bit allocation, and furthermore, spare machine allocation for the constitution control system of an array disk device in which the plural disk devices are comprised in array constitution in parallel and in which information is recorded or reproduced by processing data in parallel. CONSTITUTION:An attribute table 9 on which the master-slave attribute of a mater disk that becomes reference for the rotary synchronization of all disk devices and a slave disk rotated following the master disk is set, and an attribute table 10 on which the divisional allocation of the data to be processed in parallel, the allocation of parity information, and spare allocation are set are provided, and when a fault occurs in the disk, the array re-constitution of alternate processing is performed according to attribute information, and the information is changed to the attribute information after the re-constitution, and the next array re-constitution is awaited.

Description

[Detailed description of the invention]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a configuration control method for an array disk device in which a plurality of disk devices are arranged in parallel and data is processed in parallel to record or reproduce information. In magnetic disk devices or optical disk devices that are the input/output devices of computer systems, an array disk device is installed in which multiple disk devices are configured in an array (parallel), and data transfer is processed in parallel to facilitate information transfer. The aim is to shorten the processing time involved, speed up information processing, and increase the capacity of stored information.

[0002] In such a disk array device, in order to synchronize the rotations of multiple disk devices, a master-slave relationship is set in which a specific disk device is the master and other disk devices are slaves, and the slave rotation follows the master rotation. I'm letting you do it. Furthermore, the data to be recorded and reproduced is divided into bytes or bits including parity and allocated to each disk device, and a spare disk device is also provided to cope with the occurrence of a failure.

[0003] The master-slave relationship and bit allocation for rotational synchronization, as well as the allocation of standby machines, are fixedly set at the time of system start-up. It is desirable to be able to freely change the master-slave relationship and bit allocation as necessary.

0004

[Prior Art] In a conventional array disk device,
As shown in FIG. 8, in order to speed up data processing, a plurality of disk devices 8-a to 8-i each having disk control circuits 7-a to 7-i are configured in parallel within the device. The recording or reproduction of information is parallelly divided into disk devices 8-a to 8-i via the array control circuit section 6, and high-speed processing is performed simultaneously.

The array control circuit section 6 includes array control circuits 12-a to 12-1 corresponding to each disk device 8-a to 8-i.
2-i is provided, and a transfer circuit within the device transfers data to and from the host device via the interface via the common data control circuit 13. In the plurality of disk devices 8-a to 8-i configured in an array in this way, it is necessary to synchronize the rotations of all the disk devices, and for this rotation synchronization, one disk is used as a reference master disk. A rotation synchronization system is adopted in which a device, for example, a disk device 8-a, is set as a master, and the other disk devices 8-b to 8-i are made slaves and follow the rotation of the master.

[0006] Furthermore, in recording and reproducing data, transferred data is transferred to disk devices 8-a to 8-i arranged in an array.
Parallel partitioning of data is performed logically and physically in accordance with the data, and the partitioned data is ranked and assigned to each disk device 8-a to 8-i in an array configuration. Figure 8
In this case, bits 0 to 8-h are stored in disk devices 8-a to 8-h.
7 is assigned to each, and a parity bit is assigned to the disk device 8-i.

Furthermore, in the conventional array disk device, a plurality of disk devices 8-a to 8-i configured in an array are used.
A spare disk device 8-j is provided in order to automatically or manually replace the disk device with a spare disk device if a failure occurs in a part of the disk device. Disk device 8-a~8
-i has a failure, a process is taken in which the failed disk device is disconnected and the bit allocation is replaced with a spare disk device 8-j.

Specifically, the multiplexer 14 switches and connects the array control circuit of the failed disk device to the spare array backup control circuit 12-j.

[0009]

[Problem to be Solved by the Invention] However, in such a conventional control method for an array disk device,
Because the master-dependent settings associated with rotational synchronization of multiple disk devices and the control attributes for data division and sequencing were fixed, when a failure occurs in the master disk device or when a disk device in which a failure occurs, After removing and repairing the
When incorporating it into an array disk device, it is necessary to return it to the same position as before the exchange, with the same master-slave relationship and bit allocation attributes as before the exchange, both logically and physically.

[0010]Furthermore, if a failure occurs in a disk device and the process is being replaced with a spare disk device, there will be no spare disk device available in case the failure occurs again. Furthermore, if the disk device is returned to its original state after repair, it is necessary to transfer the data recorded on the spare disk device during replacement to the disk device after repair, and the array configuration before replacement must be transferred. There was a problem in that it took a long processing time to restore the state.

The present invention has been made in view of the above-mentioned conventional problems, and it is possible to change array configurations such as alternation or replacement without fixing the master-slave relationship of a plurality of disk devices, bit allocation, and spare device allocation. An object of the present invention is to provide a configuration control method for an array disk device that can efficiently perform array reconfiguration that flexibly responds to changes.

0012

[Means for Solving the Problems] FIG. 1 is a diagram illustrating the principle of the present invention. First, the present invention provides a plurality of disk devices 8-a to 8-a.
The present invention is directed to a configuration control method for an array disk device that records or reproduces information by configuring 8-j in a parallel array and processing data in parallel and in a sequential order.

In the present invention, with respect to such a configuration control system of an array disk device, the array control circuit 6 is provided with a master disk that serves as a reference for rotationally synchronizing all disk devices, and a slave disk that rotates following the master disk. A first attribute table 9 in which main dependencies are set, a second attribute table 10 in which divisional allocation of data to be processed in parallel, parity information allocation, and standby machine allocation are set, and first and second attribute tables 9 and 10. The present invention is characterized in that a configuration control means 11 is provided for configuring the disk devices 8-a to 8-j into an array according to attribute information and executing parallel division processing of data.

[0014] Here, the configuration control means 11 is changed based on the replacement or replacement when the disk devices 8-a to 8-j or the disk control circuits 7-a to 7-j are replaced or replaced in response to a failure. First and second attribute tables 9, 1
The array configuration of the disk devices 8-a to 8-j is reconfigured according to the attribute information of 0. In addition, one or more disk devices 8-a to 8-j or disk control circuit 7- may be damaged due to failure or the like.
When a to 7-j are excluded from the array configuration, the attribute information in the first and second attribute tables 9 and 10 is changed to attribute information that reduces the bit configuration by the remaining disk devices, and the configuration control means 11 The array configuration is reconfigured based on the attribute information whose bit configuration has been reduced.

[0015]

[Operation] According to the array disk device configuration control method of the present invention, in the array disk device replacement process, a master-slave relationship between a master disk and a slave disk is established for rotationally synchronizing a plurality of disk devices configured in an array. , data disk allocation, and backup device allocation are not fixed, but are configured as a rewritable attribute table, and the attribute information of the table can be arbitrarily changed at the time of replacement or replacement. It facilitates replacement processing and replacement of disk devices, and when a failed disk device is returned to the device after repair, the replacement process is performed by setting the disk device that will be installed after the repair is completed as a new spare disk device. It is not necessary to return data from the disk device, and reliability and maintainability can be improved.

[0016]

Embodiment FIG. 2 is a block diagram of an embodiment of an array disk device to which the configuration control method of the present invention is applied. In FIG. 2, 1 is a central processing unit as a host device, 2 is a transfer control device, and 3 is an array disk device.

In this embodiment, the array disk device 3 transfers data to and from the central processing unit 1 in byte units (one byte consists of 8 bits, and one parity bit is added). Let's take as an example a case where data is divided bit by bit in parallel and recording and playback are processed in parallel at the same time. Data transferred from the central processing unit 1 passes through the transmitting/receiving circuit 4 and the transfer control circuit 5 of the array disk device 3, and is transferred to nine disk devices 8-a to 8-i configured in an array by the array control circuit section 6. The data is divided into bits in parallel corresponding to each other, and recorded on array disks 8-a to 8-i via disk control circuits 7a to 7i.

In the case of data reproduction, the reverse route of recording and transfer is followed, and the data recorded bit by bit in the disk devices 8-a to 8-i is converted into bytes through the array control circuit section 6, and transfer control is performed. The data is transferred to the central processing unit 1 via the circuit 5, the transmitting/receiving circuit 4, and the transfer control device 2 on the upper side. Furthermore, a spare disk control circuit 7-j and a spare disk device 8-j are provided for replacement processing in the event that a failure occurs in the disk control circuits 7-a to 7-i and/or the disk devices 8-a to 8-i. It is being

FIG. 3 is a block diagram showing a specific embodiment of the array control circuit section 6 of FIG. 2 together with the disk device side. The array control circuit section 6 in FIG.
-j and a data bus switching circuit 15.

Array control circuit section 6, disk control circuit 7
-a to 7-i and disk devices 8-a to 8-i, in order to process parallel divided data to disk devices 8-a to 8-i simultaneously, all disk devices 8-
It is necessary to synchronize the rotations of a to 8-i. Therefore, the master-slave relationship between the master disk device and the slave disk device is set according to the purpose of use when the disk is installed, and the first attribute table 9 is set in the common data control circuit 13.
is stored as .

In this embodiment, as shown in FIG. 4(a),
The first attribute table 9 includes disk device machine numbers #1 to #7,
Attribute information of master M and slave S is set corresponding to #P and #SP. Here, disk devices 8-a to 8
-j originally has both functions as a master disk and a slave disk, and based on the attribute information of the master-slave relationship in the first attribute table 9 shown in FIG. 4(a), the disk device 8 with machine number #0 -a is set as the master disk, and the remaining disk devices 8-b to #7, #P, and #SP are set as the master disk.
8-j is used as a slave disk to follow the rotation of the master disk, and the rotation of the entire disk is synchronized.

Furthermore, in the array control circuit section 6, the transfer data is transferred to the disk devices 8-a to 8-8, which constitute the array.
The data is divided into bit units (or byte block units) corresponding to i and bit allocation is performed, and the divided data is controlled to be recorded and reproduced for each corresponding disk. The bit assignments for the disk drives 8-a to 8-i are stored in the second attribute table 10 of the common data control circuit.

The second attribute table 10 is shown in FIG. 4(b), for example.
As shown in the figure, bit attributes are set corresponding to machine numbers #0 to #7 and #P, and an attribute as a standby machine is set for machine number #SP. The contents of the first and second attribute tables 9 and 10 provided in the common data control circuit 6 can be changed manually or automatically, and a configuration command is given to the configuration control unit 11 in the common data control circuit 6. Then, according to the attribute information in the attribute tables 9 and 10 at that time, the bit allocation to the array control circuits 12-a to 12-j is changed and the data bus switching circuit 15 switches the disk devices for replacement processing.

Next, referring to FIG. 5, the operation of replacement processing when a disk device fails will be explained. Assume that a failure occurs in the disk device 8-b while the array configuration is in accordance with the master-slave relationship and bit assignment shown in FIG. 3. Upon this failure detection, the data bus switching circuit 15 in FIG. 3 performs bus disconnection, and the failed disk device 8-b with device number #1 is separated from the array configuration, and the spare disk device 8-j with device number #SP is disconnected from the array configuration. A replacement process is performed in which the bus is connected to the failed bus.

With this replacement process, the master-slave relationship in the first attribute table 9 in FIG. 4(a) is not changed, but the bit attribute "B" of machine number #1 in the second attribute table 10 in FIG. 4(b)
it1" is cleared, and the attribute "spare" of machine number #SP is changed to the bit attribute "Bit1." For this reason, after replacement,
The bit data previously allocated to the disk device 8-b with machine number #1 is transferred to the spare disk device 8-j.
The array function continues by operating as a bit 1 assigned disk.

The failed disk device 8-b with machine number #1 is removed from the array disk device 3, and is returned to the disk array device 3 after being repaired. At the time of this return, "spare" attribute information is written in the bit attribute of machine number #1 in the second attribute table 10 of FIG. 4(b) that has been cleared by removal. Therefore, after inserting the disk drive 8-b of machine number #1, which has been repaired,
will now stand by as a new spare disk device.

[0027] This means that when a failed disk device 8-b is
When installing the device after repair, there is no need to transfer the data recorded in the spare disk device 8-j that is undergoing replacement processing to the repaired disk device 8-b, and processing work related to data restoration can be reduced. There are benefits. on the other hand,
If the disk device 8-a with machine number #0 that has been set as a master disk fails, the first
By changing the master/slave attribute of the spare disk device 8-j with machine number #SP in the attribute table 9 to master attribute "M," other slave disks follow the spare disk device 8-j that has become the master disk. It can be rotated.

Furthermore, the replacement processing in the present invention is performed not only when the disk drives 8-a to 8-j fail, but also when the disk control circuits 7-a to 7-j correspond to the disk drives 8-a to 8-j. Since the disk control circuit 7 is provided separately, the disk control circuit 7
Even when a failure in -a to 7-j is detected, the process of automatically or manually replacing the disk control circuit with a spare circuit can be performed in the same way.

Specifically, the disc control circuit 7- in FIG.
A switching circuit similar to the data bus switching circuit 15 is added between a to 7-j and disk devices 8-a to 8-j, and disk control circuits 7-a to 7- are provided in the common data control circuit 13.
The replacement process may be performed by switching the front and rear switching circuits according to the contents of the second attribute table 10 regarding j. FIG. 6 is an explanatory diagram showing another operating state of the replacement process of the present invention, which is characterized by array reconfiguration in which the bit configuration is reduced due to failure of a plurality of disk devices.

In FIG. 6, the array configuration of nine disk devices 8-a to 8-j shown in FIG.
-f is shown. In this case, the conversion to the reconfigured disk array device 3 is performed by changing the settings of the second attribute table 10 that sets bit assignments of 2 bits for each of the disk devices 8-a to 8-d.

In this way, the array configuration can be changed by disconnecting the defective disk control circuit or disk device and reconfiguring the array according to the changed attribute information in the table. It is possible to compensate for the shortage of disk devices and to continue operating the array disk device without impairing its function as an array disk.

FIG. 7 is an explanatory diagram showing self-diagnosis performed by the array disk device of the present invention. In FIG. 7, normally,
The disk devices 8-a to 8-j used in the array configuration of the array disk device are disk devices having both master and slave functions, and the master-slave relationship between master and slave is shown in FIG. 4(a). It is set by the first attribute table 9 as shown in FIG.

The array disk device usually includes disk control circuits 7-a to 7-j and disk devices 8-a to 8-8.
-j is equipped with a self-diagnosis function to confirm normal operation, and the diagnosis is executed and operation is confirmed during maintenance of the device. In this self-diagnosis, the conventional method in which the master-slave relationship is fixed has a problem in that it is only possible to perform a diagnosis tailored to the purpose of use of either the master disk or the slave disk.

Therefore, in the self-diagnosis shown in FIG. 7, the master disk and the slave disk are not fixed in their main dependence, but the master disk is assigned to the disk devices 8-a to 8- in accordance with the diagnosis execution modes 1 to 10. j machine number #0 to #7,
By executing diagnosis while sequentially shifting to #P and #SP, for all disk devices 8-a to 8-j,
The normality of each operation of the master disk and slave disk setting circuits and the control circuits switched by the settings can be confirmed, and self-diagnosis can be performed to determine the presence or absence of an abnormality, thereby further improving diagnostic efficiency and reliability.

[0035]

[Effects of the Invention] As explained above, according to the present invention,
The main dependencies, bit assignments, and backup device assignments of the array-configured disk devices are not fixed and can be changed automatically or manually, making it easy to replace the control circuit and disk device with a spare disk device in the event of a failure. can.

[0036]Furthermore, when returning a disk device that has been repaired after replacement, it is possible to eliminate the need for data transfer processing to the replaced disk device. Furthermore, in the event of failure of a plurality of disk devices, even if the scale is reduced to some extent, operation can be continued without impairing the device functions of the array disk, and reliability and maintainability can be improved.

[Brief explanation of the drawing]

[Fig. 1] Diagram explaining the principle of the present invention

[Figure 2] Configuration diagram of an embodiment of the present invention

[FIG. 3] An embodiment configuration diagram showing details of the array control circuit section in FIG. 2.

[Fig. 4] Explanatory diagram of an attribute table used in the present invention

[Figure 5]
Diagram explaining the operation of replacement processing when one disk device fails

[Figure 6] Diagram explaining the operation of reduced-scale replacement processing when multiple disk devices fail

[Fig. 7] Explanatory diagram of self-diagnosis operation in the present invention

[Figure 8]
Illustration of a conventional device with a fixed array configuration

[Explanation of symbols]

1: Central processing unit (upper unit) 2: Transfer control device 3: Array disk device 4: Transmission/reception circuit 5: Transfer control circuit 6: Array control circuit section 7-a to 7-j: Disk control circuit 8-a to 8-i: Disk device 8-j: Spare disk control circuit 9: First attribute table 10: Second attribute table 11: Configuration control means (configuration control unit) 12-a to 12-i: array control circuit 12-j: Array preliminary control circuit 13: Common data control circuit 15: Data bus switching circuit

Claims (3)

[Claims] Claim 1: A configuration control method for an array disk device in which a plurality of disk devices (8-a to 8-j) are configured in an array in parallel, and data is processed in parallel and sequentially to record or reproduce information. Then, in the array control circuit (6), parallel processing is performed with a first attribute table (9) that sets the main dependencies of a master disk that serves as a reference for rotationally synchronizing all disk devices and slave disks that rotate following the master disk. A second attribute table (10) in which data division allocation, parity information allocation, and backup device allocation are set;
and configuration control means (11) for configuring the disk devices (8-a to 8-j) in an array according to the attribute information in the second attribute table (9, 10) and executing parallel division processing. Configuration control method for array disk devices. 2. In the configuration control method for an array disk device according to claim 1, the configuration control means (11) is configured to control a disk device (8-a to 8-j) or a disk control circuit (7-j).
- At the time of replacement or replacement due to the occurrence of a failure in a to 7-j), the first and second parts changed based on the replacement or replacement.
A configuration control method for an array disk device, characterized in that the array configuration of the disk devices (8-a to 8-j) is reconfigured according to attribute information in attribute tables (9, 10). 3. In the configuration control method of an array disk device according to claim 1, when one or more disk devices (8-a to 8-j) or a disk control circuit (7-a
~7-j) is removed from the array configuration, changing the attribute information of the first and second attribute tables (9, 10) to attribute information that reduces the byte or bit configuration by the remaining disk device, A configuration control method for an array disk device, characterized in that the configuration control means (11) reconfigures the array configuration based on attribute information with a reduced byte or bit configuration.