JPH10187357A - Disk array device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a disk array device in which the redundancy of data can be ensured even when a hot stand-by physical disk is absent at the time of transition to a degradation operating state. SOLUTION: Each physical disk 17-1-17-n is constituted of areas 17a-1-17a-n for data in which (a) blocks can be stored, and 17b-1-17b-n for saving data in which a/(n-1) blocks can be stored. Any physical disk 17-1-17-n is separated, and block data stored in the area for data of the separated physical disk are restored at the time of transition to a degradation operating state, and then divided and saved in each area 17b-1-17b-n for saving data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array device, and more particularly to a method of dividing and storing data transferred from a host device and parity data generated from the data to a plurality of small and inexpensive magnetic disk devices. The present invention relates to a disk array device that implements one logical storage device.

[0002]

2. Description of the Related Art Conventionally, in a disk array device of this type, a plurality of small and inexpensive magnetic disk devices (hereinafter, referred to as a plurality) are used.
Each magnetic disk device is a physical disk), and these multiple physical disks are processed in parallel, and data has redundancy (parity data, etc.) to provide higher reliability than a single physical disk. And high performance.

As a disk array device (hereinafter, referred to as a logical disk) configured as described above, “A Ca
se For Redundant Arrays O
fInexpensive Disks (RAID) "
(Technical Report UBC / CSD
87/391, December 1987. There is a logical disk using the RAID theory proposed in (1).

Here, the above-mentioned high reliability means that even if one of the physical disks constituting a logical disk becomes inoperable, the physical disk is disconnected from the logical disk to transit to the degraded operation state. Operation as a logical disk, that is, generating and storing data of a redundant physical disk (physical disk storing parity data), and restoring the data by using it as a logical disk. This means that the operation of the system can be continued.

[0005] The above-mentioned high performance means that a plurality of physical disks are processed in parallel so that a performance far exceeding the performance when processing one physical disk can be exhibited. .

In the above-mentioned logical disk, when one physical disk becomes inoperable, a transition is made to a degraded operation state, and a redundant physical disk (a physical disk that records parity data. The operation can be continued by using (the physical disk changes).

The degraded operation state continues until the degraded physical disk is replaced. However, if data writing to an address allocated to the degraded physical disk is instructed by the host device during the degraded operation, The logical disk regenerates the parity data based on the parity data read from the redundant physical disk and the data specified by the higher-level device.

After that, the redundant physical disk rewrites the regenerated parity data, thereby completing the data writing as the logical disk. When writing of data from an upper-level device to an address where the degenerated physical disk becomes a redundant physical disk is instructed, the physical disk on which data is to be recorded and the redundant physical disk on which parity data is recorded are The relationship is reversed.

At this time, while the data transferred from the host device remains in the data cache, the data can be read from the data cache in response to a data read request from the host device.

Further, during the degenerate operation, in order to read data at the address allocated to the degenerated physical disk, it is necessary to generate parity from all physical disks (including redundant physical disks) other than the degenerated physical disk. The unit data is read, and necessary data is restored from the data.

When the degraded physical disk is replaced with a normal physical disk, the logical disk transitions to a repaired state, and rewrites repair data to the replaced physical disk over the entire area. This data restoration operation is performed by restoring data similar to the degraded operation state. A logical disk has a redundant physical disk as described above, and by using the redundant data, availability can be maintained even when one physical disk becomes inoperable.

By the way, when the logical disk transits to the degraded operation state, the early detached physical disk is replaced with a normal physical disk, incorporated into the logical disk, and data recovery as the logical disk is executed. It is necessary to restore redundant data.

This replacement of a physical disk is usually performed by hot swapping (replacement) by a maintenance person. However, when a physical disk is separately put into a hot standby state, the physical disk is immediately replaced when the logical disk transitions to a degraded operation state. Some measures have been taken to switch. There is also a method of preventing data loss by storing a failed data block in a non-volatile memory when a physical disk is not placed in hot standby. This method is described in JP-A-5-3
No. 14660.

[0014]

In the above-described conventional logical disk, if a data read error occurs in another physical disk while the degenerate operation is already being performed, data can be read as the logical disk. Will be gone. Considering that the capacity of the physical disk itself has been increased to a gigabyte at present, it is necessary to take recovery measures for a case where a data read error occurs on another physical disk during degeneration. come. In such a case, a large-capacity nonvolatile memory must be used in the method of storing the failed data block in the nonvolatile memory.

In a logical disk, it is very important to always maintain the redundancy of data divided and stored in a plurality of physical disks in consideration of the danger of data loss and the like. It is necessary to keep driving as short as possible.

Particularly, in the case where the logical disk is in the degraded operation state and there is no hot-standby physical disk, the physical disk that has been separated becomes a normal physical disk until the replacement / data recovery is completed. It should be noted that there is a problem that there is no redundant data in the data area of the logical disk. This is also a problem that occurs when it is assumed that hot swapping (replacement) of a physical disk by a maintenance person cannot be performed immediately.

Accordingly, an object of the present invention is to solve the above-mentioned problem and provide a disk array device capable of securing data redundancy even when there is no hot standby physical disk at the time of transition to the degraded operation state. To provide.

[0018]

In the disk array device according to the present invention, data transferred from a host device and parity data generated from the data are divided and stored in each of a plurality of magnetic disk devices to form one logical storage. A disk array device realizing the device, the data to be stored in each of the plurality of magnetic disk devices and to be stored in the magnetic disk device when the magnetic disk device becomes inoperable and transitions to a degraded operation state. Is provided with a plurality of evacuation areas for dividing and evacuation.

Another disk array device according to the present invention comprises:
In addition to the above-described configuration, a judging means for judging whether or not there is a hot standby magnetic disk device capable of immediately switching the inoperable magnetic disk device when transiting to the degraded operation state, and Means for restoring all the contents of the inoperable magnetic disk device when it is determined that there is no hot standby magnetic disk device and dividing the plurality of evacuation areas into evacuation areas for evacuation. .

Another disk array device according to the present invention comprises:
In addition to the above configuration, the system further includes a cache memory for holding data transmitted to and received from the host device when a write instruction and a read instruction are input from the host device to the magnetic disk device. When the data to be stored in the changed magnetic disk device is held in the cache memory, the contents of the cache memory are divided and copied to the save area provided in the magnetic disk device other than the magnetic disk device. ing.

Still another disk array device according to the present invention includes, in addition to the above-described configuration, a repair unit that recovers data of the inoperable magnetic disk device using data of another magnetic disk device. Then, the data restored by the restoration means is divided and saved in each of the plurality of save areas.

That is, in the disk array device of the present invention, a data area and a save data area are defined in each of a plurality of physical disks, and data to be stored in the inoperable physical disk is stored in a disk other than the physical disk. The data is saved in the save data area of each physical disk.

More specifically, one logical disk
Assuming that the microprocessor is composed of n physical disks capable of storing the number of blocks, the microprocessor determines that the relationship between the data area (a) and the save data area of the physical disk is: all blocks (A) = data block Each area of the physical disk is divided and defined so as to be (a) + backup data block [a / (n-1)],
Determine the total capacity of the logical disk.

When the logical disk in which the area is defined as described above transits to the degraded operation state during operation, if there is no hot standby physical disk, the logical disk is restored from a physical disk other than the degraded physical disk. Data is sequentially written to a physical disk other than the physical disk that has been degenerated.

If a hot standby physical disk is present when the system enters the degraded operation state, the recovery data is written to the hot standby physical disk so that the redundant data (in this case, to secure the redundant physical disk). ) Will be secured.

With this, for the period from the transition of the logical disk to the degraded operation state until the detached physical disk is replaced, even if the logical disk does not have a hot standby physical disk, redundant data of the logical disk itself is secured. One means can be provided.

Also, in preparation for the case where the above-mentioned serious failure occurs in the logical disk, when data to be recorded on the degenerated physical disk is expanded on the data cache of the logical disk, background processing for the host device is performed. The data (which is data to be originally recorded on the degenerated physical disk and obtained at the time of a write instruction from the higher-level device, or data of another physical disk at the time of a data read-out instruction from the higher-level device) (Which indicates the data restored using the data) is mirrored to a save data area allocated on a plurality of physical disks and saved. As a result, it is possible to secure redundant data and provide a means for preparing for a serious failure.

[0028]

Next, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a disk array device (hereinafter, referred to as a logical disk) according to an embodiment of the present invention. In the figure, a logical disk 1 includes an upper interface control circuit 11, a microprocessor 12, a data cache 13, a nonvolatile memory 14, and a disk array data control circuit 15.
And a physical disk interface (IF) control circuit unit 1
6-1 to 16-n, magnetic disk devices (hereinafter, referred to as physical disks) 17-1 to 17-n, and a hot standby physical disk 18.

FIG. 2 shows the physical disks 17-1 to 17-17 of FIG.
It is a figure which shows the structure of -n. In the figure, each of the physical disks 17-1 to 17-n has a data area 17a-1 to 17a-n capable of storing a block and a save data area 17b-1 to which a / (n-1) block can be stored. 17b-
n.

At this time, the RAID (Redundant)
Arrays of Inexpensive Dis
ks), the number of redundant data blocks of the logical disk 1 is a block because parity data is secured.

Here, the physical disks 17-1 to 17-n
And the block length of each logical disk 1 is equal,
The number of recordable blocks of the logical disk 1 is represented by [a block × (n-1)] blocks. In addition, each physical disk 17-1 to 17-n
Data areas 17b-1 to 17b-1 assigned to
The sum of b−n is [a / (n−1)] × (n−1) = a blocks when one physical disk is disconnected. These save data areas 17b-1 to 17b-
The n management information is recorded in the nonvolatile memory 14 by the microprocessor 12 and is managed / used during operation.

The save data area 1 shown in FIG.
7b-1 to 17b-n can be fixedly assigned to the logical disk 1 itself, but the save data areas 17b-1 to 17b-n are set according to the needs of each user.
It is also possible to select not to assign. Of course, in that case, the processing method in the degenerate operation state according to the embodiment of the present invention will not be used.

FIGS. 3 and 4 are flowcharts showing a processing method in a degenerate operation state according to an embodiment of the present invention.
5 to 7 are views showing a processing method in a degenerate operation state according to an embodiment of the present invention.

FIG. 5 shows the data of the detached physical disk 17-3 and the data of the other physical disks 17-1, 17-2,.
10 shows the flow of data when restoration is performed using data of −n. FIG. 6 shows a data flow when the repair data shown in FIG. 5 is recorded on the physical disk 17-1 as an example. FIG. 7 shows restoration data of each physical disk 17-.
1, 17-2,..., 17-n, and only four physical disks 17-1 to 17-4 are shown in FIG.

A processing method in the degenerate operation state according to one embodiment of the present invention will be described with reference to FIGS.
Hereinafter, the operation when the logical disk 1 is operated will be described.

First, among the physical disks 17-1 to 17-n constituting the logical disk 1, the physical disk 17-
When a failure occurs such that the operation of the physical disk 1 becomes impossible, the microprocessor 12
Separate 7-3. As a result, the logical disk 1 transitions to the degraded operation state, and the microprocessor 1
2 is logical disk 1 and hot standby physical disk 1
8 is checked (step S in FIG. 3).
1).

When the hot standby physical disk 18 exists, the microprocessor 12 stores the data of the physical disk 17-3 restored from the logical disk 1 in the hot standby physical disk 18 (step S2 in FIG. 3).

If the hot standby physical disk 18 does not exist, the microprocessor 12 starts data recovery of the separated physical disk 17-3 (step S3 in FIG. 3).

In this case, the microprocessor 12 sends the data recovery target area of the detached physical disk 17-3 in the logical disk 1 and the physical disks 17-1, 17-2,
,..., And 17-n, select the save destination of the restoration data (step S5 in FIG. 3).

The microprocessor 12 separates the block data of the selected data recovery target area of the physical disk 17-3 from the other physical disks 17-1, 17-2,.
The data is restored from the data of 17-n (step S6 in FIG. 3), and the restored data is selected from the selected physical disks 17-1, 17-.
2,..., 17-n save data areas 17b-1 to 17b-1
7b-n (FIG. 3, step S7).

The microprocessor 12 repeats the above processing until the data recovery of the data recovery target area of the detached physical disk 17-3 is completed (step S in FIG. 3).
4 to S7).

Next, the operation during operation of the logical disk 1 will be described in detail with reference to FIG. First, when a failure occurs such that the physical disk 17-3 becomes inoperable in each of the physical disks 17-1 to 17-n constituting the logical disk 1, the microprocessor 12 transmits the logical disk 1 to the physical disk 17-3. Separate -3. As a result, the logical disk 1 transitions to the degraded operation state, and the microprocessor 12 checks whether the logical disk 1 has the hot standby physical disk 18 (step S11 in FIG. 4).

When the hot standby physical disk 18 exists, the microprocessor 12 stores the data of the physical disk 17-3 restored from the logical disk 1 in the hot standby physical disk 18 (step S12 in FIG. 4).

When the hot standby physical disk 18 does not exist, the microprocessor 12 removes the physical disk 17-1,
.., 17-n are assigned disk number values of N = 1 to (n-1) for saving the restoration data (step S13 in FIG. 4).

Thereafter, the microprocessor 12 initializes the value of the disk number N (N = 1) (step S1 in FIG. 4).
4), the detached physical disk 17 in the logical disk 1
-3 is selected (step S1 in FIG. 4).
6), other physical disks 17-1, 17-2,..., 1
The target data block is restored from the data 7-n (step S17 in FIG. 4).

In one embodiment of the present invention, another physical disk 1
The data of 7-1, 17-2,..., 17-n is divided into 16 blocks (step S17 in FIG. 4), and the data to be separated and written to the physical disk 17-3 is restored. The block length can be made variable for the sake of data control of the logical disk 1.

In FIG. 5, the physical disks 17-1,
17-2,..., 17-n data d0 to d (n-1)
The data is restored in the disk array data control circuit unit 15 from [data other than d2], and the restored data d
2 shows a flow until the data No. 2 is temporarily stored in the data cache 13.

The data d2 of the detached physical disk 17-3 restored by the above processing is continuously recorded in the save data area 17b-1 of the physical disk 17-1 selected by the microprocessor 12 (step in FIG. 4). S
21).

FIG. 6 shows the flow of the restoration data d2 at this time. That is, the microprocessor 12 controls the physical disk 1 by the disk array data control circuit 15.
7-1, 17-2,..., 17-n data d0 to d
(N-1) The data d2 restored from [data other than d2] and temporarily stored in the data cache 13 is stored in the save data area of the other physical disks 17-1, 17-2,..., 17-n. 17b-1, 17b-2, ..., 17b-
Save to n.

The save destination area for the repair data is sequentially stored in the physical disks 17-1 and 17- by the microprocessor 12.
2,..., 17-n. The microprocessor 12 performs this series of processes (step S in FIG. 4).
17 to S21) are completed, the recorded restoration data d2
Is stored in the nonvolatile memory 14 (step S22 in FIG. 4).

The data restoration process in the logical disk 1 is repeated every 16 blocks until the data restoration is completed for all the data areas of the logical disk 1 (steps S15 to S22 in FIG. 4).

That is, if the logical disk 1 is, for example,
As shown in FIG. 3, when a failure occurs such that the physical disk 17-3 becomes inoperable, the data areas 17a-1 to 17a of the physical disk 17-3 are configured. The data d02, d12, d32,... Of block a written in −n are the save data areas 17b-1, 17b-2, 17b of the other physical disks 17-1, 17-2, 17-4. -4.

Thereafter, the microprocessor 12 starts to disconnect the physical disk 17 recorded on the non-volatile memory 14.
-3, based on the recording management information of the repair data of the backup data, the backup data areas 17b-1, 17b-2,.
Operation is performed by reading from 7b-n.
It should be noted that this data restoration operation can be performed without interrupting the access of the higher-level device by performing it in the background on the higher-level device (not shown).

In the above processing method, the saving of the entire data area 17a-3 of the detached physical disk 17-3 has been described.
The data to be restored when the writing and reading of the data to and from the 7-3 are instructed is transferred to another physical disk 17-1,
17-2,..., 17-n
1, 17b-2,..., 17b-n.

On the other hand, the data on the separated physical disk 17-3 is restored and the respective saved data areas 17b-1, 17b are restored.
b-2,..., 17b-n, if a read instruction is input to the address of the physical disk 17-3 disconnected from the host device, as shown in FIG. -1, 17-2,..., 17-n data d
Data is restored from 0 to d (n-1) [data other than d2] by the disk array data control circuit unit 15, and the restored data d2 is stored in the data cache 13 and sent to the host device.

At this time, at the same time, the repair data d2 stored in the data cache 13 is transferred to the other physical disks 17-1 and 17-1 by the microprocessor 12, as shown in FIG.
17-2,..., 17-n
1, 17b-2,..., 17b-n are saved by mirroring. When a write instruction is input to the address of the physical disk 17-3 disconnected from the host device, the data to be written is stored in the data cache 1
, And the other physical disks 17-1, 17-2,..., 17-n
Saved data areas 17b-1, 17b-2,..., 1
It is saved by mirroring to 7b-n.

As described above, the plurality of physical disks 17-1
17-n for each of the save data for dividing and saving the data to be stored in the physical disk when any one of the physical disks 17-1 to 17-n becomes inoperable and transitions to the degraded operation state. By arranging the areas 17b-1 to 17b-n, data redundancy can be ensured even when there is no physical disk 18 that is in hot standby at the time of transition to the degraded operation state.

[0058]

As described above, according to the present invention, the data transferred from the upper-level device and the parity data generated from the data are divided and stored in each of a plurality of magnetic disk devices, thereby providing one logical storage device. When one of the magnetic disk devices becomes inoperable and transitions to the degraded operation state, a plurality of save areas for dividing and saving data to be stored in the magnetic disk device are provided in the disk array device realizing the above. By arranging them in each magnetic disk device, there is an effect that data redundancy can be ensured even when there is no hot-standby physical disk at the time of transition to the degraded operation state.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a disk array device according to one embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of a physical disk of FIG. 1;

FIG. 3 is a flowchart illustrating a processing method in a degenerate operation state according to an embodiment of the present invention.

FIG. 4 is a flowchart illustrating a processing method in a degenerate operation state according to an embodiment of the present invention.

FIG. 5 is a diagram illustrating a processing method in a degenerate operation state according to an embodiment of the present invention.

FIG. 6 is a diagram illustrating a processing method in a degenerate operation state according to an embodiment of the present invention.

FIG. 7 is a diagram illustrating a processing method in a degenerate operation state according to an embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Disk array device (logical disk) 12 Microprocessor 13 Data cache 14 Nonvolatile memory 15 Disk array data control circuit unit 16-1 to 16-n Physical disk interface control circuit unit 17-1 to 17-n Magnetic disk device (physical Disk) 17a-1 to 17a-n Data area 17b-1 to 17b-n Save data area

Claims (4)

[Claims] 1. A disk array device that realizes one logical storage device by dividing and storing data transferred from an upper-level device and parity data generated from the data to each of a plurality of magnetic disk devices, A plurality of save areas are provided in each of the plurality of magnetic disk devices and divide and save data to be stored in the magnetic disk device when the magnetic disk device becomes inoperable and transitions to the degraded operation state. A disk array device characterized by the above-mentioned. 2. A determination means for determining whether or not there is a hot standby magnetic disk device capable of immediately switching the inoperable magnetic disk device when transitioning to the degraded operation state, and the determination means Means for restoring all the contents of the inoperable magnetic disk device when it is determined that the hot standby magnetic disk device does not exist, dividing the data into the plurality of save areas, and saving the data. The disk array device according to claim 1. 3. A cache memory for holding data transmitted to and received from the host device when a write command and a read command are input from the host device to the magnetic disk device, wherein the inoperable magnetic device is provided. When data to be stored in a disk device is held in the cache memory, the content of the cache memory is divided and copied to the save area provided in a magnetic disk device other than the magnetic disk device. 2. The disk array device according to claim 1, wherein: 4. A restoration means for restoring data of the inoperable magnetic disk device using data of another magnetic disk device, and restoring the data restored by the restoration means to each of the plurality of save areas. 4. The disk array device according to claim 1, wherein the disk array device is divided and saved.