JP2000357058A - Disk array device - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To disperse the concentration of processing load when a disk fails. A disk array for restoring data stored in a failed disk based on storage information of another disk and writing the data into a spare disk when one of a plurality of disks fails. In the device, means 20 for setting the recovery mode to the occurrence of a disk failure, a mapping table 19 for storing an already accessed area of each area of the failed disk, and an access request for the failed disk When the specified area is not stored in the mapping table, this area is newly registered in the mapping table, and data in this area is restored based on the storage information of another disk and written into the spare disk. , 24, and when the area is stored in the mapping table, Means for accessing the disk.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array apparatus comprising a plurality of disks connected in parallel, and more particularly, when one of the plurality of disks fails, the data is stored in the failed disk. The present invention relates to a disk array device having a data recovery function of restoring data that has been stored based on storage information of another disk.

[0002]

2. Description of the Related Art Redundant arrays of inexpenses (RAID) incorporated in computer systems.
ive disks), the data of the write request to the HDD output from the host computer is divided into a plurality of parts, and the data is written in parallel to a plurality of disks (hard disks) connected in parallel and read from the HDD. By reading out the data divided and stored on a plurality of disks in parallel in response to a request, it is possible to speed up access to the HDD.

In a disk array device (RAID) having such a basic function, in order to ensure the reliability as a data storage device, when one of a plurality of disks fails, the failed disk A data recovery function for restoring the data stored in the disk based on the storage information of another disk is incorporated.

FIGS. 9A and 9B show the operation of a disk array device (RAID) incorporating this data recovery function.
This will be described with reference to FIG. 9B and FIG. This disk array device includes a plurality of disks (hard disks) 1a,
1b and 1c and an access control unit 2. FIGS. 9A to 9C show a case where a read access request is issued from an external host computer.

FIG. 9A shows a normal state, in which data d ₁ , d ₂ and d ₃ are normally read from the disks 1a, 1b and 1c. FIG. 9B shows a state where the third disk 1c has failed. In this case, the data d ₃ stored on the third disk 1c is replaced with another normal disk 1c.
The data is restored using information such as parity stored in a and 1b, and is output as a read access result.

Thereafter, the failed disk 1c is replaced with a spare disk 3 as shown in FIG. Then, without stopping the computer system in which the disk array device is incorporated, each data stored in the failed disk 1c is stored in the disk 1c using the above-described method.
Is sequentially written from the head area of the spare disk 3 while sequentially restoring the data from the head area of the spare disk 3.

When all the data stored in the failed disk 1c is written to the spare disk 3, an external access request is switched to the spare disk 3, and
Return to the normal access operation mode.

[0008]

However, FIG.
The disk array device having the data recovery function for the data stored on the failed disk described with reference to FIGS. 9A to 9C has the following problems to be solved.

In other words, this disk data recovery method requires a great deal of processing until all data in the failed disk 1c is restored and writing to the spare disk 3 is completed. In addition, during the recovery period, it is necessary to execute a process corresponding to a normal access request input from the outside. As a result, the processing load on the entire disk array device is greatly increased.

Further, since all data in all areas of the failed disk 1c are to be restored and unused areas of the disk 1c are treated equally, wasteful processing occurs.

In addition, due to the data restoration, temporary concentrated disk accesses occur to the normal disks 1a and 1b other than the failed disk 1c during data restoration. As a result, the load on each of the disks 1a and 1b and the load on the access control unit 2 are large, and the performance of the entire disk array device is reduced. Further, the risk of failure (2) for normal disks 1a and 1b other than the failed disk 1c There is a concern that the reliability of the entire disk array device will decrease.

Further, these problems increase in proportion to the number of disks incorporated in the disk array device. In particular,
The effect is great when the above-mentioned disk array device is incorporated in a video server that requires a huge amount of disks, a stable disk response speed, and reliable data.

[0013] The present invention has been made in view of such circumstances, and even if one of a plurality of disks incorporated in the apparatus fails, each of the data stored in each area of the failed disk is lost. It is an object of the present invention to provide a disk array device capable of performing data restoration processing in a time-dispersed manner, uniforming the processing of the entire apparatus, and improving the performance and reliability of the entire apparatus.

[0014]

SUMMARY OF THE INVENTION According to the present invention, a plurality of disks for storing various data are incorporated, a normal access to each disk is performed in response to an external access request, and a plurality of disks are stored. When one of the disks fails, the disk array device having a data recovery function of restoring the data stored on the failed disk based on the storage information of another disk and writing the data to a spare disk is applied. You.

In order to solve the above problem, in the disk array device of the present invention, a means for setting an operation mode to a recovery mode in response to occurrence of a disk failure,
A mapping table for storing the already accessed area of each area of the failed disk, and when an access request for the failed disk occurs during the setting of the recovering mode, the area specified by the access request is stored in the mapping table. If not, this area is newly registered in the mapping table, data in this area is restored based on the storage information of another disk and written to the spare disk, and the area is stored in the mapping table. Means for accessing the spare disk when the device is in the standby state.

In the disk array device configured as described above, when a disk failure occurs, the operation mode of access is set to the recovering mode. In general, when a disk failure occurs, it is necessary to restore each data stored in each area of the failed disk using storage information such as parity stored in another normal disk. Does not restore all the data at once, but restores the data stored in the relevant area using the storage information of the other disk every time an access is made to a new area in the failed disk. Processing is performed.

Therefore, since the data restoration processing does not occur intensively, the data processing of the entire disk array device is made uniform.

In another aspect of the invention, a means for monitoring the number of areas stored in the mapping table is provided for the disk array apparatus of the invention, and the number of monitored areas is predetermined. When the specified value is reached,
A means for restoring each data of all the remaining areas not stored in the mapping table of the failed disk based on the storage information of the other disks and writing the data to the spare disk is added.

Conversely, in the disk array device configured as described above, when the number of areas on the failed disk where data has not been restored decreases below a certain value, the data in these areas are forcibly combined. Will be restored. The processing load on the disk array device in this case is not so large. Therefore, the access operation mode returns to the later normal mode within a certain period after the disk failure.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. (First Embodiment) FIG. 1 is a block diagram showing a schematic configuration of a disk array device according to a first embodiment of the present invention.
In the disk array device of the first embodiment, n disks 11 and n spare disks 12 for storing various data are respectively stored in the memory unit 14 via the bus 13.
Are connected to each buffer 15.

The memory unit 14 is connected to a control unit 16 composed of a kind of computer. This control unit 16
Inside, an access management unit 17, a data restoration management unit 18, and a mapping table 19 are provided.

In the mapping table 19, as shown in FIG. 2, a plurality of areas 19a for storing the identification information of the access area in the failed disk and the identification information of the spare disk 12 are formed.

The access management unit 17 includes a mode switching unit 20 for switching the operation mode of the disk access in the disk array device between the normal mode and the recovering mode, and a new area 19 in the mapping table 19.
The access area mapping unit 21 that writes the identification information of the access area of the failed disk and the identification information of the spare disk 12 into the a, and the access area of the access request input from the outside is already stored in the area 19 a in the mapping table 19. Mapping determination unit 2 for determining whether or not the mapping has been completed
2 are provided.

Further, in the data restoration management unit 18, a data restoration unit 23 which is stored in an accessed area of the failed disk and restores data based on the storage information of another normal disk, and a restored data A restoration data writing unit 24 for writing data to the spare disk 12 is provided.

In the disk array device having such a configuration, when each of the n disks 11 is in a normal state,
The access operation mode is the normal mode. In the normal mode, direct access to each disk 11 is performed in response to an access request input from outside.

When it is detected that one of the n disks 11 has failed, the access management unit 17 and the data restoration management unit 18 of the control unit 16 execute the failed disk recovery processing shown in FIG. Is carried out. Here, as shown in FIG. 2, the third disk 11
It is assumed that a failure has occurred in (D ₃ ).

When a failure occurs in the disk 11 (D ₃ ) and access becomes impossible, the mode switching unit 20 of the access management unit 17 changes the access operation mode from the normal mode to the data restoration using the mapping table 19. The mode is switched to the recovery mode (S1).

After switching the operation mode to the recovering mode, the access management unit 17
It waits until an access request to (D ₃ ) is input. Then, when an access request to the failed disk 11 (D ₃ ) is input (S 2), the mapping determination unit 22 checks whether an area to be accessed is already registered in any one of the areas 19 a of the mapping table 19. (S
3).

If the corresponding access area has already been stored in the mapping table 19 (4), since the data stored in the access area of the failed disk 11 (D ₃ ) has already been restored, the mapping table 19 Referring to the spare disk identification information, access to the corresponding spare disk 12 is performed (S5).

In S4, if the corresponding access area is not stored in the mapping table 19, the access area mapping unit 21 uses the identification information of the current access area as the new area 19 in the mapping table 19.
a is registered (S6).

If the corresponding access request is a data write access (S7), the corresponding defective disk 11
The relevant data is immediately written to the spare disk 12 corresponding to (D ₃ ) (S8). In S7, if the access request is a data read access, the data restoration management unit 18
Processing proceeds to FIG. 4 shows the processing operation on the data restoration management date 18.

First, the data restoring unit 23 is activated, and
Disk 11 (D_ThreeOther normal disks 11 other than)
Data such as parity of the corresponding access area from
The data is read into each buffer 15 of the memory unit 14 and the memory unit 1
4 on the failed disk 11 (D _Three) Memorized in the access area
The restored data is restored (S9). And this
It is necessary to access the restored data as a read access result.
The data is output to the requesting host computer (S10). Next
Then, the restored data writing unit 24
Disk 11 (D_Three) Of the spare disk 12 corresponding to
Writing to this area (S11). Then, returning to S2, the outside
From failed disk 11 (D_Three) Access request to enter
Wait until you do.

The failed disk 11 (D ₃ ) is externally supplied.
When an access request to a normal disk 11 other than the above is input, without referring to the mapping table 19,
Immediately, access to the relevant disk 11 is executed.

In the disk array device configured as described above, when a disk failure occurs, if an access is made to the corresponding failed disk 11 (D ₃ ) after the occurrence of the disk failure, the area specified by the access is specified. Only the data is correctly restored when the access occurs, and stored in the spare disk 12.

Of course, at this time, the failed disk 11 (D
If the entire area ₃ ) is accessed, all data is restored to the spare disk 12. For example, some disk array devices have a function of sequentially monitoring the entire area of each disk 11 at regular time intervals. In this case, all the data is restored by monitoring each area of the failed disk 11 (D ₃ ) in order.

When all the data of the failed disk 11 (D ₃ ) has been restored to the spare disk 12, the disk array device switches between the failed disk 11 (D ₃ ) and the spare disk 12 to perform an access operation. The mode can be returned to the normal mode.

How to define the access area of the failed disk 11 (D ₃ ) to be written into each area 19a in the mapping table 19 depends on the access mode to the disk 11 and the amount of resources available to the disk array device. May be changed. For example, if the required storage capacity of the mapping table 19 increases and there is no problem, the access area may be defined as the minimum access unit (sector) of the disk, and conversely, the required storage capacity of the mapping table 19 is desired to be reduced. In this case, a set of several sectors may be defined as an access area. In particular,
In the latter case, it works effectively when an access mode in which access to adjacent sectors frequently occurs or when a disk storage method is adopted.

As described above, when one disk 11 out of the plurality of disks 11 incorporated in the disk array device has failed, data is stored only in the accessed area in the failed disk 11 (D ₃ ). The data is restored and written to the spare disk 12.

Therefore, it is possible to avoid a temporary concentration of the data restoration process and the data writing process, so that the processing load on the control unit 16 and the disk 11 can be reduced, and a decrease in access performance and a decrease in reliability during data restoration can be suppressed. can do.

Further, since the access to the spare disk 12 is always possible, the data restored and written into the spare disk 12 can be used immediately without waiting for the restoration of the entire disk, so that the access process can be efficiently performed. .

(Second Embodiment) FIG. 5 is a block diagram showing a schematic configuration of a disk array device according to a second embodiment of the present invention. The same parts as those of the disk array device of the first embodiment shown in FIG. Therefore, the detailed description of the overlapping part will be omitted. Data restoration management unit 1 in the disk array device of the first embodiment
In addition to the data restoration unit 23 and the restoration data writing unit 24 described above, the remaining data restoration unit 25 and the restoration adjustment unit 26
Is provided.

The restoration adjustment unit 26 is configured to restore the failed disk 11 stored in the mapping table 19 shown in FIG.
The number N _A of accessed areas in (D ₃ ) is monitored, and the number N _{A of} areas is monitored at a predetermined threshold (predetermined value).
, The remaining data restoring unit 25 is started.

The remaining data restoring unit 25 is provided for the failed disk 11
The data stored in all the remaining areas not stored in the mapping table 19 of (D ₃ ) are restored collectively based on the storage information of the other normal disks 11 and written to the spare disk 12. .

When it is detected that one of the n disks 11 has failed, the access management unit 17 and the data restoration management unit 18 of the control unit 16 are controlled.
Performs the failed disk recovery process shown in FIG. In addition,
Here, it is assumed that a failure has occurred in the _third disk 11 (D ₃ ), as in the above-described first embodiment.

When a failure occurs in the disk 11 (D ₃ ) and access becomes impossible, the mode switching unit 20 of the access management unit 17 changes the operation mode of access from the normal mode to the data recovery using the mapping table 19. The mode is switched to the recovering mode (Q11).

After switching the operation mode to the recovering mode,
The access management unit 17 receives the failed disk 11 (D
₃ ) Wait until an access request is input. Then, when an access request to the failed disk 11 (D ₃ ) is input (Q2), the restoration adjustment unit 2 of the data restoration management unit 18
6 starts, and detects the number N _A of the accessed areas in the failed disk 11 (D ₃ ) stored in the mapping table 19 (Q3).

If the number N _{A of} detected areas does not reach the predetermined threshold value (predetermined value) (Q4) and there is no operation for instructing the operator (user) to restore the remaining data (Q5), Proceeding to Q6, the process of restoring only the data stored in the area corresponding to the current access and writing the restored data to the spare disk 12 is performed as described above. Specifically, S3 in the flowchart of FIG.
To S11. Then, the process returns to Q2, and waits until the next access request to the failed disk 11 (D ₃ ) is input from outside.

When the number N _{A of} detected areas reaches a predetermined threshold value (predetermined value), or when an operator (user) performs a restoring instruction operation on remaining data (Q5). Starts the remaining data restoring unit 25. That is, as described above, the remaining data restoring unit 25 replaces each data stored in all remaining areas not stored in the mapping table 19 of the failed disk 11 (D ₃ ) with other normal disks. 11 and write them to the spare disk 12 collectively based on the storage information of Q11 (Q7, Q7).
8).

Thereafter, an actual access to the currently input access request is executed for the spare disk 12 on which all data has been written (Q9). Finally, the mode switching unit 20 is activated to return the access operation mode of the disk array device from the recovering mode to the original normal mode (Q10).

The disk array device configured as described above
, The failed disk 11 (D _Three) Data
When the number of unrestored areas falls below a certain value,
The data in these areas are collected together and the data is forcibly restored.
Is written to the spare disk 12 and the operation mode
Returns from the recovering mode to the normal mode.

As described above, when the number of data to be restored is reduced to a certain value or less, the data restoration processing is performed collectively, but the processing load on the disk array device in this case is not so large. Accordingly, the access processing mode of the entire disk array device can be improved by returning the access operation mode to the normal mode later within a certain period after the disk failure.

(Third Embodiment) FIG. 7 is a block diagram showing a schematic configuration of a disk array device according to a third embodiment of the present invention. The same parts as those of the disk array device of the first embodiment shown in FIG. Therefore, the detailed description of the overlapping part will be omitted. In the disk array device of the third embodiment, the first array shown in FIG.
In the disk array device of the embodiment, each disk 11
Each of the spare disks 12 attached to the above has been removed. Then, only one spare disk 12a is prepared as a spare when a failure occurs in each disk 11. Naturally, the spare disk 12a is not connected to the memory unit 14. However, this spare disk 12a is
It has the same configuration as each disk 11 bus-connected to the memory unit 14.

When each disk 11 fails,
The operator (user) can easily replace the failed disk with the spare disk.

In such a disk array device in which the spare disk 12a is simulated, the access management unit 17 and the data restoration management unit 18 of the control unit 16 when one of the plurality of disks 11 fails are shown in FIG. The failed disk recovery process shown in FIG. Here, it is assumed that a failure has occurred in the _third disk 11 (D ₃ ) as in the case of the disk array device of the first embodiment.

When a failure occurs in the disk 11 (D ₃ ) and access becomes impossible, the access management unit 16 displays the failure occurrence and the name (number) of the failed disk on the display unit, and notifies the operator (user). Warning.

When the operator (user) enters the failed disk 11
(D ₃ ) is removed from the bus 13 and the failed disk 1
1 (D ₃ ) and the spare disk 12 a
3 (S0).

When the spare disk 12a is mounted, the mode switching unit 20 of the access management unit 17 switches the access operation mode from the normal mode to the recovering mode in which data is restored using the mapping table 19 (S).
1).

Subsequent restoration processing of each data stored in each area of the failed disk 11 (D ₃ ) and writing processing of each restored data to the spare disk 12a S1
Steps S11 to S11 are the same as those of the disk array device of the first embodiment shown in FIG.
Since the processing is the same as that of S1 to S11 shown in FIG.

Therefore, in the disk array device of the third embodiment configured as described above, the data restoration of the data stored in each area of the failed disk 11 (D ₃ ) is performed collectively. Since the present invention is implemented in a distributed manner, almost the same effects as in the disk array device of the first embodiment can be obtained.

Further, in the disk array device of the third embodiment, since there is only one spare disk 12a, the size of the entire disk array device can be reduced and the manufacturing cost can be reduced.

[0061]

As described above, in the disk array device of the present invention, the data restoration of each data stored in each area of the failed disk is not performed collectively, but the access to the corresponding area is not performed. It is implemented each time it occurs.

Therefore, even if one of a plurality of disks fails in the apparatus, the restoration processing of each data stored in each area of the failed disk can be executed in a time-dispersed manner. The entire processing is made uniform, and the performance and reliability of the entire apparatus can be improved.

[Brief description of the drawings]

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

FIG. 2 is a diagram showing storage contents of a mapping table formed in the disk array device;

FIG. 3 is a flowchart showing the operation of the disk array device.

FIG. 4 is a schematic diagram for explaining a data restoration operation in the same disk array device.

FIG. 5 is a block diagram showing a schematic configuration of a disk array device according to a second embodiment of the present invention.

FIG. 6 is a flowchart showing the operation of the disk array device.

FIG. 7 is a block diagram showing a schematic configuration of a disk array device according to a third embodiment of the present invention.

FIG. 8 is a flowchart showing the operation of the disk array device.

FIG. 9 is a schematic diagram for explaining a data restoration operation in a conventional disk array device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 ... Disk 12, 12a ... Spare disk 13 ... Bus 14 ... Memory part 15 ... Buffer 16 ... Control part 17 ... Access management part 18 ... Data restoration management part 19 ... Mapping table 20 ... Mode switching part 21 ... Access area mapping part 22 ... Mapping determination unit 23 ... Data restoration unit 24 ... Restoration data writing unit 25 ... Remaining data restoration unit 26 ... Restoration adjustment unit

Claims (2)

[Claims] 1. A plurality of disks for storing various data are incorporated, a normal access to each of the disks is performed in response to an external access request, and one of the plurality of disks fails. In this case, a disk array device having a data recovery function of restoring the data stored on the failed disk based on the storage information of another disk and writing the data to a spare disk operates in response to the occurrence of the disk failure. Means for setting the mode to the recovering mode, a mapping table for storing the already accessed area of each area of the failed disk, and an access request to the failed disk during the setting period of the recovering mode. When the area specified by the access request is not stored in the mapping table, Means for newly registering this area in the mapping table, restoring the data in this area based on the storage information of another disk and writing the data to the spare disk, and the area being stored in the mapping table A means for accessing the spare disk. 2. A means for monitoring the number of areas stored in the mapping table, and when the number of monitored areas reaches a predetermined value, a remaining part of the failed disk which is not stored in the mapping table. 2. The disk array device according to claim 1, further comprising means for restoring each data of all the areas based on the storage information of another disk and writing the data to the spare disk.