JP2003167687A - Method and device for disk array control and disk array control program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve reliability of a disk array 20. <P>SOLUTION: A disk array controller 10 is provided with a means which generates parity data for write data received from a host processor 12 and writes the write data and the parity data in logical disks #1 to #n and writes the write data in a data storage disk 14, a means which reads the write data and the parity data from logical disks #1, etc., to perform a parity check, and a means which reads the write data from the data storage device 14 to generate parity data when finding a parity error as the result of the parity check and writes the write data and the generated parity data in logical disks #1, etc., again. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array control device and method for controlling a disk array composed of a plurality of disks as a storage medium, and a disk array control program.

[0002]

2. Description of the Related Art In conventional magnetic disk array control,
Parity data is generated for the write data received from the host device, and the write data and the parity data are written to the logical disk. However, since the data check is not performed until re-reading, it is difficult to immediately identify that the data fraud has occurred.
In particular, if write-out of divided data or a parity generation error occurs in only one of the physical disks that make up the logical disk, data corruption will occur during re-reading, but which disk data is invalid. Could not be identified. Further, if the data is not read for a long time after the data is written, there is a high possibility that the data backup does not exist, and thus it is difficult to recover the data even if the data is fraudulent.

On the other hand, the present inventor has already invented a disk array control device which solves this problem (Japanese Patent Laid-Open No. 2001-142650). That is, this disk array control device enables early detection of the occurrence of illegal data and disk failure, and enables reliable storage of even data that has not been read for a long period of time.

FIG. 6 is a block diagram showing the disk array control device disclosed in the above publication. Hereinafter, description will be given with reference to this drawing.

The disk array control device 102 divides the data transmitted from the host device 101 into a cache memory 106 for temporarily storing the data transmitted from the host device 101 and the physical disks 121 ,. In addition, the data dividing unit 103 that generates parity data of the divided divided data and the data transfer unit 104 that stores the divided data and parity data generated by the data dividing unit 103 in each physical disk 121 ,. There is.

Further, the disk array controller 102 is
Divided data and parity data are sent to each physical disk 12
A data comparison unit 108 that reads out the data from each physical disk 121, and then compares the data with the corresponding data stored in the cache memory 106.
The data comparing unit 108 is provided with a failure determining unit 110 that determines that a failure has occurred in the physical disks 121, ...

The disk array control device 102, via the data transfer unit 104, each physical disk 121, 122, 12 constituting the host device 101 and the logical disk 120.
3 is connected to. Here, the physical disk 12
1, ... N from physical disk 1 to physical disk N
It is provided individually.

The disk transfer unit 104 is a host device 101.
Data is read and written to the cache memory 106 and data is transferred to the logical disk 120. The cache memory 106 uses a semiconductor memory such as a RAM that is faster to access than the physical disks 121.

In the data comparison unit 108, the write data received from the host device 101 is stored in the cache memory 106.
, The data stored in each physical disk 121, ... Is read and compared with the data stored in the cache memory 106. At this time, if there is a mismatched portion, it means that the data has not been written normally. Then, which physical disk 121, ... Is the data written in the portion that was not normally written is determined based on the division rule by the data division unit 103. Therefore, if there is a mismatched portion in the compared data, it becomes clear to which physical disk 121, ... Which illegal data causing the mismatch was stored.

Therefore, if there is a mismatched portion in the data comparison unit 108, the failure determination unit 110 determines that a failure has occurred in the physical disks 121, ... That store the mismatched portion. As a result, although the data can be read, it is possible to specify the physical disks 121, ... In which a failure such that the data cannot be written normally occurs and to perform the recovery process. Then, even if the data is not read for a long period of time, in order to confirm whether or not the data was correctly stored at the time of storing the data, for example, a plurality of physical disks 121, ... Have a failure at the same time and the data is recovered. It is possible to prevent the situation that cannot be done. Also, the physical disk 1 that could not be written normally when storing data
21 are identified, the physical disks 121,
By notifying the host device 101 of the device number and so on as the write failure information, the cause of the illegal data can be clarified while maintaining the environment in which the illegal data occurs. Therefore, as compared with the case where illegal data is generated after a long period of time, the cause of the failure can be easily understood.

When the failure determining unit 110 identifies the failed physical disk 121, ..., The recovery processing unit 112 reads the divided data including the invalid data portion from the cache memory 106, and performs the rewriting process. The cache memory 106 is repeated once or several times.
It is determined whether illegal data has occurred by comparing it with the data stored in.

FIG. 7 is a flowchart showing the operation of the disk array controller 102. Hereinafter, FIG. 6 and FIG.
The operation of the disk array controller 102 will be described based on FIG.

First, the data transfer unit 104 is used by the host device 1.
01 receives write data (step S1),
The data is stored in the cache memory 106 (step S2). Further, the data division unit 103 divides the data to generate parity data (step S).
3). The data transfer unit 104 stores the divided data and the parity data in each of the physical disks 121 to 123 (step S4). Then, the data transfer unit 104 reports the completion of storage to the higher-level device 101 (step S5).
Specifically, the command indicating the storage completion report is sent to the higher-level device 1
01 to the input / output interface.

Then, the data stored in the physical disks 121, ... Is compared with the data stored in the cache memory 106 (steps S6 to S8). The data transfer unit 104 reads the data stored in the physical disks 121, ... (Step S7). On the other hand, the data comparison unit 108 reads the same data from the cache memory 106 and generates parity data (step S8). Then, the data comparison unit 108 compares the data read from the physical disks 121, ... With the data stored in the cache memory 106 (step S9). In this example, the data comparison step includes a step of generating parity data of the data read from the cache memory 106. As a result, it is possible to determine whether or not the data storing the parity data is invalid.

As a result of the comparison of the two data, if they match each other, the writing is normally performed. Therefore, the data writing process is ended. On the other hand, if there is a difference between the two data,
The physical disks 121 storing the non-matching portions, ...
It is determined that a failure has occurred (step S10). On the other hand, if there is a difference between the two data, the recovery processing unit 112 is activated to perform recovery processing (step S1).
1).

[0016]

However, the disk array controller 102 has the following problems.

The cache memory 106 is not originally provided for backing up data, but is provided for terminating the writing from the host device 101 at high speed. Therefore, the capacity of the cache memory 106 with respect to the total capacity of the physical disk 102 becomes extremely small. In that case, if the data writing from the higher-level device 101 is concentrated, the cache memory 106 is not checked before the parity check of the logical disk 120 is completed.
Will be overwritten.

[0018]

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to detect the occurrence of illegal data and a disk failure at an early stage without causing a storage capacity shortage and a processing speed delay, even if the data cannot be read for a long time. (EN) It is possible to provide a disk array control device and method and a disk array control program that enable reliable storage.

[0019]

A disk array control device according to the present invention generates parity data for write data received from a higher-level device, writes the write data and the parity data to a logical disk, and Means for writing write data to a data storage disk; means for reading the write data and its parity data from the logical disk to perform a parity check; and, if the result of the parity check is a parity error, from the data storage disk A means for reading the write data to generate parity data, and writing the write data and the parity data again to the logical disk is provided (claim 1). At this time, when the write data and the parity data thereof are written to the logical disk, the write data and the parity data are divided into a plurality of physical disks that form the logical disk and written.
(Claim 2) Further, the logical disk and the data storage disk may be magnetic disks (claim 3).

The write data sent from the host device is written to the logical disk together with the parity data after the parity data is created. On the other hand, this write data is also written on the data storage disk. Then, this write data and its parity data are read from the logical disk and their parity check is executed. As a result of the parity check, if there is a parity error, it is possible that incorrect data is generated.
Then, the write data is read from the data storage disk, the parity data is subsequently created, and these are written again to the logical disk. in this way,
Since the occurrence of illegal data is detected early, the reliability of the disk array is improved.

The major difference between the prior art disk array control device (FIG. 6) and the present invention disk array control device is that the prior art cache memory (semiconductor memory) is a data storage disk in the present invention. is there. Therefore, in the present invention, a sufficient capacity for temporarily storing the write data can be obtained, so that various problems caused by the insufficient storage capacity can be solved. The reason is that a disk storage medium such as a magnetic disk or an optical disk can easily increase the capacity by an order of magnitude more than a semiconductor memory.

On the other hand, in the prior art, the data is checked by reading the data from both the logical disk and the cache memory and comparing the data. Therefore, simply replacing the cache memory in the conventional technique with the data storage disk causes another problem that the processing time becomes longer than that in the conventional technique. This is because the data read speed from the data storage disk is much slower than that of the cache memory.
Therefore, in the present invention, the data is checked by performing the parity check only on the data read from the logical disk. As a result, it is not necessary to read the data from the data storage disk when checking the data, so that the processing speed equivalent to that of the conventional technique is realized.

The disk array control method according to the present invention is
It is used in the disk array control device according to the present invention and is substantially the same as the disk array control device according to the present invention (claims 4 to 6). A disk array control program according to the present invention causes a computer to function as the disk array control device according to the present invention.
It is substantially the same as the disk array control device according to the present invention (claims 7 to 9).

In other words, the disk array control according to the present invention is characterized in that the data comparison means, the data storage disk and the stored data management table are provided, and the parity check of the data written in the disk is executed to Correct data can be rewritten from the data storage disk to the logical disk when a fraud is detected.
Further, in the present invention, the use of a magnetic disk for data backup makes it difficult to cause a capacity shortage. Further, since the stored data management table is accessed at the time of data check (that is, the data storage disk is not accessed at the time of data check), the data check processing time is almost the same as that of the conventional technique.

[0025]

FIG. 1 is a block diagram showing an embodiment of a disk array control device according to the present invention. Hereinafter, description will be given with reference to this drawing.

The disk array controller 10 of this embodiment
Generates parity data for the write data received from the host processor 12 as a higher-level device, writes the write data and the parity data to the logical disks # 1 to #n, and stores the write data in the data storage disk 14 And a means for writing the write data and its parity data to the logical disk # 1,
A means to read from ... and execute a parity check,
As a result of the parity check, if there is a parity error, the write data is read from the data storage disk 14 to generate parity data, and the write data and the parity data are rewritten to the logical disks # 1 ,. It is a thing.

In the present embodiment, when the write data and the parity data thereof are written to the logical disks # 1, ..., The write data and the parity data thereof are written in a plurality of physical disks 1 to 1 constituting the logical disks # 1 ,. N
Divide into and write. The logical disks # 1, ... And the data storage disk 14 are magnetic disks. Furthermore,
Each means constituting the disk array control device 10 is realized by a disk array control program on a microprocessor.

The disk array controller 10 is incorporated in the disk array 20. The disk array 20 is divided into a disk control unit 22 for controlling data with the host processor 12 and a logical disk unit 24 for recording data. The disk control unit 22 has a storage data management table 16 and a data transfer means 1 in addition to the disk array control device 10 and the data storage disk 14.
Eight. The logical disk unit 24 is composed of n logical disks # 1 to #n. Each logical disk # 1, ...
Are each composed of N physical disks 1 to N.

In the disk controller 22, the disk array controller 10 and the host processor 12 and the logical disk #
, And data transfer with the data storage disk 14 are controlled. The disk control unit 22 also includes a stored data management table 16 that manages information on the data transferred to the data storage disk 14.

A more detailed description will be given. Disk control unit 2
2 transfers data with the host processor 12,
A data transfer unit 18 for reading / writing data from / to the data storage disk 14 and transferring data to the logical disk unit 24, a data storage disk 14 for storing data from the host processor 12 without dividing the data, and a data storage Stored data management table 16 for storing management information of data written in disk 14, transfer control of data transfer means 18, writing data to stored data management table 16, and data written in logical disks # 1 ,. The disk array controller 10 executes the validity check of the parity and the like. The data transfer means 18 is composed of, for example, a data bus, an interface circuit and the like. The stored data management table 16 is composed of, for example, a semiconductor memory or the like.

Next, the outline of the operation of the disk array controller 10 will be described. The write data sent from the host processor 12 is written to the logical disks # 1, ... Together with the parity data after the parity data is created. On the other hand, this write data is also written in the data storage disk 14. Subsequently, the write data and the parity data are the logical disks # 1,
... and their parity checks are performed. As a result of the parity check, if there is a parity error, it is possible that incorrect data is generated. Therefore, the write data is read from the data storage disk 14, the parity data is subsequently created, and these are written again to the logical disks # 1 ,. in this way,
Since the occurrence of illegal data is detected early, the reliability of the disk array 20 is improved.

Further, the failed physical disk 1, ...
If is specified, the following operation may be performed. For example, by reading the data including the invalid data portion from the data storage disk 14 and repeating the rewriting process on the physical disks 1, ... judge. At this time, writing may be performed in different sectors of the physical disks 1, ... If illegal data is generated even after such rewriting, the physical disks 1, ... Which caused the illegal data are degenerated. The degeneration process is a process of separating the physical disks 1, ... After restoring all other data.

FIG. 2 is an explanatory diagram showing a recording format of write data received from the host processor. Less than,
A description will be given based on FIGS. 1 and 2.

The data received from the host processor 12 is stored in the data storage disk 14 as it is (FIG. 2 [1]). On the other hand, the data received from the host processor data 12 is divided into a prescribed size and recorded in the logical disks # 1, ... (FIG. 2 [2]). The divided data are written to the physical disks 1, ... Corresponding to the addresses of the logical disks # 1 ,. Further, when data is recorded on the logical disks # 1, ..., Which are composed of N physical disks 1 ,.
Parity data is generated for each piece of data (see FIG. 2).
[3]). The divided data and the generated parity data are written to each physical disk 1, ...
[3]).

FIG. 3 is an explanatory diagram showing the format of the stored data management table. Hereinafter, description will be given with reference to FIGS. 1 and 3.

After writing the data from the host processor 12 to the logical disk # 1 and the data storage disk 14, the disk array control device 10 writes the logical disk number, logical disk address, and storage disk of the data written to the stored data management table 16. The address and data length are written and the check flag is set. Then, when the parity check is completed, the check flag is reset.

4 and 5 are flow charts showing the operation of the disk array controller 10. Hereinafter, the operation of the disk array controller 10 will be described in detail with reference to FIGS.

First, referring to FIG. 4, the host processor 1
A process associated with writing data from No. 2 will be described.
When the disk array control device 10 receives a data write command from the host processor 12 (step 100),
The stored data management table 16 is checked to find a free area in the data storage disk 14 (step 110), and the data is stored in the data storage disk 14 in the received format.
(Step 120). At the same time, the disk array controller 10 divides the data to generate the parity data (step 130), and writes the divided data and the generated parity data to the corresponding addresses of the logical disks # 1, ... (Step 140). . When all data writing to the data storage disk 14 and the logical disks # 1, ... Is completed, the disk array control device 10 writes the data shown in FIG. 3 to the stored data management table 16 and sets the check flag (step 150. When the writing to the stored data management table 16 is completed, the disk control unit 22 sends a write completion report to the host processor 12. This completes the data write command processing (step 160).

Next, the data check of the logical disks # 1, ... In the background and the correction processing when an error is detected will be described with reference to FIG. Host processor 1
When there is no input / output command from 2, the disk array control device 10 checks the stored data management table 16,
Data is read from the addresses of the flagged logical disks # 1, ... And the presence or absence of a parity error is checked. (Steps 500, 510). Step 520 shows a branch depending on the check result. Step 5
When the parity error is detected in 20, the disk array control device 10 uses the stored data management table 16 to store the data storage disk 1 in which the data is stored.
The address and the data length of No. 4 are referred to and the data is read from the data storage disk 14 (step 530).
Subsequently, the disk array control device 10 proceeds to step 53.
The data read at 0 is divided to regenerate the parity data (step 540), and the divided data and the regenerated parity data are written to the logical disks # 1, ... (Step 550). When the writing to the logical disk # 1 is completed, the disk array control device 10 clears the corresponding check flag of the stored data management table 16 and ends the series of operations (step 560). On the other hand, if no parity error is detected in step 520, the process branches to step 560 to end the series of operations.

Next, differences between the conventional disk array control device 102 and the disk array control device 10 of this embodiment will be described with reference to FIGS. 1 and 6.

The cache memory (semiconductor memory) 106 in the prior art is the data storage disk (magnetic disk) 14 in this embodiment. Therefore, in this embodiment, a sufficient capacity for temporarily storing the write data can be obtained, so that various problems caused by the insufficient storage capacity can be solved. The reason is that the magnetic disk can easily increase the capacity by an order of magnitude more than the semiconductor memory.

On the other hand, in the prior art, the logical disk 120
The data is checked by reading the data from both the cache memory 106 and the cache memory 106 and comparing these data. Therefore, simply replacing the cache memory 106 in the conventional technique with the data storage disk 14 causes another problem that the processing time becomes longer than that in the conventional technique. Data storage disk 14
Data read speed from the cache memory 106
This is because it is much slower than Therefore, in this embodiment, the data is checked by performing the parity check only on the data read from the logical disk # 1. As a result, it is not necessary to read the data from the data storage disk 14 at the time of checking the data, so that the processing speed equivalent to that of the conventional technique is realized.

Needless to say, the present invention is not limited to the above embodiment. For example, the logical disks # 1, ... And the data storage disk 14 are not limited to magnetic disks, but may be optical disks, magneto-optical disks, or the like.

[0044]

According to the present invention, write data and its parity data are written to a logical disk, and
Write the write data to the data storage disk,
Incorrect data is generated by reading the write data and its parity data from the logical disk and executing a parity check, and if a parity error occurs, read the corresponding write data from the data storage disk and write it back to the logical disk together with the parity data. Etc. are discovered early, the reliability of the disk array can be improved.

Moreover, since the data storage disk is used, a sufficient capacity for temporarily storing the write data can be obtained, so that various problems caused by the insufficient storage capacity can be solved. In addition to this, by performing the parity check only on the data read from the logical disk, it is not necessary to read the data from the data storage disk at the time of checking the data, so that the delay in the processing speed can be prevented.

In other words, according to the present invention, the data storage disk and the storage data management table are provided, and the data for which the parity check in the background has not been completed is left in the data storage disk, so that the write data from the host processor is written. Make sure that is written to the disk array correctly, and if it was not, you can rewrite the correct data.
Therefore, it is possible to prevent the writing failure of the divided data, and even if the writing failure occurs, the data can be restored.

[Brief description of drawings]

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

FIG. 2 is an explanatory diagram showing a recording format of write data received from a host processor in the disk array control device of FIG.

FIG. 3 is an explanatory diagram showing a format of a stored data management table in the disk array control device of FIG.

FIG. 4 is a flowchart showing an operation of the disk array control device of FIG.

5 is a flowchart showing an operation of the disk array control device of FIG.

FIG. 6 is a block diagram showing a conventional disk array control device.

7 is a flowchart showing the operation of the disk array controller of FIG.

[Explanation of symbols]

10 Disk array controller 12 Host processor (upper device) 14 Data storage disk 16 Stored data management table 18 Data transfer means 20 disk array 22 Disk controller 24 Logical disk part # 1 to #n logical disk 1-N physical disk

Claims (9)

[Claims] 1. A unit for generating parity data for write data received from a host device, writing the write data and the parity data to a logical disk, and writing the write data to a data storage disk, and the write data. And a means for reading the parity data from the logical disk to perform a parity check, and if the result of the parity check is a parity error, the write data is read from the data storage disk to generate parity data, and the write data is written. And a means for rewriting the parity data thereof to the logical disk, a disk array controller. 2. When writing the write data and the parity data thereof to the logical disk, the write data and the parity data are divided and written to a plurality of physical disks that constitute the logical disk. Disk array controller. 3. The disk array control device according to claim 1, wherein the logical disk and the data storage disk are magnetic disks. 4. Parity data is generated for write data received from a higher-level device, the write data and the parity data are written to a logical disk, and the write data is written to a data storage disk. Parity data is read from the logical disk and a parity check is executed. If the result of the parity check is a parity error, the write data is read from the data storage disk to generate parity data. The write data and its parity data To write the data to the logical disk again. 5. The write data and its parity data are divided into a plurality of physical disks constituting the logical disk and written when the write data and its parity data are written to the logical disk. Disk array control method. 6. The disk array control method according to claim 4, wherein the logical disk and the data storage disk are magnetic disks. 7. A unit for generating parity data for write data received from a host device, writing the write data and the parity data to a logical disk, and writing the write data to a data storage disk, the write data and Means for reading the parity data from the logical disk and executing a parity check; and, if the result of the parity check is a parity error, the write data is read from the data storage disk to generate parity data, and the write data is written. And a disk array control program for causing a computer to function as means for rewriting the parity data to the logical disk. 8. The write data and the parity data thereof are written into the plurality of physical disks configuring the logical disk while writing the write data and the parity data thereof to the logical disk. Disk array control program. 9. The disk array control program according to claim 7, wherein the logical disk and the data storage disk are magnetic disks.