JP2002169660A - Data storage array apparatus, its control method, program recording medium and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a data array storage apparatus capable of storing error information not to lose real time properties and also not to interfere reading/ writing of data when reading/writing the data in real time or continuously from outside. SOLUTION: In the data storage array apparatus comprising a plurality of disk devices having storage regions composing of more than one block and writing or reading striped data and parity data into or from a group composed of blocks of each disk device, at every time of completing of writing, it is decided whether the error information stored in a memory should be written or not into a specified group, according to the decided result, the error information is read out from the memory to write into the specified group.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data storage array device comprising a plurality of data storage devices and having stored data having redundancy to improve reliability.

[0002]

2. Description of the Related Art Conventionally, there is a data storage array device (hereinafter abbreviated as an array device) in which the speed and reliability of data access are improved by distributing and storing data with a plurality of hard disks having redundancy. The array storage device is a so-called RAID (Redundant Array of Inexpensive Disks)
Well known as For example, in RAID-3,
The storage area of each hard disk is divided into a plurality of blocks, and one hard disk is used for parity data. Data is striped and recorded in groups of corresponding blocks on a plurality of hard disks. At this time, the parity data is recorded on the parity data hard disk. The parity data is usually generated as an exclusive OR of block data other than the parity block in the group.

[0003] If there is corrupted or invalid data in one of the blocks in the block group, the array device will break based on the paridiy data in the group and the correct block data. Alternatively, restore and overwrite invalid data. The array device described in US Pat. No. 5,581,690 records status information (error information) indicating the presence or absence of broken or invalid data for each block of each hard disk on one of the hard disks. There is disclosed a technique for preventing use of corrupted or invalid data as it is. That is, this array device records status information on the hard disk at the time of writing data, checks the status information just before reading data,
If there is no corrupted or invalid data, the data is read. If there is, the data in the group is restored if the data is corrupted or invalid.

[0004]

However, according to the above-mentioned prior art, the size of the error information has been increased to several megabytes on the order of the storage capacity and the data transfer rate of the hard disk. There is a problem in that the processing load for writing error information to the device or the nonvolatile memory is increasing.

For example, a material for digital broadcasting (DV (Dig)
In the case where external data is written to the array storage device in real time (or continuously) from the outside, there is a problem that the real-time property of the writing process of the material data is impaired (or not in time). This is because every time data and parity data are written to blocks in a group, a process of writing or updating error information to a disk device or a non-volatile memory occurs. Therefore, a process of writing or updating error information occupies the bus. This is because the rate at which data is written increases, and as a result, the bus occupation time permitted for the writing process of the material data is reduced.

SUMMARY OF THE INVENTION The present invention provides a data array storage device for storing error information so as not to impair the read / write operation of the data when the data is read / written from outside in real time or continuously. I do.

[0007]

According to a first aspect of the present invention, there is provided a data storage array device having a plurality of disk devices each having a storage area including a plurality of blocks, and striping a group of blocks of each storage device. A data storage array device for writing or reading the read data and parity data, comprising: a read / write means for writing or reading striping data and parody data to / from a group in accordance with an access request input from the outside; Error information generating means for generating error information indicating the presence / absence of an error generated as a result of writing for each block, and storing the generated error information in a memory, and an error stored in the memory each time writing by the reading / writing means is completed. Information should be written to a specific group Comprising determination means for determining whether, in accordance with the determination result of the determining means, and control means for controlling the reading and writing means to write the error information in the read specific group from memory. Here, a nonvolatile memory may be used instead of the specific group.

[0008] The error information may be bitmap data including a flag for each block indicating the presence or absence of a write timeout error or a medium error. Further, the determination unit may be configured to determine that the error information stored in the memory should be written to a specific group when there is no external access request.

The determination means counts the number of times of writing by the read / write means, and when the count reaches a threshold value, determines that error information stored in the memory should be written to a specific group. The count number may be configured to be initialized each time error information is written. Further, the determination means may change the threshold value according to the access frequency of the read / write means.

[0010] The data storage array device may further comprise a restoring means for restoring data of a block having an error from data of another block in the same group when the error information indicates the presence of an error. Means are
During restoration by the restoration unit, it may be configured to determine that the error information stored in the memory should not be written to a specific group.

Further, the determination means is configured to change the frequency of determining that the error information stored in the memory should be written to a specific group in accordance with the frequency of writing or reading by the reading / writing means. Is also good.

[0012]

(Embodiment 1) FIG. 1 shows a configuration of a data storage system according to Embodiment 1 of the present invention. The data storage system includes an external access device 22, an input device 23, and a data storage array device (hereinafter abbreviated as an array device) 20.

The external access device 22 issues a write request to the array device 20 based on data input from the input device 23, issues a read request to the array device 20, and transmits the read data to another device. Output (not shown). The input device 23 is, for example, a broadcast DV (Digital V).
(ideo) A camera that outputs video data (such as DV data) serving as a material of a broadcast program to the external access device 22.

The array device 20 includes an external access device 22
Access request (write request or read request)
In response to this, data writing / reading required for the internal disk array (hereinafter, referred to as external data writing / reading) is performed. At that time, when writing data, the array device 20 generates error information indicating the presence or absence of broken or invalid data, and temporarily stores it in the internal memory.
The error information stored in the internal memory is
The data is transferred to the disk array at any time. Hereinafter, writing of error information is referred to as bitmap writing to distinguish it from external data writing. The array device 20 is configured not to write error information directly to a disk every time external data is written, but to transfer the information to an internal disk array as needed while there is no access request. As described above, in the array device of the present embodiment, the process of writing error information to the disk is performed when there is no access request, when the frequency of data access is low, or when the bus to which the disk is connected is free. Configured to do so.
As a result, the external data write processing and the external data read processing are not pressed.

The specific hardware configuration of the array device 20 includes magnetic disk devices 1 to 5, a processor 6, and a memory 7, as shown in FIG. Each component is a bus 11
~ 16. The buses 11 to 15 are, for example, S
CSI (Small Computer System Interface) buses for connecting magnetic disk devices 1 to 5 (disk arrays) and SCSI controllers (not shown), respectively. The bus 16 is connected to a PCI (Peripheral Component Interco
nnect) bus and external interface 21 or SCSI
Connect to controller (not shown). The processor 6 and the memory 7 are connected on a bus 16.

As shown in FIG. 2, each of the storage areas of the magnetic disk drives 1 to 5 includes a number of partial areas (hereinafter, referred to as blocks). In FIG. 1, the magnetic disk 1 includes 100 blocks of blocks 101a to 200a. The same applies to the magnetic disks 2 to 5. One block is 512K
It is about the size of a byte. Corresponding blocks with the same numeral in the reference numerals are grouped among the magnetic disk devices to form parity groups 101p to 200p. The magnetic disk device 5 is a parity disk that stores parity data in each parity group in each block.

The processor 6 has a queue for queuing access requests input from the external access device 22. The processor 6 sequentially takes out access requests from the queue and accesses the magnetic disk devices 1 to 5 (external write / read processing). ). At this time, the processor 6 generates error information every time the external data write processing is performed, stores the error information in the memory 7, and transfers the error information from the memory to the magnetic disk device 5 when there is no access request.
The reason for transferring the error information is to ensure that the error information is preserved even when the power is turned off, even when reset. In addition, the processor 6 performs
Refer to the error information and execute data restoration (data reconstruction) as necessary.

The memory 7 includes, in addition to the program to be executed by the processor 6, information (error information) indicating the result of the external data write process to the magnetic disk devices 1 to 5, the write count of the external data, and a data reconstruction flag. Etc. are stored. Here, the error information is bitmap data indicating the presence / absence of a time-out error in external data write processing or a catalyst write error in the magnetic disk device by a flag for each block.
Is stored in the error table 7a. The write count number indicates the number of times external data write processing has been performed on the magnetic disk device since the previous time when the bitmap data of the error table 7a was written to the magnetic disks 1 to 5. The data reconstruction flag indicates whether data reconstruction is in progress. All of this information is updated by the processor 6.

FIG. 3A shows an example of bitmap data stored in the error table 7a. As shown in FIG. 3A, the bitmap data is stored in the magnetic disk devices 1 to 5.
Of each block includes 2-bit data (2 flags). The left flag indicates the presence or absence of a write error (that is, the presence or absence of invalid data) in the external data write process, and the right flag indicates the presence of a catalyst write error (such as the presence of a defective sector) in the external data write process. (The presence or absence of corrupted data)
Indicates that there is an error and there is no error that is 0). For a write error, it is necessary to restore the data of the parity group. For a catalyst write error, replace the magnetic disk device in which the error occurred (evacuation of data from the magnetic disk device,
(Replacement of a new magnetic disk, restoration of saved data to a new magnetic disk, and restoration of group data by parity data).

FIG. 3B shows another example of the bitmap data.
Shown in The bitmap data in FIG. 6B has one flag indicating whether or not there is a write error for each block of the magnetic disk devices 1 to 5. In another example, one flag may be represented by a logical sum of a write error and a catalyst write error. Hereinafter, for convenience of explanation, the error table 7a stores the bitmap data shown in FIG. 3B.

The operation of the data storage system configured as described above will be described. FIG. 4 is a flowchart showing an operation of the processor 6 for processing an external access request. The processor 6 constantly monitors whether there is an access request in the queue (whether it is not empty or empty) (step 41). If the access request is not empty and the first access request is a write request (step 42),
In accordance with the write request, a process of writing external data to the magnetic disk devices 1 to 5 is executed, a flag value is written in the error table 7a based on the write result, and the write count is incremented (step 44).
Each time the external data write processing is executed, the processor 6 writes a flag corresponding to a block in the written group in the error table 7a in the memory 7.

Further, when the write count is 3 (threshold) or more (step 45), there is no external access request in the queue (step 46), and no data reconstruction has occurred (step 47). And processor 6
Reads the bitmap data stored in the error table 7a and writes it to the parity group 200p of the magnetic disk devices 1 to 5 (step 48). At this time, the processor 6 determines whether or not data reconstruction has occurred based on the state of the data reconstruction flag. If the writing of the bitmap data to the parity group 200p succeeds, the processor 6 clears the write count to zero (steps 49 and 50). If the write count fails, the processor 6 leaves the write count as it is to the external access device 22. On the other hand, a failure is notified by returning an error status (step 51). Here, successful writing means that neither a writing error (timeout error) nor a catalyst writing error has occurred, and failure means that at least one error has occurred.

If the first access request in the queue is a read request (step 42: No), the processor 6 executes a read sequence (step 4).
3). In this read sequence, the processor 6
Refers to the flag of each block in the group to be read in the error table 7a, reads data from the magnetic disk devices 1 to 5 if the flag is not set, and if one error flag is set, , The data from other blocks (data and parity) is restored, and the restored data is output to the external access device 22 as read data.

As described above, according to the array device of the present embodiment, writing of bitmap data from the error table 7a to the magnetic disk device is not performed every time external data is written, but is written in the external device. This is performed only when the number of times of data write processing is smaller than the number of times of data write processing, and is performed when there is no access request. Therefore, it is possible to prevent the external data write processing and read processing from being overwhelmed.

In step 44 of FIG. 4, the processor 6 sets the value of the flag corresponding to the block in the group to be written in the error table 7a every time the external data is written to the magnetic disk device. However, the flag may be set to 1 only when a timeout error occurs in the external data write processing.

Further, in step 48, if no write error has occurred in the write processing of the external data three times in the past (that is, the number of times of the threshold value), the processor 6 stores the bitmap data in the magnetic disk drive 1
Writing to the parity group 200p within the range of ~ 5 may be omitted. In the flowchart of FIG. 4, the following conditions (a) to (b) are used as conditions for writing bitmap data to the parity group 200p of the magnetic disk devices 1 to 5.
Although (c) is described, at least one of (a) to (d) may be used as a condition.

(A) The write count is equal to or greater than the threshold value. (B) No external access request exists. (C) No data reconstruction has occurred. (D) The same number of times of writing the external data of the past as the threshold value. A write error occurred during processing. Also, the threshold value of the write count was not fixed,
The frequency of writing bitmap data to the parity group 200p may be adjusted by measuring the external access frequency (the number of external accesses per unit time) and increasing the threshold value of the write count when the external access frequency increases. .

Further, in this embodiment, the bitmap data is stored in the parity group 2 of the magnetic disk devices 1 to 5.
Although the data is written to 00p, the data may be written to one magnetic disk device instead, or may be written to a nonvolatile memory such as a flash memory. In the present embodiment, the description has been made on the assumption that the error table shown in FIG. 3B is used. However, in the case of the error table shown in FIG. If the write error flag is set, it is sufficient to notify the necessity of replacing the disk.

(Embodiment 2) FIG. 5 shows the configuration of a data storage system including an array device according to Embodiment 2 of the present invention. This figure is different from the array device 50 shown in FIG. 1 in that a processor 56 and a memory 57 are provided instead of the processor 6 and the memory 7, and a flash memory 58 is newly provided. The components having the same reference numerals are the same as those in FIG. 1, and the description thereof will be omitted.

The processor 56 executes the program in the memory 57 for the processor 6 to execute
The point that bitmap data is written to the flash memory 58 and the point that bitmap data stored in the flash memory 58 is checked immediately after power-on (when the system is started) and immediately after reset are added. Other than this, it is the same as the processor 6.

The memory 57 is different from the memory 7 in that the memory 57 stores an execution program for the processor 56 instead of the execution program for the processor 6. The flash memory 58 is a non-volatile memory for holding bitmap data. It should be noted that a battery 82 as shown in FIG.
May be provided.

FIG. 6 is a flowchart showing the operation of the processor 56 for processing an external access request. The flow of FIG. 13 is different in that step 68 is replaced with step 68 in FIG. Step 68
In, the processor 56 executes a write process to the magnetic disk device, sets the value of the flag corresponding to each block in the write target group to the bitmap data in the error table 7a based on the write result, Bitmap data is transferred to the magnetic disk devices 1 to
5 is written to the parity group 200p, and the parity group address (address of 200p) and the bitmap data to which the parity group is written are written to the flash memory 58. Other steps are the same as those in FIG.

FIG. 7 is a flowchart showing the operation of the processor 56 immediately after power-on and immediately after reset. In the figure, the processor 56 refers to the parity group address and the bitmap data stored in the flash memory 58 (step 71), and if the error flag is not set (step 72: No),
The process ends (to the processing in FIG. 6).

If two or more error flags are set (step 74: No), the processor 56 notifies the external access device 22 of the failure (steps 74 and 7).
5). If one write error flag is set (step 74: Yes), the processor 56 temporarily inhibits reading from the block in which the error has occurred, and simultaneously sets another block in the group to which the erroneous block belongs. Recover data from blocks (data and parity). (Step 76).

As described above, according to the array device 50 of the present embodiment, since the bitmap data is also redundantly stored in the flash memory 58, an error is found at the time of system startup and immediately after reset. If can be restored faster, and also improve the reliability of the system. The operation of the processor 56 at the time of starting the system and immediately after the reset may be performed as shown in the flowchart of FIG. 8 instead of the flowchart shown in FIG.

<U> 8 </ U> in FIG.
Reads the bitmap data from the parity group 200p of each of the magnetic disk devices 1 to 5, creates a read bitmap in which a read error (timeout error) and a catalyst read error in this read are flagged (step 81), and performs flashing. Memory 58
The logical sum of the bitmap data stored in the memory and the created read bitmap is calculated (step 82). If only one magnetic disk device has the error flag set (steps 83 and 84), the data is restored (step 82). Step 8
6). The logical sum may be obtained by calculating the logical sum of the bit map for one group included in the bit map data in the flash memory 58 and the read bit map for each group.

As a result, not only the bit map data of the flash memory 58 but also the read error flag included in the read bit map is reflected, so that the reliability can be further improved. In addition, in step 68 of FIG. 6, the processor 56 writes the bitmap data in the memory 57 into the parity group 200p of the magnetic disk device, and writes the parity group address (address of 200p) into which the bitmap data was written and the bitmap data. Although the data is written in the flash memory 58, the data may not be written to the flash memory 58 at this point and may be written immediately before the system is powered off.

Further, in the array device of the first embodiment, a flash memory is provided, and in step 48 of FIG.
Instead of storing the data in the flash memory, the data may be written in the flash memory. (Embodiment 3) In the array device according to Embodiment 2, the flash memory 58 stores bitmap data (hereinafter, referred to as first bitmap data) having the same contents as the parity group 200p. The flash memory 58 in the embodiment stores the second bitmap data. Here, the second bitmap data is data indicating whether or not an error has occurred when the processor 56 writes the bitmap data of the error table (that is, the first bitmap data) into the parity group 200p. It consists of flags for each block constituting 200p.

The hardware configuration of the data storage system according to the present embodiment is the same as that shown in FIG. 5, and a description thereof will not be repeated. <U></U> Figure 10
Is a flowchart showing an operation of the processor 56 for processing an external access request. The flow shown in FIG. 12 is different from step 68 in FIG.
8a and 68b.

At step 68a, the processor 56
Executes a write process to the magnetic disk device, sets a flag value corresponding to each block in the write target group to the first bitmap data in the error table 7a based on the write result, The bitmap data is transferred to the parity group 2 of the magnetic disk devices 1 to 5
Write to 00p.

In step 68b, the processor 68
Generates second bitmap data indicating the presence or absence of an error that has occurred during the writing of the first table to the parity group 200p, and stores the generated second bitmap data and the address of the parity group 200p in the flash memory. Write. FIG. 11 is a flowchart showing the operation of the processor 56 immediately after power-on and immediately after reset. FIG. 7 is different from FIG.
The difference is that steps 71a and 76a are provided instead of 1 and 76, and that step 73 is added.

In step 71a, the processor 56
Reads the parity group address and bitmap data stored in the flash memory 58. If the error flag is not set (step 72: N
o) In step 73, the processor 56 reads out bitmap data from the parity group of the magnetic disk devices 1 to 5 indicated by the parity group address and stores it in the error table 7a. <BR CLEAR="LEFT"> In step 76a, in addition to the operation in step 76, the processor 56 reads out bitmap data from the parity groups of the magnetic disk devices 1 to 5 indicated by the parity group addresses and stores them in the error table 7a.

However, steps 72, 74 to 76 in FIG.
Are performed on the flag group for each group corresponding to each parity group, while steps 72 to 72 in FIG.
In steps 75 and 76a, the processing is performed on the flag group corresponding to the parity group 200p. As described above, in the array device according to the present embodiment, the second bitmap data indicating the presence or absence of an error in the parity group storing the error information (first bitmap) is stored in the flash memory, and immediately after the power is turned on and immediately after the reset. Since the second bitmap data is checked first, the reliability of the error information can be improved.

FIG. 8 may be used instead of FIG.
In this case, in step 82, the logical sum of the read bitmap and the second bitmap may be obtained. Further, the present invention can be realized by the processor 6 or 56 executing a program in the memory 7 or 57 in the hardware configuration shown in FIG. 1 or FIG. The details of the program are as already described using the flowchart. This program can be downloaded (or installed) to another general disk array device similar to that shown in FIG. 1 or FIG. 5 via a program storage medium or a network, and can be executed.

<U></U> Instead of data restoration (data reconstruction) in each of the above embodiments, the following (a) or (b) may be selectively used.
(A) At the time of reading in response to a read request, if the error flag is set, the processor 6 or 56 restores the data of the block having an error from the data of another block in the same group, and reads the restored data. Respond externally as data. In this case, the restored data is not written to the block (magnetic disk device).

(B) When the error flag is set, the processor 6 or 56 restores the data of the erroneous block from the data of another block in the same group and restores the data to the block (magnetic disk device). Write the written data. <U></U> For example, if (a) is performed at the time of reading a read request, it is possible to respond to the read request at high speed. In this case, the process (b) may be performed before the power is turned off or during idle time (when no access request is continuously generated).
(B) may be performed immediately after the power is turned on or immediately after the reset.

[0047]

The data storage array device of the present invention has a plurality of disk devices each having a storage area consisting of a plurality of blocks, and strips data and parity data for a group of blocks of each storage device. A data storage array device for writing or reading data, comprising: a read / write means for writing or reading striping data and parity data to / from a group according to an externally input access request; and an error generated as a result of writing by the read / write means. Generating error information indicating the presence / absence of each block and storing the generated error information in a memory; and storing the error information stored in the memory in a specific group each time the writing by the reading / writing means is completed. Determination means for determining whether to write , According to the judgment result of the judging means, and control means for controlling the reading and writing means to write the error information in the read specific group from memory. Here, a nonvolatile memory may be used instead of the specific group.

According to this configuration, the error information is not written in a specific group every time data is written, but may not be written immediately after data writing according to the determination result. This makes it possible to read and write data externally in real time or continuously,
The error information can be stored so as not to impair the real-time property and hinder the data writing.

Here, the error information may be bitmap data including a flag for each block indicating the presence or absence of a write timeout error or a medium error. The determination means may be configured to determine that the error information stored in the memory should be written to a specific group when there is no external access request.

According to this configuration, the data storage array device writes the error information to the specific group when no external access request exists, so that the next access request already input by the writing of the error information is written. Has the effect of not oppressing the execution. Further, the determination unit counts the number of times of writing by the read / write unit, and determines that the error information stored in the memory should be written to a specific group when the count reaches a threshold value. The configuration may be such that initialization is performed every time error information is written.

According to this configuration, the data storage array device writes the error information to the specific group at a rate of once each time the same number of writes as the threshold value is performed, so that the execution of the access request is not hindered. effective. Here, the determination means may change the threshold value according to the access frequency of the read / write means.

According to this configuration, since the threshold value is dynamically changed according to the access frequency, the balance between data access according to the access request and writing of error information can be adjusted. Further, the data storage array device further includes a restoring means for restoring data of an erroneous block from data of another block in the same group when the error information indicates the presence of an error, wherein the determining means comprises: During restoration by the restoration unit, it may be configured to determine that the error information stored in the memory should not be written to a specific group.

According to this configuration, the data storage array device does not write error information to a specific group while restoring a block having an error, so that an access request can be prioritized. Here, the determination unit is configured to write or read by the read / write unit in accordance with the frequency of
The frequency of determining that the error information stored in the memory should be written to a specific group may be changed.

According to this configuration, the data storage array device dynamically changes the frequency of the affirmative determination in accordance with the frequency of the access request, thereby enabling data access according to the access request and writing of error information. Balance can be adjusted. The data storage array device may further include a non-volatile memory having an area for storing error information, and the control unit may write the error information to the non-volatile memory immediately before turning off the power of the data storage array device. Good.

Here, the error information is bitmap data including a flag for each block indicating the presence or absence of a write timeout error, and the data storage array device is further provided immediately after power-on or immediately after reset of the data storage array device. When the bitmap data is read from the non-volatile memory and only one of the flags corresponding to the blocks constituting the group indicates an error, the reading of the block is prohibited and the restoring means for restoring the block data is provided. It may be provided.

According to this configuration, if the power is once turned off before the data is restored while there is an error, the operation after the restart can be prevented. Here, the error information is bitmap data including a flag for each block indicating the presence / absence of a write timeout error, and the data storage array device further includes the data storage array device immediately after power-on or immediately after reset. By reading the bitmap data from the specific group, a read bitmap including a flag indicating a read error for each block in the specific group is created, and the created read bitmap and the bitmap data stored in the nonvolatile memory are created. The logical sum with the partial bitmap corresponding to each of the groups is taken, and if only one bitmap obtained as a result of the logical sum indicates an error, the reading of the block is prohibited and the block data May be provided with restoring means for restoring.

According to this configuration, higher reliability can be obtained only by calculating the logical sum of the error information (bitmap data) and the read bitmap. The error information is bitmap data including a flag for each block indicating the presence or absence of a write timeout error. The data storage array device has a non-volatile memory, and the control unit controls an error to a specific group. A configuration may be employed in which a second bitmap indicating the presence or absence of a write error in writing information is generated for each block of a specific group, and the generated second bitmap data is written to a nonvolatile memory.

According to this configuration, since the error information is duplicated in the nonvolatile memory and the specific group, the reliability of the error information itself can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram of a data storage system including a data storage array device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing blocks and parity groups in a magnetic disk device.

FIG. 3A shows an example of a write result bitmap table. FIG. 13B is a diagram illustrating another example of the write result bitmap table.

FIG. 4 is a flowchart showing an operation of the processor 6;

FIG. 5 is a block diagram of a data storage system including a data storage array device according to a second embodiment of the present invention.

FIG. 6 is a flowchart showing the operation of the processor 56.

FIG. 7 is a flowchart showing an operation immediately after power-on and immediately after reset in a processor 56 according to the second embodiment.

FIG. 8 is a flowchart showing another operation of the processor 56 immediately after power-on and immediately after reset.

FIG. 9 is a block diagram of a battery backup SRAM.

FIG. 10 is a flowchart showing an operation of a processor 56 according to the third embodiment.

FIG. 11 is a flowchart showing an operation of the processor 56 immediately after power-on and immediately after reset.

[Explanation of symbols]

 1-5 magnetic disk device 6 processor 7 memory 7a error table 11-16 bus 20 array device 21 external interface 22 external access device 23 input device 50 array device 56 processor 57 memory 58 flash memory 68 processor 68a step 68b step 81 SRAM 82 battery 101a-200a block 101p-200p parity group

Claims (28)

[Claims] 1. A data storage array having a plurality of disk devices having a storage area consisting of a plurality of blocks, and writing or reading striped data and parity data to and from a group consisting of blocks of each storage device. An apparatus, comprising: a read / write unit for writing or reading striping data and parody data to and from a group in accordance with an externally input access request; and error information indicating for each block whether or not an error has occurred as a result of writing by the read / write unit Error information generating means for generating the error information and storing the generated error information in a memory; and determining whether or not the error information stored in the memory should be written to a specific group each time writing by the reading and writing means is completed. Means according to the judgment result of the judgment means Data storage array system, characterized in that it comprises a control means for controlling the reading and writing means to write the error information from the memory to the read specific group. 2. The data storage array device according to claim 1, wherein the error information is bitmap data including a flag for each block indicating the presence or absence of a write timeout error or a medium error. 3. The method according to claim 1, wherein the determination unit determines that the error information stored in the memory should be written to a specific group when no external access request exists. A data storage array device as described. 4. The data storage array device includes a queue for queuing an external access request, the read / write unit inputs an external access request via the queue, and the determination unit determines that the access request is 2. The data storage array device according to claim 1, wherein when no queue is present, it is determined that the error information stored in the memory should be written to a specific group. 5. The determination unit counts the number of times of writing by the read / write unit, and determines that error information stored in a memory should be written to a specific group when the count reaches a threshold value. 2. The data storage array device according to claim 1, wherein the count number is initialized each time error information is written. 6. The data storage array device according to claim 5, wherein said determination means changes a threshold value according to an access frequency of said read / write means. 7. The data storage array device further comprises a restoring means for restoring data of an erroneous block from data of another block in the same group when the error information indicates the presence of an error; 2. The data storage array device according to claim 1, wherein during the restoration by the restoration unit, it is determined that the error information stored in the memory should not be written to a specific group. 8. The method according to claim 1, wherein the determining unit writes the error information stored in the memory to a specific group when there is no external access request and the recovery unit is not performing the recovery. The data storage array device according to claim 7, wherein the determination is performed. 9. The method according to claim 1, wherein the determining unit changes a frequency of determining that the error information stored in the memory should be written to a specific group according to a frequency of writing or reading by the reading / writing unit. The data storage array device according to claim 1. 10. The data storage array device further includes a non-volatile memory having an area for storing error information, and the control unit writes the error information to the non-volatile memory immediately before turning off the power of the data storage array device. 2. The data storage array device according to claim 1, wherein: 11. The error information is bitmap data including a flag for each block indicating whether or not a write time-out error has occurred. The data storage array device further includes: If the bitmap data is read from the non-volatile memory and only one of the flags corresponding to the blocks constituting the group indicates an error, readout of the block is inhibited and the block data is restored. The data storage array device according to claim 10, wherein: 12. The error information is bitmap data including a flag for each block indicating the presence or absence of a write timeout error. The data storage array device further includes: By reading bitmap data from the specific group, a read bitmap including a flag indicating a read error for each block in the specific group is created, and the created read bitmap and a bitmap stored in the non-volatile memory Logically OR with a partial bitmap corresponding to each group in the data, and if only one bitmap obtained as a result of the logical sum indicates an error, reading of the block is prohibited, and It is characterized by comprising a restoration means for restoring data 11. The data storage array device according to claim 10, wherein: 13. The error information is bitmap data including a flag for each block indicating the presence / absence of a write timeout error. The data storage array device has a non-volatile memory; 2. A second bitmap indicating whether or not there is a write error in writing error information to a specific group is generated for each block of a specific group, and the generated second bitmap data is written to a nonvolatile memory. Data storage array device. 14. The data storage array device further reads the second bitmap data from the non-volatile memory immediately after turning on the power of the data storage array device or immediately after resetting, and reads out only the flags in the second bitmap. 14. The data storage array device according to claim 13, further comprising a restoring unit that, when one of them indicates an error, prohibits reading of the block and restores the block data. 15. The data storage array device further includes a flag indicating a read error for each block in the specific group by reading bitmap data from the specific group immediately after power-on or reset of the data storage array device. And a logical sum of the created read bitmap and the second bitmap data stored in the non-volatile memory is obtained. Only one error is found in the bitmap obtained as a result of the logical sum. 14. The data storage array device according to claim 13, further comprising a restoring means for prohibiting the reading of the block and restoring the block data in the case of indicating. 16. The data storage array device further includes a non-volatile memory having an area for storing error information, wherein the control unit reads the error information from the memory and writes the error information to a specific group according to a result of the determination by the determination unit. 2. The data storage array device according to claim 1, wherein the error information is written into the nonvolatile memory at the same time as controlling the read / write means. 17. The data storage array device according to claim 1, further comprising a notifying means for notifying to the outside when the reading / writing means fails to write the error information to the specific group. A data storage array device as described. 18. An array storage device having a plurality of disk devices having a storage area made up of a plurality of blocks, and writing or reading striped data and parity data to and from a group made up of blocks of each storage device. Reading and writing means for writing or reading striping data and paridiy data to and from a group in accordance with an access request input from the outside; and error information indicating, for each block, the presence or absence of an error generated as a result of writing by the reading and writing means. Error information generating means for generating and storing the generated error information in a memory; and determining means for determining whether or not the error information stored in the memory should be written to a non-volatile memory every time writing by the reading and writing means is completed. According to the judgment result of the judgment means, Data storage array system, characterized in that it comprises the information from a memory and control means for writing the read non-volatile memory. 19. The error information is bitmap data including a flag for each block indicating the presence / absence of a write timeout error or a medium error. 19. The data storage array device according to claim 18, wherein: 20. The method according to claim 18, wherein the determination unit determines that the error information stored in the memory should be written to the nonvolatile memory when there is no external access request. A data storage array device as described. 21. The data storage array device has a queue for queuing an external access request, the read / write unit inputs an external access request via the queue, and the determining unit determines that the access request is 19. The data storage array device according to claim 18, wherein when no queue exists, it is determined that the error information stored in the memory should be written to the nonvolatile memory. 22. The determination means counts the number of times of writing by the read / write means, and when the count reaches a threshold value, determines that error information stored in the memory should be written to the nonvolatile memory; 19. The data storage array device according to claim 18, wherein the count number is initialized each time error information is written. 23. The data storage array device further comprises a restoring means for restoring data of an erroneous block from data of another block in the same group when the error information indicates the presence of an error; 19. The data storage array device according to claim 18, wherein during the restoration by the restoration unit, it is determined that the error information stored in the memory should not be written to the nonvolatile memory. 24. The determination means, when the read / write means has not received at least one access request from outside and is not being restored by the restoration means, writes the error information stored in the memory to the nonvolatile memory. 24. The data storage array device according to claim 23, wherein the data storage array device is determined to be: 25. The method according to claim 25, wherein the determining unit changes the frequency of determining that the error information stored in the memory should be written to a specific group according to the frequency of writing or reading by the reading / writing unit. Claim 18
A data storage array device as described. 26. An array storage device having a plurality of disk devices having a storage area made up of a plurality of blocks, and writing or reading striped data and parity data to and from a group made up of blocks of each storage device. A writing / reading step of writing or reading striping data and parity data to / from a group in accordance with an externally input access request, and indicating, for each block, the presence or absence of an error generated as a result of writing by the reading / writing step. An error information generating step of generating error information and storing the generated error information in a memory, and determining whether or not the error information stored in the memory should be written to a specific group each time the writing in the read / write step is completed. A determining step of According to the judgment result of the constant unit, the control method characterized in that it comprises a write step of writing the error information from the memory to the read specific group. 27. An array storage device having a plurality of disk devices having a storage area made up of a plurality of blocks, and writing or reading striped data and parity data to and from a group made up of blocks of each storage device. A program recording medium for recording a program to be executed by a processor in the apparatus, the program comprising: a read / write step of performing writing or reading of striping data and parity data with respect to a group according to an externally input access request; An error information generation step of generating error information indicating, for each block, the presence or absence of an error generated as a result of the writing by the step, and storing the generated error information in the memory; D Causing the processor to execute a determining step of determining whether or not the error information should be written to a specific group, and a writing step of reading error information from a memory and writing the error information to a specific group according to the determination result of the determining means. Program recording medium. 28. An array storage device having a plurality of disk devices having a storage area made up of a plurality of blocks, and writing or reading striped data and parity data to and from a group made up of blocks of each storage device. A program executed by a processor within the group, the program comprising: a read / write step of writing or reading striping data and parity data to / from a group according to an externally input access request; and a generation of a write result by the read / write step. An error information generation step of generating error information indicating the presence / absence of each error in each block, and storing the generated error information in a memory; and, each time the writing by the read / write step is completed, the error information stored in the memory is identified. To the group A determination step of determining whether or not to burn them, according to the determination result of the determining means, a program for causing to execute a write step of writing error information from the memory to the read specific group to the processor.