JP2003303057A - Method for data recovery and disk array controller in disk array apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve quick detection and recovery of media failures by executing media check processing with a distinction between areas actually in use and not in use at a file system among disk areas of a disk array. <P>SOLUTION: In a media check processing for checking partial failures of a plurality of HDDs (hard disk drives) composed of the disk array, it is determined whether each stripe in the disk areas of the disk array is used by the file system or not based on a disk source management table (S41, S42), and a media check including data reading from the HDDs is executed only for a stripe (a first stripe) used by the file system (S43, S44). <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a RAID (Redundant
The present invention relates to a data recovery method in a disk array device having an ant array of Inexpensive Disks), and particularly a data recovery method suitable for detecting and repairing a partial failure (media failure) of a disk drive of a member forming the disk array. And a disk array controller.

[0002]

2. Description of the Related Art Generally, a disk array device is composed of a plurality of disk drives, that is, a magnetic disk device (hereinafter referred to as HD
D)) and a disk array controller for controlling access to each HDD (member HDD) in this disk array, and each HDD is moved in parallel to distribute read / write. It is known as an external storage device in which the speed of access is increased by executing it and the reliability is improved by a redundant configuration.

The disk array controller generates redundant data as data correction information for the write data transferred from the host computer, and the plurality of H
It is adapted to write to any one of the DDs. As a result, when one of the plurality of HDDs fails, the redundant HDD data and the remaining HDD data can be used to restore the failed HDD data.

As one of data redundancy methods, RAI
The D method is known. In the RAID method, RAID
Various RAIs in relation to redundant data and redundant data
Classified as D level. Levels 3 and 5 are typical RAID levels.

At level 3 (RAID level 3), update data (write data) transferred from the host computer
To generate parity data as redundant data by performing an exclusive OR operation between the divided update data, and the original parity written in any of the plurality of HDDs with the parity data. Update the data.
On the other hand, at level 5 (RAID level 5), the update data (new data) transferred from the host computer and the pre-update data (old data) stored in the HDD internal area where the update data is stored. , The parity data updated by performing an exclusive OR operation with the parity data before update (old parity data) stored in the area of another HDD corresponding to the storage destination of the update data. New parity data) is generated, and the original parity data is updated with the updated parity data.

In a disk array device having such a RAID structure, when a member HDD in the disk array fails, data is read from the HDDs other than the failed HDD for each stripe, which is a unit for managing the disk area of the disk array. By the RAID function of reading and performing an exclusive OR operation of those data, data in all areas of the failed HDD can be restored in the HDD used in place of the failed HDD. The HDD used in place of the failed HDD is an HDD used by replacing the failed HDD, or a spare HDD that is previously connected to the disk controller and is assigned as a substitute for the failed HDD.

[0007]

As described above, the RAID
In the disk array device with the configuration, even if the HDD fails
The data of the failed HDD can be restored to the original state.

However, in the conventional disk array device,
In order to restore the data in the failed HDD to its original state, the data in all HDD areas are restored by the RAID function. Therefore, as the HDD capacity has increased in recent years, there has been a problem that it takes a very long time to restore data.

During data restoration, data redundancy due to RAID is generally lost. For this reason, the longer the restoration time, the more likely it is that another HDD failure will occur, making the data unrecoverable and increasing the risk of data loss. Further, in order to recover the data of the failed HDD, it is necessary to read all the areas of the other HDDs. If a media failure (partial failure of the HDD) occurs in the HDD that is the target of reading, multiple failures of the HDD occur and data cannot be recovered. Therefore, the disk area (disk volume) of the disk array is blocked. Or, the data restoration process cannot be continued. This reduces the reliability of the disk array device.

However, the inventor of the present invention has come to the idea that the data recovery process is not necessarily required except for the HDD area actually used by the file system of the host computer that uses the disk array device. Therefore, if the data of the HDD (disk drive) is restored only for the HDD area (disk area) actually used by the file system, the time required for the data restoration process is shortened, and the risk of HDD failure or the like is reduced. It is possible to reduce the sex. However, the conventional disk array device cannot know the HDD area actually used by the file system of the host computer, and therefore the HDD data is restored only for the HDD area used by the file system. It is not possible.

In addition, a HDD media failure, that is, HD
For the purpose of early detection and repair of a sector block failure, which is a partial failure of D, an HDD media inspection process for periodically reading and inspecting the contents of the HDD is generally known. However, in the conventional disk array device, since the HDD area used by the file system is unknown, it is necessary to perform a read test on the entire area of the HDD. For this reason, there is a problem that the inspection takes a very long time, as in the case of the restoration processing of the data of the failed HDD.

Further, when a media failure is found, it is necessary to perform processing for replacing the sector block at the failure location with another sector block and restore the replacement source data in the replacement destination block. . However, this data restoration can be performed only when the redundancy of data is secured by the RAID function. Therefore, the HDD
While data is being restored due to a failure, the HDD is
Even if a media failure is found in the media inspection process, it cannot be repaired, and the reliability of the disk array device is reduced.

The present invention has been made in consideration of the above circumstances, and an object thereof is to perform a media inspection by discriminating an area actually used in a file system and an area not used in a disk area of a disk array. By executing the processing, it is possible to realize the early detection and early repair of the media failure, thereby improving the reliability of the disk array device.

[0014]

The present invention includes a logical block used by the file system or a logical block used by the file system from a host computer having a file system for managing the disk area of the disk array in logical block units. A step of acquiring first disk resource management information indicating a stripe and indicating whether each stripe in the disk area includes a logical block used by a file system from the first disk resource management information Generating a second disk resource management information and holding it in the disk array device; and when receiving a data write request from the host computer,
The step of identifying the stripe to which the data write destination specified by the request belongs, and the identified stripe is the first stripe including the logical block used by the file system or is used by the file system. A step of determining whether the stripe is a second stripe that does not include a logical block based on the second disk resource management information held in the disk array device; and the specified stripe is the second stripe. And the storage content of each disk drive in the disk array, the step of updating the second disk resource management information so as to indicate that the specified stripe is the first stripe. To detect a partial failure of the disk drive. When performing the media inspection process, it is determined whether all the stripes in the disk area of the disk array are the first stripes or the second stripes based on the second disk resource management information. To perform the media check including reading data from the disk drive only for the first stripe, that is, the stripe used by the file system, and the data at the location where the failure is detected by this media check. And a step of repairing by a RAID function.

In such a configuration, the first disk resource management information is acquired from the host computer, for example, by the disk array controller in the disk array device, and the second disk resource management information is generated from the information. It is held in, for example, a disk array controller in the disk array device. The media inspection process here is
Only the stripe (second stripe) indicated by the second disk resource management information that the file system contains the logical block actually used is executed.

In this way, by extracting only the stripes used in the file system and inspecting the medium, the processing time required for the inspection can be shortened. Further, by shortening the processing time, as a result, it becomes possible to detect the media failure earlier, and the reliability of the disk drive can be improved. This also improves the reliability of the disk array device.

The first disk resource management information acquired from the host computer represents the usage status of the logical block until immediately before the acquisition of the information. Therefore, file updates (data writing to the disk array device) that occurred after that are not reflected. However, acquiring the latest first disk resource management information from the host computer every time the file is updated results in a significant decrease in performance.

Therefore, when the disk array device receives a data write request from the host computer, that is, when a file update occurs in the disk array device,
Based on the data write request, the disk array device itself may update the second disk resource management information held in the disk array device so as to represent the latest usage status of the stripe.

Further, according to the present invention, the above-mentioned media inspection process is sequentially executed for all stripes, and when a partial failure of the disk drive is detected,
It is determined whether the stripe including the location where the failure is detected is the first stripe or the second stripe, and in the case of the first stripe, that is, the stripe used by the file system, the failure is detected. When a disk drive other than the detected disk drive is normal, the data at the location where the failure is detected is restored by the RAID function, and in the case of the second stripe, that is, the stripe not used by the file system. In this case, the feature is that the data at the location where the failure is detected is restored by the fixed data.

As described above, the failure occurs only when the stripe including the location where the failure is detected by the media inspection is used by the file system and when the disk drive other than the disk drive where the failure is detected is normal. The data at the detected location is restored by the RAID function. On the other hand, if the stripe containing the location where the failure was detected is not used by the file system, it is not necessary to hold the data in that stripe, so a disk other than the disk drive where the failure was detected is not required. Irrespective of whether the drive is normal, ie R
Regardless of whether or not data redundancy by the AID function is secured, by forcibly repairing the data at the location where the failure is detected with fixed data, the possibility of being able to repair the media failure is greatly improved. The reliability of the disk drive is improved, and thus the reliability of the disk array device is also improved.

The invention related to the above method is also realized as an invention related to a device (disk array controller, or a disk array device including the same disk array controller, or a computer system including the same disk array device).

[0022]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a computer system including a disk array device according to an embodiment of the present invention. The computer system shown in FIG. 1 includes a host computer 10 and a disk array device 20 used by the host computer 10.

The host computer 10 is the host computer 1
A file system 11 for managing files stored in (a disk area of) the disk array device 20 connected to the disk array device 0 is provided. This file system 11 is O
It is a part of the function provided by S (operating system).

The host computer 10 has a file system 11 for all logical blocks (fixed length blocks composed of a plurality of consecutive physical sector blocks) in the disk area of the disk array device 20.
Management table (hereinafter, disk resource management table) that indicates whether it is used (that is, valid data is stored) or is not used (that is, no data is stored and new data can be stored) 12) is a storage device of the computer 10, for example, H
It is held in DD (not shown). The storage device may be the disk array device 20. Also, the disk resource management table 12 may indicate whether or not all physical sector blocks in the disk area are used by the file system 11.

A disk resource management table 12 is acquired from the file system 11 in the host computer 10 (a storage device of the host computer) at a predetermined timing, and a disk resource management information list 90 (FIG. 9)) is installed in the disk array device 20. The disk resource management information list 90 is composed of a set of logical block numbers as identification information of all logical blocks used by the file system 11 as described later.

The disk array device 20 comprises a disk array 21 and a disk array controller 22. The disk array 21 includes a plurality of disk drives connected to the disk array controller 22, for example, four HDDs (magnetic disk devices) 210-0 to 210.
It consists of 0-3. Disk array controller 22
Also connected to this is a spare HDD (not shown) that is assigned as a backup disk when any of the HDDs 210-0 to 210-3 fails.

The disk array controller 22 controls access to each HDD 210-0 to 210-3 in the disk array 21. Disk array controller 2
2 includes a microprocessor 221 which is a main control unit of the controller 22 and a memory 222. The memory 222 stores a control program 222a executed by the microprocessor 221. A disk resource management table area 222b is secured in the memory 222. This disk resource management table area 2
22b is used to store the disk resource management table 120 generated based on the disk resource management information list 90 transmitted from the host computer 10.

In this embodiment, the disk array device 20
Shall be used at the RAID 5 level. In this case, the HDDs 210-0 to 210-3 are used for storing data and parity data (redundant data) (for data / parity disk). When the disk array device 20 is used at the RAID3 level, H
Three of the DDs 210-0 to 210-3 are allocated for data storage (for data disk), and the other one is allocated for parity data storage (for parity disk).

In the disk array device 20 (inside the disk array controller 22), the HDDs 210-0 to 210
The disk area of the disk array 21 realized by -3 is managed by being divided into a plurality of stripes 23 as shown in FIG. The size of this stripe 23 is 1 HDD
It is generally set to about 64 K (kilo) bytes to 256 K bytes. The stripe 23 is composed of at least one logical block 24. The logical block 24 is the minimum unit when the file system 11 of the host computer 10 manages the disk area of the disk array device 20 (the disk array device 20 therein). That is, the disk area of the disk array device 20 is managed in units of stripes 23 in the disk array device 20, and in the host computer 10, the logical block 2 is managed.
It is managed in units of 4. Usually, the size of the logical block is about 1 Kbyte to 8 Kbytes. Logic block 24
Is composed of a plurality of consecutive physical sector blocks 25. The size of this sector block 25 is generally 512 bytes.

FIG. 3 shows the disk resource management table 12 held in the host computer 10 and the disk resource management table 12 stored in the disk resource management information area 222b in the memory 222 of the disk array controller 22.
An example of the data structure of 0 is shown.

The entry number determined by the arrangement order of the entries in the disk resource management table 12 directly represents the logical block number. A flag indicating whether or not a logical block represented by a logical block number unique to the entry is used by the file system 11 is set in each entry of the table 12. A pair of the logical block number and the flag may be set in each entry of the disk resource management table 12.

On the other hand, the entry number determined by the order of arrangement of the entries in the disk resource management table 120 directly represents the stripe number. A flag indicating whether or not a stripe (at least one of logical blocks included in the stripe) represented by a stripe number unique to the entry is used by the file system 11 is set in each entry of the table 120. There is. In addition, in each entry of the disk resource management table 120,
A pair of a stripe number and the above flag may be set.

Next, the operation of the computer system configured as shown in FIG. 1 will be described as follows: (1) processing when the disk resource management information list is transmitted from the host computer 10, (2) host computer 10
The process for issuing a data write request from (1), (3) data recovery process in the disk array device 20, and (4) HDD media failure inspection process in the disk array device 20 will be sequentially described as an example.

(1) Processing when transmitting the disk resource management information list from the host computer 10 First, the processing when transmitting the disk resource management information list from the host computer 10 will be described with reference to the flowchart of FIG.

The host computer 10 is held in (the storage device of) the computer 10 at a predetermined timing according to a dedicated software 13 installed in (the storage device of) the computer 10 at a predetermined timing. The disk resource management table 12 is acquired from the file system 11. Then, the host computer 10 creates a disk resource management information list 90 shown in FIG. 9 from the disk resource management table 12 and sends the list 90 to the disk array device 20. At this time, in order to prevent the contents of the disk resource management table 12 from changing during transmission of the disk resource management information list 90, it is preferable that the host computer 10 considers that no file update occurs. Further, the disk resource management information list 90 may have an extremely large size, in which case it takes a long time to transmit the list 90. Therefore, in order to prevent the transmission of the disk resource management information list 90 from affecting the efficiency of the host computer 10, the transmission timing is set to a fixed period such as when the host computer 10 is started up or at night when the load of the host computer 10 is low. Good to set.

Now, the disk resource management information list 9 transmitted from the host computer 10 to the disk array device 20
As shown in FIG. 9, 0 is a logical block 24 when the file system 11 in the host computer 10 handles the disk area (disk volume) of the disk array device 20.
, And a set of block numbers (logical block numbers) 92, 92 ... Of the logical blocks 24 used by the file system 11 among all the logical blocks 24 in this disk area. Thus, in the disk resource management information list 90, the logical blocks 24 that are not used by the file system 11 are
Information (logical block number) is not included because the size of the list 90 is reduced to shorten the time required to transmit the list 90 from the host computer 10 to the disk array device 20. is there. Normal,
The ratio of the area used by the file system 11 to the disk area of the disk array device 20 is small. In such a case, not including the information of the logical block 24 not used by the file system 11 in the disk resource management information list 90 is particularly effective in reducing the transmission time of the list 90. The disk resource management table 12 itself may be transmitted from the host computer 10 to the disk array device 20. Further, one of the disk resource management information list 90 or the disk resource management table 120 having the smaller data amount may be transmitted. in this case,
Information indicating which of the list 90 and the table 120 is to be transmitted may be added and transmitted.

When the disk resource management information list 90 is transmitted from the host computer 10, the disk array controller 22 (microprocessor 221 provided therein) in the disk array device 20 receives the list 90 (step S1). . Then, the disk array controller 22 determines the logical block number of the logical block not used by the file system 11 based on all the logical block numbers included in the disk resource management information list 90, and the disk array device 20. For all the logical blocks in the disk area, the disk resource management table 12 shown in FIG. 3 in which entries indicating whether or not the block is used are arranged in ascending order of the logical block numbers indicating the block, for example. It is restored (step S2).

Next, the disk array controller 22
The logical block number managed by the host computer 10 and the stripe number managed by the disk array device 20 are associated (step S3). This association is performed as follows.

First, the disk array controller 22
The "number of logical blocks per stripe" is managed by the "stripe size" managed by itself and the host computer 10
From the “logical block size” 91 included in the disk resource management information list 90 transmitted from the device, the number of logical blocks per stripe = “stripe size” / “logical block size” is calculated.

Next, the disk array controller 22
For every "logical block number", its "logical block number" and "number of logical blocks per stripe"
From the logical block 24 indicated by the "logical block number"
The "stripe number" indicating the stripe 23 including "is included" is calculated by the integer part of "stripe number" = {"logical block number" / "number of logical blocks per stripe"}. For example, when the “number of logical blocks per stripe” is “4”, the “stripe number” of the stripes including the logical blocks whose “logical block number” is “0” to “3” is “0”. As a result, the logical block number managed by the host computer 10 and the stripe number managed by the disk array device 20 are associated with each other.

When the disk array controller 22 associates the logical block number with the stripe number,
From the association result and the restored disk resource management table 12, for all the stripes in the disk array device 20, is the stripe used by the file system 11 in the ascending order of the stripe numbers indicating the stripes? A disk resource management table 120 shown in FIG. 3 in which an entry indicating whether or not is arranged is arranged.
Is created (step S4). Here, a stripe including at least one logical block used by the file system 11 is determined to be used by the file system 11, and a flag indicating busy is set in the corresponding entry. On the other hand, a stripe that does not include a logical block used by the file system 11 is determined not to be used by the file system 11, and a flag indicating unused (not used) is set in the corresponding entry. .

After creating the disk resource management table 120, the disk array controller 22 overwrites the table 120 on the disk resource management table area 222b in the memory 222 (step S5).

As described above, when the logical block number and the stripe number are associated with each other in the disk array device 20, it is not necessary for the host computer 10 to consider the stripe size unique to the disk array device 20. However, in the disk array device 20, it is necessary to consider the size of the logical block unique to the host computer 10.

On the other hand, if the disk resource management information list 90 having the data structure shown in FIG. 10 is used in place of the disk resource management information list 90 shown in FIG. 9, the disk array device 20 has a logic unique to the host computer 10. There is no need to consider the block size. The disk resource management information list 100 having the structure shown in FIG. 10 is composed of a set of stripe numbers 101, 101 ... Which indicate stripes including the logical block 24 used by the file system 11. However, in order to prepare the disk resource management information list 100 having the structure shown in FIG. 10 in the host computer 10, the host computer 10 needs to prepare the disk array device 20.
It is necessary to acquire the stripe size in advance from the disk array controller 22 in the internal disk array controller 22 and to associate the logical block number with the stripe number in accordance with the dedicated software 13 installed in the host computer 10.

(2) Processing when issuing a data write request from the host computer 10 Next, in the disk resource management table area 222b in the memory 222, the disk resource management table 12 shown in FIG.
With 0 stored, the disk array controller 2 in the disk array device 20 from the host computer 10
A process when a data write request is issued to the data 2 will be described with reference to the flowchart of FIG.

First, the disk array controller 22 (the microprocessor 221 therein) is connected to the host computer 10.
When a data write request is issued from (the file system 11 in) to the disk array device 20, the request is received (step S11). This request includes the address (starting address) and the size.

Next, the disk array controller 22
From the start address and size included in the received data write request, the stripe number indicating the stripe to be written is calculated (step S12).

Next, the disk array controller 22
Among the entries in the disk resource management table 120 stored in the disk resource management table area 222b, by referring to the entry specified by the stripe number calculated in step S12, the write target stripe indicated by the stripe number is a file. System 1
It is determined whether or not it is already used in 1 (step S13). If the stripe to be written has not been used until then, the disk array controller 22 indicates the content (flag state) of the entry in the disk resource management table 120 referred to in step S13 from unused to used. To (step S1
4). As described above, when it is determined by the data writing specified by the data writing request from the host computer 10 that the state of the stripe including the logical block specified by the request changes from unused to busy, Disk resource management table 120 corresponding to stripes
The contents of the entry in are updated to indicate busy.

After executing step S14, the disk array controller 22 writes the data specified by the data write request from the host computer 10 to the disk array 21 (step S15).

Further, the disk array controller 22 is
When the write target stripe indicated by the stripe number calculated in step S12 is already used in the file system 11 (step S13), step S1
4 is skipped and the process proceeds to step S15 to write the data specified by the data write request from the host computer 10.

(3) Data Restoration Process in Disk Array Device 20 Next, the data restoration process in the disk array device 20 will be described with reference to the flowchart of FIG.

Now, the HDD 210 in the disk array 21
-Since the HDD 210-3 out of 0 to 210-3 has failed,
Replace the failed HDD (old HDD) 210-3 with a new HD
Replace the D (new HDD) 210-3 with the old HDD 210-3
It is assumed that the data in the HDD is restored to the new HDD 210-3.
Here, for convenience, the new HDD, that is, the HD that is the restoration destination
D also has the same code "210" as the old HDD 210-3 that failed.
-3 "is added. The HDD that is the restoration destination may be a spare HDD that is connected in advance to the disk array controller 22.

When the HDD 210-3 fails, the disk array controller 22 (inside the microprocessor 221) stripes the processing for restoring the data in the failed HDD (old HDD) 210-3 to the new HDD 210-3. The following processing is executed in order from the first stripe with the number n of 0 (step S21).

First, the disk array controller 22
Stripe 23 with stripe number n (the initial value of n is 0)
To recover the data of the stripe number n, the entry in the disk resource management table 120 designated by the stripe number n is referred to, and the stripe 23 indicated by the stripe number n is referred to.
It is determined whether (at least one of the logical blocks included in) is used by the file system 11 (step S22).

If the stripe 23 indicated by the stripe number n is used, the disk array controller 22 performs the normal data restoration process according to the conventionally known RAID function as shown in FIG. Do as shown.

First, the disk array controller 22
Reads data from all normal HDDs 210-0 to 210-2 for the stripe 23 to be repaired 111
Is performed (step S23). Next, the disk array controller 22 uses the read data to acquire the data restored as the operation result by the RAID function, that is, the exclusive OR operation 112 (step S24). And the disk array controller 22
Is the new HD included in the stripe 23
The operation 113 of writing to the area in D210-3 is executed (step S25). As a result, the data in the old HDD 210-3 is restored in the new HDD 210-3.

On the other hand, if the stripe 23 indicated by the stripe number n is not used by the file system 11, the disk array controller 22 determines that valid data to be restored is not stored in the stripe 23. to decide. In this case, the disk array controller 22 performs the operation shown in FIG. That is, the disk array controller 22 writes the predetermined fixed data into each area in all the normal HDDs 210-0 to 210-2 included in the stripe 23 (operation 21)
1 is executed, and the operation 212 of writing the exclusive OR value of the fixed data to the area in the new HDD 210-3 included in the stripe 23 is executed (step S26).
Here, the exclusive OR value of the fixed data may be obtained by actually performing an exclusive OR operation or may be a predetermined fixed value. That is, in step S26, fixed data is written to the stripe 23 that is not used by the file system 11. The operation of this step S26 is performed by the normal HDD 21.
Stripe 2 used by file system 11 because it does not need to read from 0-0 to 210-2
When recovering data of No. 3 (steps S23 to S25)
It can be executed in a shorter time than In addition, HDD210-0
Since the reading from ~ 210-2 is not performed, H
The risk of multiple failures of DD is greatly reduced.

When the disk array controller 22 finishes step S25 or S26, the stripe number n
Is incremented by 1 (step S27), and it is determined whether or not the restoration processing has been completed up to the final stripe depending on whether or not the stripe number n after the increment exceeds the final stripe number (step S28). If it is determined in step S28 that the disk array controller 22 has not ended, the disk array controller 22 re-executes the operations in step S21 and subsequent steps.

In the above-described processing for recovering data in the failed HDD, it is assumed that fixed data is written to the stripe (step S26) that is not used by the file system 11. explained. However, file system 1
Since valid data to be restored is not stored in the stripe which is not used by No. 1, the operation of this step S26 (stripe restoration operation with fixed data) as indicated by the broken line 60 in the flowchart of FIG. You may skip.

However, if step S26 is skipped, the consistency between the data in RAID level 5 and the redundant data (parity data) cannot be obtained for the stripes not used by the file system 11. That is, the correct parity data is not generated for the data. Therefore, step S
When applying the method of skipping the operation of No. 26, it is necessary to process the data write request from the host computer 10 as shown in FIG. 13 according to the procedure shown in the flowchart of FIG.

First, the disk array controller 22
When the host computer 10 issues a data write request to the disk array device 20, the stripe number of the stripe 23 to which the data specified by the request is written is calculated, and the stripe 23 is used by the file system 11. It is determined based on the disk resource management table 120 (step S
31-S33). The process up to this point is the same as steps S11 to S13 in the flowchart of FIG.

If the stripe 23 to which data is to be written is not used by the file system 11, the disk array controller 22 determines that the stripe 23 does not have parity data consistency according to RAID level 5. To do. In this case, first, the disk array controller 22 sets the host computer 10
From the write data (new write data) 131 specified by the data write request from the HDD, the number of the HDDs constituting the disk array 21-2 HDs.
The exclusive OR value 135 with the fixed data 133 for D is acquired as correct parity data (redundant data) (step S34). Here, if the fixed data 133 is data in which all bits are “0”, the exclusive OR value 135
Matches the write data 131. In this case, the write data 131 is converted into the exclusive OR value (redundant data) 135.
Therefore, the exclusive OR operation is not required.

Next, the disk array controller 22
In each HDD 210-0 to 210-3 area included in the stripe 23 to which data is to be written, each HD
The write data 131, the fixed data 133, or the exclusive OR value (redundant data) 135 is written for each D (step S35). Here, the data write destination HDD designated by the data write request from the host computer 10 is the HDD 210-0, and the HDD storing redundant data in the stripe 23 is the HDD 210-3.
, The write data 131 is written 132 to the HDD 210-0, and the fixed data 133 is written 134 to the HDD 210-1 and 210-2.
And writing 136 of the exclusive OR value (redundant data) 135 to the HDD 210-3 are performed.
As a result, data redundancy can be assured even for stripes for which repair has been skipped when repairing data on a failed HDD.

When the disk array controller 22 finishes the step S35, the contents of the entry in the disk resource management table 120 referred to in the step S33, that is, the disk resource management table indicating whether or not the stripe 23 to which the data is written is used. The contents of the entry in 120 are updated to indicate unused to in use (step S36).

When the stripe 23 to which data is to be written is used by the file system 11, the disk array controller 22 operates as a normal R disk.
Data writing is performed by the AID method (step S3)
7). Here, the data (new data) specified by the data write request and the HDD that is the storage destination of the new data
An exclusive OR operation is performed between the data (old data) stored in the inner area and the parity data (old parity data) stored in another HDD inner area included in the same stripe 23. Thus, new parity data (new redundant data) is generated, and the old parity data is updated with the new parity data.

(4) H in the disk array device 20
DD Media Failure Inspection Processing Next, HDD media failure inspection processing in the disk array device 20 will be described with reference to the flowchart of FIG.

The disk array controller 22 is an HD
The D media inspection is periodically executed by the patrol function of the controller 22, for example. Here, the disk array controller 22 executes the HDD media inspection in the following order from the first stripe with the stripe number n of 0 (step S41).

First, the disk array controller 22
Stripe 23 with stripe number n (the initial value of n is 0)
Stripe number n for HDD media inspection of
By referring to the entry in the disk resource management table 120 designated by, it is determined whether or not the stripe 23 indicated by the stripe number n is used by the file system 11 (step S42).

If the stripe 23 indicated by the stripe number n is used, the disk array controller 22 makes all HDD
Data is read from 210-0 to 210-3 (step S43).

Next, the disk array controller 22
The result of reading the data from the HDDs 210-0 to 210-3 is checked to determine whether the reading has been successful (step S44).

If the data read from any of the HDDs 210-0 to 210-3 fails, the disk array controller 22 sets the same sector block as the failed sector block, that is, the defective sector block in which the media failure is detected. Substitution processing is performed to substitute another sector (replacement sector) in the HDD (step S45). For example, in FIG.
As shown in FIG. 4, sector block 1 in HDD 210-1
If 41 is detected as a bad sector block, a substitution process 143 is performed for substituting the sector block 141 with an arbitrary replacement sector 142 in the same HDD 210-1.

Next, the disk array controller 22
Using the RAID function, the bad sector block 141
Of the restored data in step S4 and write the data in the alternate sector 142 (step S4).
6). Then, the disk array controller 22 proceeds to step S47.

On the other hand, if the data read from the HDDs 210-0 to 210-3 for the stripe 23 indicated by the stripe number n is successful, the disk array controller 22 skips steps S45 and S46 and proceeds to step S47. .

If the stripe 23 indicated by the stripe number n is not used by the file system 11, the disk array controller 22 does not inspect the stripe 23 and skips steps S43 to S46 and goes to step S47. move on.

The disk array controller 22 increments the stripe number n by 1 in step S47, and until the incremented stripe number n exceeds the final stripe number (step S48).
The operation after step S41 is repeated.

In the HDD media failure inspection process described above, the case where the inspection (media inspection) by reading the data from the HDD is executed only for the stripes used by the file system 11 has been described.
It is not limited to this. For example, as shown in the flowchart of FIG. 15, all stripes in the disk area of the disk array device 20 may be inspected by reading data from the HDD regardless of whether they are used (steps S52 and S53). ). Here, when the sector block (defective sector block) in which the media failure has occurred is detected by the inspection by reading the data, the disk resource management table 120 is referred to and the stripe including the sector block is used by the file system 11. It is determined whether it has been done (step S54).

If the stripe is not used by the file system 11, it is not necessary to hold the data in the stripe. Therefore, in this case, the replacement process (1) of the defective sector block (141) as shown in FIG. 12 is performed regardless of the presence or absence of RAID redundancy.
After performing step 43) (step S55), a predetermined fixed data (not restoration data) is written (144) to the alternate sector (142) to restore (step S56). In this step S56, fixed data is written in the sector in another normal HDD corresponding to the alternate sector, as in the case of the stripe in step S26.

On the other hand, if the stripe is used by the file system 11, no failure occurs in another HDD (step S57), and therefore, if data redundancy is ensured, a defect occurs. After performing the process (143) of substituting the sector block (141) with the alternate sector (142) (step S58), the data of the defective sector block (141) is restored using the RAID function, and the restored data is restored. Writing (144) to the alternate sector (142) is performed (step S59).

On the other hand, when the data redundancy is lost because another HDD has failed (step S57), the data of the defective sector block (141) is set to R.
It cannot be repaired due to the function of AID. In this case, steps S58 and S59 are skipped and the defective sector block (141) is left as it is.

In the above embodiments, the disk array device 20 is described as being used at the RAID 5 level, but the present invention is not limited to RAID 3 level and other R disks.
Although the data recovery method is different, it can be applied to the disk array device used at the AID level as in the case of the RAID5 level.

The present invention is not limited to the above-described embodiment, but can be variously modified in an implementation stage without departing from the scope of the invention. Furthermore, the embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some constituent elements are deleted from all the constituent elements shown in the embodiment, the problem described in the section of the problem to be solved by the invention can be solved, and the effect described in the section of the effect of the invention can be solved. When at least one of the above is obtained, the configuration in which this constituent element is deleted can be extracted as the invention.

[0082]

As described above in detail, according to the present invention, of the disk areas of the disk array, the area actually used by the file system is determined and is actually used by this file system. Since the disk drive media failure inspection is performed focusing on the area, the processing time can be shortened, and as a result, the media failure can be detected early and the reliability of the disk array device can be improved. .

Further, according to the present invention, when a data write request is received from the host computer, the stripe to which the data write destination specified by the request belongs includes the first block including the logical block used by the file system.
Disk stripe management information (second disk resource management information) held in the disk array device, indicating whether the stripe is the second stripe or the second stripe that does not include the logical block used by the file system. If it is the second stripe, the second disk resource management information is updated so as to indicate that the stripe to which the data write destination belongs is the first stripe. The second disk resource management information can represent the latest usage status of the stripe, and it is not necessary to obtain the latest first disk resource management information from the host computer every time the file is updated.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a computer system including a disk array device 20 according to an embodiment of the present invention.

FIG. 2 is a diagram for explaining the relationship between stripes, logical blocks, and sector blocks used to manage the disk area of the disk array 21.

FIG. 3 is a diagram showing an example of a data structure of a disk resource management table 12 stored in a host computer 10 and a disk resource management table 120 stored in a disk resource management information area 222b in a memory 222 of a disk array controller 22.

FIG. 4 is a flowchart showing a processing procedure when transmitting a disk resource management information list.

FIG. 5 is a flowchart showing a processing procedure when a data write request is issued from the host computer 10.

FIG. 6 is a flowchart showing a data recovery processing procedure in the disk array device 20.

FIG. 7 is a flowchart showing a modification of the processing procedure when a data write request is issued from the host computer 10.

FIG. 8 is a flowchart showing a processing procedure of an HDD media failure inspection in the disk array device 20.

FIG. 9 shows a host computer 10 to a disk array device 20.
The figure which shows the example of a data structure of the disk resource management information list transmitted to FIG.

FIG. 10 is a diagram showing a modified example of the disk resource management information list.

FIG. 11 is a diagram for explaining a normal data restoration process in RAID level 5.

FIG. 12 is a diagram for explaining an operation at the time of writing fixed data to an unused stripe.

FIG. 13 is a diagram for explaining an operation at the time of writing new data to an unused stripe.

FIG. 14 is a diagram for explaining a data recovery operation when a defective sector block is detected in the HDD media failure inspection.

FIG. 15 is a flowchart showing a modification of the processing procedure of the HDD media failure inspection shown in FIG.

[Explanation of symbols]

10 ... Host computer 11 ... File system 12 ... Disk resource management table 13 ... Dedicated software 20 ... Disk array device 21 ... Disk array 22 ... Disk array controller 23 ... Stripe 24 ... Logical block 25 ... Sector block 90, 100 ... Disk resource Management information list (first disk resource management information) 120 ... Disk resource management table (second disk resource management information) 210-0 to 210-3 ... HDD (disk drive) 222b ... Disk resource management table area

Claims (4)

[Claims] 1. A RAID (Redundant Arrays of Inexpensive Disks) composed of a plurality of disk drives.
A method for recovering data in a disk array device having a disk array having a configuration, wherein a logical block used by the file system from a host computer having a file system for managing the disk area of the disk array in logical block units Or obtaining a first disk resource management information indicating a stripe including the logical block, and using the file system for each stripe in the disk area of the disk array from the first disk resource management information. Generating a second disk resource management information indicating whether or not it contains a logical block and holding it in the disk array device; and when a data write request is received from the host computer, it is specified by the request. The data write destination belongs to Specifying a slice, and the specified stripe is a first stripe including a logical block used by the file system or a second stripe not including a logical block used by the file system Determining based on the second disk resource management information held in the disk array device, and if the identified stripe is the second stripe, Updating the second disk resource management information held in the disk array device to indicate that the specified stripe is the first stripe; and each disk in the disk array. By reading the contents stored in the drive, the part of the disk drive A case to perform media inspection processing of detecting a failure, for all the stripes of the disk areas of the disk array, or a first stripe comprising a logic block used by the file system,
Alternatively, a step of determining whether the second stripe does not include a logical block used by the file system based on the second disk resource management information, and based on the determination result of the determination step, , Performing a media inspection including data reading from the disk drive only for the first stripe, and R for data of a location where a failure is detected in the media inspection.
And a step of recovering by the AID function. 2. A RAID (Redundant Arrays of Inexpensive Disks) composed of a plurality of disk drives.
A method for recovering data in a disk array device having a disk array having a configuration, wherein a logical block used by the file system from a host computer having a file system for managing the disk area of the disk array in logical block units Or obtaining a first disk resource management information indicating a stripe including the logical block, and using the file system for each stripe in the disk area of the disk array from the first disk resource management information. Generating a second disk resource management information indicating whether or not it contains a logical block and holding it in the disk array device; and when a data write request is received from the host computer, it is specified by the request. The data write destination belongs to Specifying a slice, and the specified stripe is a first stripe including a logical block used by the file system or a second stripe not including a logical block used by the file system Determining based on the second disk resource management information held in the disk array device, and if the identified stripe is the second stripe, Updating the second disk resource management information held in the disk array device to indicate that the specified stripe is the first stripe; and each disk in the disk array. By reading the contents stored in the drive, the part of the disk drive A step of sequentially performing a media inspection process for detecting various faults for all stripes in the disk area of the disk array, and when a partial fault of the disk drive is detected in the media inspection process, The stripe containing the location where the failure was detected is the first stripe containing the logical block used by the file system,
Alternatively, a step of determining whether the second stripe does not include a logical block used by the file system, based on the second disk resource management information held in the disk array device, If the disk drive other than the disk drive in which the failure is detected is normal among the plurality of disk drives in the disk array determined to be the first stripe in the determination step, the failure is detected. And a step of recovering the data of the location where the failure is detected by the fixed data when the determination step determines that the stripe is the second stripe. A method for recovering data in a disk array device, comprising: 3. A RAID (Redundant Arrays of Inexpensive Disks) composed of a plurality of disk drives.
In a disk array controller for controlling a disk array having a configuration, a logical block used by the file system or the logical block transmitted from a host computer having a file system that manages the disk area of the disk array in logical block units. From the first disk resource management information indicating the stripe including the block, the second disk resource management indicating whether or not each stripe in the disk area of the disk array includes the logical block used by the file system. A means for generating information, a memory for storing the second disk resource management information, and a means for specifying a stripe to which a data write destination designated by the request belongs when a data write request is received from the host computer And the identifying means When the stripe to which the data write destination belongs is specified by, the stripe is the first stripe including the logical block used by the file system, or the logical block used by the file system. Second not including
If the first determination means determines that the stripe identified by the identification means is the second stripe, the stripe is the first stripe. The second stored in the memory to indicate that it is a stripe of
And a first stripe including a logical block used by the file system for all stripes of the disk area of the disk array, or used by the file system. The second stripe that does not include the logical block
Second judging means for judging based on the disk resource management information, and media checking means for detecting a partial failure of the disk drive by reading the storage contents of each of the disk drives in the disk array. A media inspection unit that executes a media inspection including data reading from the disk drive only for the stripe determined to be the first stripe by the second determination unit; and a failure by the media inspection unit. A disk array controller, comprising: a data recovery unit that recovers data at a location where is detected by a RAID function. 4. A RAID (Redundant Arrays of Inexpensive Disks) composed of a plurality of disk drives.
In a disk array controller for controlling a disk array having a configuration, a logical block used by the file system or the logical block transmitted from a host computer having a file system that manages the disk area of the disk array in logical block units. From the first disk resource management information indicating the stripe including the block, the second disk resource management indicating whether or not each stripe in the disk area of the disk array includes the logical block used by the file system. A means for generating information, a memory for storing the second disk resource management information, and a means for specifying a stripe to which a data write destination designated by the request belongs when a data write request is received from the host computer And the identifying means When the stripe to which the data write destination belongs is specified by, the stripe is the first stripe including the logical block used by the file system, or the logical block used by the file system. Second not including
If the first determination means determines that the stripe identified by the identification means is the second stripe, the stripe is the first stripe. The second stored in the memory to indicate that it is a stripe of
Means for updating the disk resource management information, and a media inspection for executing, for each stripe, a media inspection process for detecting a partial failure of the disk drive by reading the storage contents of each disk drive in the disk array. Means, and when the media inspection means detects a partial failure of the disk drive, the stripe including the location where the failure is detected is the first stripe including the logical block used by the file system. Yes, or a second stripe that does not contain a logical block used by the file system,
A second determining unit that determines based on the second disk resource management information; and a plurality of disk drives that are determined to be the first stripe by the second determining unit and that are in the disk array. When a disk drive other than the disk drive in which the failure is detected is normal, first data recovery means for recovering data at the location where the failure is detected by a RAID function, and the second determination means A disk array controller comprising: a second data repairing unit that repairs the data at the location where the failure is detected with fixed data when it is determined to be the second stripe.