JP2003167688A - Disk array device and consistency recovering method for logical drive having redundant data - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a disk array device and a consistency recovering method for a logical dive having redundant data, which are expected to perform smooth processing after restart by surely avoiding a medium error accompanied with processing interruption in the middle of write with respect to a disk array where the fault tolerance of disks is realized by redundant data. <P>SOLUTION: At restart after write processing interruption, a RAID controller 11 writes a predetermined value in an area (X, Y) registered in a non-volatile memory 12 and generates a corresponding parity in accordance with each data stripe in the same parity group including this data area and writes the parity in a corresponding disk of a disk array 14. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array device which realizes disk fault tolerance by redundant data, and a method for recovering consistency of a logical drive having redundant data.

[0002]

2. Description of the Related Art In a disk array system composed of a plurality of small disks, RA is used as a technology for realizing disk fault tolerance by having redundant data.
ID technology is known.

In a RAID system to which this RAID technology is applied, when the RAID controller is stopped due to a power failure or the like during writing to a disk, the consistency between the data on the disk device and the redundant data is lost. There is. For example, in the case of a logical drive that realizes fault tolerance by providing redundant data with RAID level 3 or RAID level 5, etc., if processing stops while data and parity (redundant data) are being written to the disk, some data will be lost. However, the parity is not updated even though the update of the above is completed, and the data and the parity may be inconsistent in the parity group of the disk array.
If the system is restarted in this state and processing continues, if a disk drive belonging to a logical drive fails and the logical drive is in a degraded state, incorrect parity is used on the failed disk drive. Since the data is restored, the wrong data will be restored.

In a RAID controller of a certain type, an area currently being updated is registered in the nonvolatile memory when the writing process is executed in the logical drive, and when the writing process is interrupted, the area is updated at the next restart. After performing the stripe group consistency recovery process on the former area, the normal operation is resumed. As a consistency recovery processing technique associated with the interruption of the writing process at this time, Japanese Patent No. 307101
No. 7 is known. Here, when the writing process is interrupted, the redundant data is reconstructed from the data of the data disk corresponding to the data area being updated at the time of the next restart.

In this prior art, when the writing process is interrupted, the operation ends while the area is still in the middle of writing, so there is a risk of a media error such as a sector defect at the time of restart. That is, at the time of restarting after the above-mentioned write processing is interrupted, data including the area that is still in the state of being written is written, and then when a read request is made, the consistency of data and parity is still lost at this point. Since it remains in the opened state, the I / O access becomes invalid (error) and data cannot be read or written. In this case, once the above-mentioned consistency recovery processing is performed, the redundant data is reconstructed from the data on the data disk corresponding to the data area that was being updated, the consistency recovery processing is performed, and then the processing according to the request. Must be performed.

Therefore, at the time of restarting after the interruption of the writing process, when there is a writing request including an area which is still in the state of being written, prior to the writing process, the original data and the redundant data are temporarily changed. There is a wasteful process of newly generating redundant data and performing a consistency recovery process.

[0007]

As described above, in the prior art, when the writing process is interrupted, the operation ends while the area is still in the middle of writing, so that when the writing process is restarted, the operation is terminated. There is a risk that a media error will occur, and if it is attempted to avoid this, useless processing to reconstruct redundant data from the data of the data disk corresponding to the data area that was being updated prior to access request processing There was a problem of being involved.

The present invention has been made in view of the above circumstances,
In a disk array that achieves disk fault tolerance with redundant data, a disk array device and redundant data that can reliably avoid media errors due to processing interruption during writing and expect smooth processing after restarting An object of the present invention is to provide a consistency recovery method for a logical drive having

[0009]

SUMMARY OF THE INVENTION The present invention is a disk array in which disk fault tolerance is realized by redundant data, and when the writing process is interrupted during the writing process, the writing process is restarted during the writing process. It is possible to avoid access failure such as media error after reboot by performing forced write (overwrite processing) for consistency recovery using the default value or existing data in the area that remains in the state. And

The present invention is also characterized in that the consistency of redundant data is recovered by a method different from the conventional consistency recovery processing described above.

That is, the present invention provides a disk array device which realizes disk fault tolerance by redundant data.
Non-volatile storage means for storing at least a logical address and a size when a write request is made until the processing according to the request is completed, and the logic stored in the storage means when the write processing according to the write request is interrupted. A means for writing a default value in the area of the data stripe indicated by the address and size, reconstructing redundant data from the default value and the data of the other corresponding data stripe, and writing the redundant data in the corresponding parity stripe is provided. It is characterized by having done.

Also, the present invention provides a disk array device which realizes disk fault tolerance by redundant data,
Non-volatile storage means for storing at least a logical address and a size when a write request is made until the processing according to the request is completed, and the logic stored in the storage means when the write processing according to the write request is interrupted. A default value is written in the redundant data area of the parity stripe corresponding to the data area of the data stripe indicated by the address and size, and the default value and the data of other data stripes corresponding to it are used to write the default value to the area Means for reconstructing data and writing the data in the data area of the data stripe.

Further, according to the present invention, in the disk array device which realizes the disk fault tolerance by the redundant data provided for each stripe group, the process of complying with at least the logical address and the size at the time of the write request is completed. And a parity stripe redundancy corresponding to the data area of the data stripe indicated by the logical address and size stored in the storage means when the write processing according to the write request is interrupted. It is characterized by further comprising means for restoring data in an area in which the writing process is interrupted from data and data in another data stripe corresponding thereto and writing the data in the data area of the data stripe.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a disk array system for explaining an embodiment of the present invention.

Here, R provided in the host device 10
An example is shown in which a disk array 14 including a plurality of hard disk drives (HDD0, HDD1, HDD2) is connected to the AID controller 11 via a SCSI bus at RAID level 5. A non-volatile memory 12 and a cache 13 are provided in the RAID controller 11.

In the non-volatile memory 12, the logical address range (address and size (X, Y)) requested to be written, which is referred to in the consistency recovery process described later, is registered until the writing process is completed. In order to realize high-speed access to the cache 13, similar to the existing system,
The data read from the disk array 14 is temporarily stored.

FIG. 2 is an operation explanatory diagram of the write processing executed in the RAID controller 11. Here, the write processing is executed in an area (offset X, size Y) where the logical drive is located. Is shown. At the time of this writing process, the area (X, Y) indicating the access range is registered in the nonvolatile memory 12 and held until the writing is completed. Using the area (X, Y) registered in the non-volatile memory 12, a consistency recovery process, which will be described later, is executed at the time of restart after the interruption of the write process.

FIGS. 3 to 9 are operation explanatory views and flowcharts for explaining the consistency recovery processing operation in each embodiment of the present invention, and here, for the sake of simplicity of explanation, one operation is explained. Parity group (PG)
Is composed of two data stripes (DS) and one parity stripe (PS). In each embodiment, the data stripe (DS) and the parity stripe (PS) are shown in the same arrangement.

Of these, FIGS. 3 and 4 are for explaining the consistency recovery processing operation according to the first embodiment of the present invention. FIG. 3 is an operation explanatory view, and FIG. 4 is a flowchart showing a processing procedure. Is. Here, when restarting after the interruption of the writing process, a default value (for example, 0) is written in the area (X, Y) registered in the nonvolatile memory 12,
Consistency recovery is performed by generating and overwriting the corresponding parity from the data stripe including the data area.

FIG. 5 and FIG. 6 respectively show the second aspect of the present invention.
FIG. 5 is an operation explanatory diagram and FIG. 6 is a flowchart showing a processing procedure for explaining the consistency recovery processing operation according to the embodiment. Here, when restarting after the interruption of the writing process, a default value (for example, 0) is written in the parity area corresponding to the area (X, Y) registered in the non-volatile memory 12, and this parity and the area (X , Y) using the same stripe data other than the above area (X,
Consistency recovery is performed by reconstructing the data in Y).

FIG. 7 and FIG. 8 respectively show a third embodiment of the present invention.
FIG. 7 is an operation explanatory diagram and FIG. 8 is a flowchart showing a processing procedure for explaining the consistency recovery processing operation according to the embodiment. Here, at the time of restarting after the interruption of the write processing, the consistency recovery is performed by overwriting the area (X, Y) registered in the nonvolatile memory 12 with the data generated from the other stripe of the corresponding parity group. Is going.

FIG. 9 is a diagram for explaining the operation of the consistency recovery process according to the fourth embodiment of the present invention. Here, at the time of restarting after the interruption of the write process, the area (X , Y) straddles two data stripes, the data in the areas (X, Y) registered in the non-volatile memory 12 is separately generated for each parity group, and the consistency is restored by overwriting the data. It is carried out.

Now, the operation of the embodiment of the present invention will be described with reference to the drawings.

Under the control of the host device 10, the processing procedure when the RAID controller 11 executes the writing process in the area of the logical drive (offset X, size Y) is shown below.

A. The write-requested logical address and size (X, Y) are registered in the nonvolatile memory 12.

B. A parity (P) corresponding to the data to be updated is generated.

C. The data (D) and the parity (P) are written in the disk drive of the disk array 14.

D. When the writing process is completed, the registration (X, Y) on the non-volatile memory 12 is deleted.

In the RAID 5 logical drive, since it is necessary to update the data and the corresponding parity data, the areas to be written are the data area and the parity area.

Here, it is assumed that the RAID controller 11 loses power supply and the processing is stopped before the writing of data or parity is completed.

In this state, the redundancy data becomes inconsistent with respect to the parity group for which writing has been interrupted. Therefore, when the power supply to the RAID controller 11 is restarted next, the consistency of the corresponding parity group is resumed. Restore sex.

In other words, according to the embodiment of the present invention, the nonvolatile memory 1 is used when restarting after the interruption of the writing process.
By using the logical address / size (X, Y) registered in No. 2, the consistency recovery process accompanied by the write (overwrite) process is executed, and the write (overwrite) process at this time causes the subsequent access, It reliably eliminates media errors, and avoids useless temporary redundancy data consistency recovery processing that is performed in existing systems when writing data after a reboot.

The consistency recovery processing operation in each embodiment of the present invention will be described below with reference to FIGS. 3 to 9.

(First Embodiment) First, FIG. 3 and FIG.
The consistency recovery processing operation in the first embodiment of the present invention will be described with reference to FIG.

At the time of system startup, the non-volatile memory 12 is accessed and the disk I / O for which writing has been interrupted
It is determined whether there is any (step S11 in FIG. 4). At the time of restarting after the interruption of the writing process, since the area (X, Y) at the time of the interruption of the writing process is registered in the non-volatile memory 12, it is possible to determine that the writing has been interrupted (FIG. 4 step). S11 YES).

At this time, a default value (for example, 0) is written in the data area of the data stripe indicated by the area (X, Y) registered in the nonvolatile memory 12 (step S12a in FIG. 4). And the default value (0) of the data area
And the data of each data stripe (DS) in the same parity group (PG) corresponding to the data area are read to generate the corresponding parity (P). The generated parity (P) is set to the parity stripe (P
S) is written in the corresponding parity area (step S12b in FIG. 4). The value of the area (X, Y) registered in the non-volatile memory 12 is erased upon completion of this reconfiguration writing process (step S13 in FIG. 4). This ends the consistency recovery process.

(Second Embodiment) Next, FIG. 5 and FIG.
The consistency recovery processing operation in the second embodiment of the present invention will be described with reference to FIG.

At the time of restart after the interruption of the writing process (step S21 YES in FIG. 6), the parity corresponding to the data of the data stripe (D) indicated by the area (X, Y) registered in the nonvolatile memory 12 is written. A default value (for example, 0) is written in the parity area of the stripe (step S22a in FIG. 6). The default value (0) of this parity area and the same parity group (P
The data of the other data stripe (DS) in G) is read, the data of the data stripe (D) is reconstructed, and the data is written in the data area (X, Y) (step S22b in FIG. 6). Upon completion of this reconfiguration writing process, the value of the area (X, Y) registered in the non-volatile memory 12 is deleted (step S23 in FIG. 6). This ends the consistency recovery process.

(Third Embodiment) Next, FIG. 7 and FIG.
The consistency recovery processing operation in the third embodiment of the present invention will be described with reference to FIG.

At the time of restarting after the interruption of the writing process (step S31 YES in FIG. 8), the same parity group corresponding to the data of the data storage pipe indicated by the area (X, Y) registered in the nonvolatile memory 12 above. The parity is read from the parity area of the (PG) parity stripe and the data is read from the data area of another data stripe to restore the data. The restored data is written into the data area of the data stripe indicated by the area (X, Y) and the restoration processing is completed (step S3 in FIG. 8).
2). The value of the area (X, Y) registered in the non-volatile memory 12 is erased upon completion of this reconfiguration writing process (step S33 in FIG. 8). This ends the consistency recovery process.

(Fourth Embodiment) Next, the consistency recovery processing operation in the fourth embodiment of the present invention will be described with reference to FIG.

The data reconstruction shown in FIG. 9 shows an example of the consistency recovery processing when the area registered in the non-volatile memory 12 extends over a plurality of data stripes (DS). The area (X,
When Y) is a range that spans a plurality of data stripes (disk drives), the write access range is divided at stripe boundaries, and in the same method as in the third embodiment, in step 1, the first data stripe The first data (D) is restored from the parity and data corresponding to (DS), and written in the area of one drive designated by the area (X, Y). Then in step 2,
The data (D) is restored from the parity and the data corresponding to the second data stripe (DS), and the area (X,
Write to the area of another drive specified in Y). In this way, the area (X, Y) that straddles the disk drive
Restore each write data of. In addition, although the restoration processing of the third embodiment is applied in the fourth embodiment, the first and second embodiments described above can be applied similarly.

As described above, according to the embodiments of the present invention, at least the logical address / size (X, Y) registered in the non-volatile memory 12 is used at the time of restart after the interruption of the writing process. , Consistency recovery processing with write (overwrite) processing is executed, and by this write (overwrite) processing, media errors are reliably eliminated during subsequent access, and write processing after restart is performed. In this case, the wasteful recovery process of the temporary redundant data in the existing system is avoided. As a result, it is possible to reliably avoid media errors due to processing interruption during writing and expect smooth processing after restarting.

[0045]

As described above in detail, according to the present invention, it is possible to surely avoid a media error due to interruption of processing during writing and to expect smooth processing after restarting.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a disk array system for explaining an embodiment of the present invention.

FIG. 2 is an operation explanatory diagram illustrating a write processing operation of the RAID controller in the above embodiment.

FIG. 3 is an operation explanatory diagram of the consistency recovery processing according to the first embodiment of the present invention.

FIG. 4 is a flowchart showing a procedure of consistency recovery processing according to the first embodiment of the present invention.

FIG. 5 is an operation explanatory diagram of the consistency recovery processing according to the second embodiment of the present invention.

FIG. 6 is a flowchart showing a procedure of consistency recovery processing according to the second embodiment of the present invention.

FIG. 7 is an operation explanatory diagram of consistency recovery processing according to the third embodiment of the present invention.

FIG. 8 is a flowchart showing a procedure of consistency recovery processing according to the third embodiment of the present invention.

FIG. 9 is an operation explanatory diagram of the consistency recovery processing according to the fourth embodiment of the present invention.

[Explanation of symbols]

10 ... Host device 11 ... RAID controller 12 ... Nonvolatile memory 13 ... Cache (cache memory) 14 ... Disk array PG ... parity group DS ... Data stripe PS: Parity stripe

Claims (8)

[Claims] 1. A non-volatile storage means for storing at least a logical address and a size when a write request is made until the processing according to the request is completed in a disk array device which realizes disk fault tolerance by redundant data. And when the write process according to the write request is interrupted,
A default value is written in the area of the data stripe indicated by the logical address and the size stored in the storage means, and redundant data is reconfigured from the default value and the data of another data stripe corresponding to the default value, and the corresponding And a means for writing to the parity stripe. 2. A non-volatile storage means for storing at least a logical address and a size when a write request is made until the processing according to the request is completed in a disk array device which realizes disk fault tolerance by redundant data. And when the write process according to the write request is interrupted,
A default value is written in the redundant data area of the parity stripe corresponding to the data area of the data stripe indicated by the logical address and size stored in the storage means, and the default value and the data of other data stripes corresponding thereto are written. And a means for reconstructing data in an area in which the write processing is interrupted and writing the data in the data area of the data stripe. 3. In a disk array device that realizes disk fault tolerance by redundant data for each stripe group, at least a logical address and size when a write request is made are stored until the processing according to the request is completed. Non-volatile storage means, when the writing process according to the write request is interrupted,
Data in the area in which the writing process is interrupted from redundant data of the parity stripe corresponding to the data area of the data stripe indicated by the logical address and size stored in the storage means and data of other corresponding data stripes And a unit that restores the data and writes the data in the data area of the data stripe. 4. When the data area of the data stripe indicated by the logical address and size extends over a plurality of data stripe areas, the data or the redundant data is reconstructed separately for each data stripe. 4. The disk array device according to claim 1, 2 or 3. 5. A consistency recovery method for a logical drive comprising a plurality of physical drives and having a fault-tolerant function by having redundant data for each stripe group, wherein at least a logical address at the time of a write request and a logical address The size is stored in the nonvolatile storage means until the process according to the request is completed, and when the write process according to the write request is interrupted,
A default value is written in the area of the data stripe indicated by the logical address and size stored in the storage means, and redundant data is reconfigured from the default value and the data of other data stripes corresponding to the default value, and the corresponding Consistency recovery method of logical drive characterized by writing to parity stripe. 6. A consistency recovery method for a logical drive comprising a plurality of physical drives and realizing a fault tolerance function by having redundant data for each stripe group, and at least a logical address when a write request is made and The size is stored in the nonvolatile storage means until the process according to the request is completed, and when the write process according to the write request is interrupted,
A default value is written in the redundant data area of the parity stripe corresponding to the data area of the data stripe indicated by the logical address and size stored in the storage means, and the default value and the data of other data stripes corresponding thereto are written. A method of recovering a consistency of a logical drive having redundant data, comprising reconstructing data in an area in which the write processing is interrupted and writing the data in the data area of the data stripe. 7. A consistency recovery method of a logical drive comprising a plurality of physical drives and having a fault tolerance function by having redundant data for each stripe group, wherein at least a logical address and a logical address when a write request is made are provided. The size is stored in the nonvolatile storage means until the process according to the request is completed, and when the write process according to the write request is interrupted,
Data in the area in which the writing process is interrupted from redundant data of the parity stripe corresponding to the data area of the data stripe indicated by the logical address and size stored in the storage means and data of other corresponding data stripes Is restored and written in the data area of the data stripe. 8. When the data area of the data stripe indicated by the logical address and the size extends over a plurality of data stripe areas, the data or the redundant data is reconstructed separately for each data stripe. 8. The consistency recovery method according to claim 5, 6 or 7.