JPH09146717A - Information storage device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it unnecessary to load a new free storage device even when a certain storage device is destructed, to maintain the redundancy of other storage devices in a disk array RAID group and to hold the reliability of the information storage device. SOLUTION: Plural RAID groups are constituted of hard disks 7 to 10. When one of hard disks constituting a certain RAID group is destructed, remaining hard disks in the RAID group to which the destructed hard disk belongs are integrated to another RAID group by the use of a free storage device information transmitting/receiving, part 3, a stored data managing part 4, a stored data RAID processing part 5, and a storage device trouble detecting part 6, so that a hard disk with free storage capacity is secured and destructed data are restored to the secured disk.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information storage device for storing information by connecting a plurality of information storage media such as a disk array (RAID: Redundant Array of Disks) device in parallel.

[0002]

2. Description of the Related Art Recently, as information storage devices, for example, information storage media (hereinafter referred to as disks) such as hard disks and optical disks are connected in parallel, and all of them are controlled in synchronization as one disk storage device. Therefore, a disk array (RAID) device is often used for the purpose of speeding up data access (reading and writing) and improving durability against failures (U).
SP52333618).

In such an information storage device, information is distributed and stored in a plurality of disks, and a parity disk is used to realize high reliability of data management. For example, if one of a plurality of disks necessary for constructing one disk array cannot read data due to a disk failure, etc. In addition, it is possible to reconstruct the disk array by restoring the information stored in the disk whose data cannot be read.

Conventionally, a failure occurs in a storage device under the control of a storage system in which the reliability of the storage data is improved by making the storage data redundant among a plurality of media as described above. When restoration of the data is required, the same storage device as the defective device is replaced with a new one, and the storage data is restored to the new storage device to restore the original redundant and reliable storage data.

That is, when a failure occurs in a storage device under the control of a storage system having redundancy among a plurality of media and when this stored data is restored, the same storage device as the failure device is replaced with a new one. Then
I was restoring the stored data to a new storage device.

Therefore, if there is no new same storage device, even if the lost storage data can be restored, it cannot be restored to the redundant storage form, and the reliability of the storage data cannot be restored.

FIG. 13 shows an example of a general information storage device which has a plurality of storage devices, stores them as redundant data in a distributed manner, and responds to a data access request from the outside.

This information storage device includes a plurality of hard disks 97 to 100, a data transmission / reception section 92, and storage data RA.
It is composed of an ID processing unit 93, a storage device failure detection unit 94, and a storage data management unit 94, which are connected by a common bus 90.

The external system 91 includes a data transmitter / receiver 92.
The read data and the write data are transferred to the hard disks 97 to 100 via the. Hard disk 97-
The data stored in 100 is the stored data management unit 9
4. The storage data management unit 94
It has a management table 96 for data storage locations.

FIG. 14 shows an example of a management table 96 relating to data storage locations. Data storage locations are represented by record numbers (RN), hard disk numbers (HDN), and hard disk block numbers (BN).

For example, as shown in FIG. 15, it can be seen that the data of the record number RN2 is at the hard disk number HDN = 1 and at the hard disk block number BN = 2.

An access request comes from the external system 91 with a record number (RN) corresponding to the data. Also,
In this information storage device, data has redundancy and is distributed and stored as a RAID group in a plurality of storage devices.

FIG. 16 shows an example of the structure of a RAID group. RAID groups RG0, RG1, RG2
It is composed of groups such as…. That is, RAID
One hard disk is provided as a parity device (PD) for each group, and the data of all the remaining hard disks of the RAID group is set as E for each block.
The XORed value is written. By this, R
Even if one hard disk in the AID group malfunctions and the stored data is destroyed, the damaged data can be recovered from the stored data of the remaining hard disks constituting the RAID group.

FIG. 17 shows a management table 96 of the constituent hard disk numbers of each RAID group and the parity device numbers therein. The management table 96 can be used to identify which hard disk constitutes each RAID group and which of them is a parity device. Such a series of RAID-related processing is performed by the storage data RAID processing unit 93.

The storage device failure detection unit 95 monitors whether or not a failure has occurred in the hard disk, which is a storage device, and the stored data has been destroyed. If the stored data is destroyed, this is stored. Data RAID
Notify the processing unit 93.

In response to this, the storage data RAID processing unit 93 performs a restoration process for the destroyed storage data based on the RAID group management table 96 in the storage data management unit 94.

FIG. 18 shows the restoration of the stored data. The defective hard disk is discarded as shown in FIG. 18 (a), and a new empty hard disk is replaced as shown in FIG. 18 (b). Instead, add and restore the stored data that is destroyed.

As described above, in the conventional system, when the data in the storage device is destroyed, a new empty storage device exists in the system, and the data cannot be completely recovered unless it is mounted. Moreover, since the other storage devices in the RAID group to which the defective storage device belongs loses the redundancy of the storage data, the reliability is extremely deteriorated.

Further, in the conventional information storage device, the number of storage devices in the information storage device increases in the process of gradually increasing the storage data, and accordingly, the number of storage devices constituting the RAID group increases and the storage data increases. May reduce the reliability of. Further, when a RAID group is newly established without considering the data arrangement in the storage device, a large amount of data is moved between the RAID groups during the data optimum arrangement process, which may result in deterioration of performance.

[0020]

As described above, when the data in the storage device is destroyed, a new empty storage device exists in the system, and if it is not installed, the data can be completely restored. Since the other storage devices in the RAID group to which the storage device which cannot be performed and the faulty storage device belongs lose the redundancy of the storage data, the reliability is extremely deteriorated, and the storage data gradually increases. Then, the number of storage devices in the information storage device increased,
Along with this, the number of storage devices that make up the RAID group may increase and the reliability of the stored data may decrease. By creating a new RAID group without considering the data arrangement in the storage device, There has been a problem in that a large amount of data is moved between RAID groups during the optimal placement process, leading to a drop in performance.

Therefore, according to the present invention, when a storage device is destroyed, it is not necessary to mount a new empty storage device, and the redundancy of other storage devices in the RAID group is maintained to reduce reliability. It is an object of the present invention to provide an information storage device that does not have the above.

In the process of increasing the stored data, R
Prevents deterioration of reliability in the AID group and
It is an object of the present invention to provide an information storage device capable of reducing the occurrence of data movement between RAID groups due to the establishment of a new group and preventing performance deterioration.

[0023]

An information storage device of the present invention is an information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group. In the above, in the group and a plurality of information storage media forming the group, storage information management means for redundantly storing and managing information, detection means for detecting the occurrence of a defect in the information storage medium, and this detection means When detecting the occurrence of a defect in the information storage medium in, in the case of investigating means for investigating whether there is an empty information storage medium in which information is not stored, and if there is no empty information storage medium in this investigating means, A determination means for selecting and determining a restoration method for the stored information of the information storage medium detected by the detection means, and a restoration process using the stored information management means according to the restoration method determined by the determination means It is composed of a power sale processing means.

The information storage device of the present invention is an information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group. Detecting means for detecting the occurrence of the problem of, and a detecting means for investigating whether or not there is an empty information storage medium in which information is not stored when detecting the occurrence of the problem of the information storage medium by this detecting means, When there is no empty information storage medium in this investigation means, depending on the determination means for selectively determining the restoration method of the stored information of the information storage medium detected by the detection means, and the restoration method determined by this determination means , And a processing means for performing a process for recovering the information stored in the information storage medium in which the above-mentioned trouble has occurred and a writing process for redundantly distributing the stored information.

The information storage device of the present invention is an information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group. Detecting means for detecting the occurrence of the problem of, and a detecting means for investigating whether or not there is an empty information storage medium in which information is not stored when detecting the occurrence of the problem of the information storage medium by this detecting means, If there is no empty information storage medium in this investigation means, it is composed of an integration means for integrating the information storage medium belonging to the group of the information storage medium in which the defect has been detected by the above detection means into another group. .

The information storage device of the present invention is an information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group. Detecting means for detecting the occurrence of the problem of, and a detecting means for investigating whether or not there is an empty information storage medium in which information is not stored when detecting the occurrence of the problem of the information storage medium by this detecting means, If there is no empty information storage medium in this investigation means, the empty areas of the plurality of information storage media corresponding to the amount of stored information stored in the information storage medium in which the defect detected by the detection means has occurred are investigated. And the storage information stored in the information storage medium in which the defect detected by the detection unit has occurred is restored and stored in the empty area secured by this securing means. It is composed of a 憶 means.

The information storage device of the present invention is an information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group. Detecting means for detecting the occurrence of the problem of, and a detecting means for investigating whether or not there is an empty information storage medium in which information is not stored when detecting the occurrence of the problem of the information storage medium by this detecting means, If there is no empty information storage medium in this investigation means, the empty areas of the plurality of information storage media corresponding to the amount of stored information stored in the information storage medium in which the defect detected by the detection means has occurred are investigated. Whether a new information storage medium is required from the reliability of the group that is redundantly distributed and stored when the free space is not secured by this securing means and the securing means. Determining means for determining, and a notifying means for notifying to the outside when it is determined to require a new information storage medium in this determination means.

In the information storage device of the present invention, a plurality of portable information storage media are loaded and unloaded to read and write information, and a group is composed of the plurality of portable information storage media. Information is redundant for each group. In an information storage device that stores data in a distributed manner, a management unit that manages an access frequency for reading and writing each portable information storage medium that forms the group, and a portable information storage that forms a group managed by the management unit Depending on the access frequency of each medium,
It comprises optimum placement means for optimally placing information stored in a plurality of portable information storage media forming a group.

The information storage device of the present invention is an information storage device in which a group is composed of a plurality of portable information recording media, and the information stored in the portable information storage media forming this group is redundantly distributed and stored. Determining means for determining division of the group when the number of portable information storage media forming the group exceeds a predetermined number, and processing for dividing the group when the determination means determines division of the group And means.

In the information storage device of the present invention, a plurality of portable information storage media are loaded / unloaded to read / write information, a group is composed of the plurality of portable information storage media, and information is redundant for each group. In an information storage device that stores data in a distributed manner, a management unit that manages an access frequency for reading and writing each portable information storage medium that constitutes the group, and a portable information storage medium that constitutes the group are predetermined. Based on the access means for each portable information storage medium constituting the group managed by the management means when the group division is judged by this judgment means And a processing means for dividing the group into groups.

In the information storage device of the present invention, a plurality of portable information storage media are loaded / unloaded to read / write information, a group is composed of the plurality of portable information storage media, and information is redundant for each group. In an information storage device that stores data in a distributed manner, a management unit that manages an access frequency for reading and writing each portable information storage medium that constitutes the group, and a portable information storage medium that constitutes the group are predetermined. Based on the access means for each portable information storage medium constituting the group managed by the management means when the group division is judged by this judgment means Processing means for dividing the group into groups and the information stored in the plurality of portable information storage media forming the groups divided by the processing means. And a best placement means for optimum placing.

[0032]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a first embodiment of the present invention.
1 illustrates a schematic configuration of an information storage device according to an embodiment.

That is, the information storage device includes a data transmission / reception unit 2, an empty storage device information transmission / reception unit 3, a storage data management unit 4, a storage data RAID processing unit 5, a storage device failure detection unit 6, and a plurality of hard disks 7-10. And are connected by a common bus 15.

The external system 1 uses the data transmitting / receiving section 2 to read data to the plurality of hard disks 7-10.
Transfer write data. The empty storage device information transmission / reception unit 3 transmits an empty device addition request notification from the empty storage device addition request determination unit 13 in the storage data management unit 4 to the external system, and receives information that an empty device is attached. Data recovery and storage data RA
It serves to inform the ID processing unit 5.

The stored data management unit 4 is a management unit that manages various data stored in the plurality of hard disks 7 to 10. In the storage data management unit 4, a storage data recovery method determination unit 11, a free storage device free area investigation processing unit 12, and a free storage device addition request determination unit 1
3, a data storage location and a RAID group management table 14.

The storage device failure detection unit 6 monitors whether a failure occurs in the hard disk, which is a storage device, and the stored data is destroyed. If the stored data is destroyed, it is stored. It is a detection unit that notifies the data recovery method determination unit 11.

The storage data RAID processing section 5 performs a redundant distributed storage processing of normal write data. In addition, according to the stored data recovery method determined by the stored data recovery method determination unit 11 as the determination means, the stored data management unit 4
Based on the RAID group management table 14 therein, the recovery processing of the destroyed storage data is performed.

The empty storage device information transmission / reception unit 3, the storage data management unit 4, the storage data RAID processing unit 5, and the storage device failure detection unit 6 are composed of, for example, a CPU and a memory.

FIG. 2 shows an example of the structure of a RAID group. RAID groups RG0, RG1, RG2, ...
And so on. In the RAID group RG0, the parity disk is disk number 0, the disk number is 1,
It is composed of two and three. RAID group RG1
The parity disk has a disk number of 5 and is composed of disk numbers 4, 6, and 7. In the RAID group RG2, the parity disk is disk number 11, and is composed of disk numbers 8, 9, and 10.

FIG. 3 shows a RAID group management table 1
4 shows a configuration example of No. 4. That is, the RAID group management table 14 has a group number RGN, a constituent hard disk number HDN, and a parity device number PD.
N, the number M of constituent hard disks, and whether or not the number M of constituent hard disks exceeds a certain value (B) (○ × for M ≦ B) is managed.

The number M of hard disks constituting the RAID group and MT representing the reliability of the data stored therein
The relationship of BF can be expressed by the following formula. However, E is the failure rate of the hard disk.

MTBF = 1 / (1- {ME (1-E)
^M-1 + (1-E) ^M }) From this equation, it is understood that the larger M is, the smaller MTBF is, and the reliability is lowered. That is, as the number of hard disks forming the RAID group increases, the reliability of the data stored therein decreases. When it is desired to keep the data reliability MTBF of this information storage device above a certain MTBF ₀ , M may be set to be equal to or less than a certain maximum M (= B) that satisfies the following equation.

MTBF ₀ ≤ 1 / (1- {ME (1-E)
^M-1 + (1-E) ^M }) In the conventional system, if the data in the storage device is destroyed, it must be completely restored if a new empty storage device exists in the system and is not installed. I can't. If it is not mounted in the system, the other storage devices of the RAID group to which the defective storage device belongs loses the redundancy of the stored data, resulting in extremely low reliability.

In the present invention, two methods are mentioned for the purpose of avoiding such a state where reliability is extremely lowered. One is to use the remaining storage devices of the RAID group to which the defective storage device belongs to other RAID
This is a method of securing an empty storage device by integrating each device into a group, restoring the destroyed data to this, and making the stored data redundant again to secure reliability.

The other is to investigate and secure an empty area for one storage device from a storage device that does not belong to the RAID group in which the failure has occurred, and restore the data destroyed in this free area in a distributed manner, and to perform the RAID in which the failure has occurred. This is a method of preventing parity deterioration by re-parity matching in the remaining storage devices of the group.

First, the first method will be described with reference to FIG. In FIG. 4, the RAID group RG0 includes a hard disk HD0 as a parity disk and hard disks HD1, HD2, and HD3. Also, RA
The parity disk of the ID group RG1 is a hard disk HD5 and is composed of hard disks HD4, 6, and 7.

As shown in FIG. 4A, when the hard disk HD6 of the RAID group RG1 malfunctions and the stored data is destroyed, it is considered to integrate it with the RAID group RG0, for example. The RAID group integration process is shown in FIG.

First, the hard disk H in which the defect has occurred
The data of the other hard disks HD4, HD5, and HD7 of the RAID group to which D6 belonged is E in the horizontal direction.
XOR operation 20 is performed. When the HD 6 is not the parity device but the normal storage data, the result of the EXOR operation 20 is overwritten on the hard disk HD 5 which is the parity device of the RAID group RG 1. If the hard disk HD6 is a parity device, it is not necessary to perform such processing.

The result of this EXOR operation 20 and R
The data of the hard disk HD0 which is the parity device of the AID group RG0 is also subjected to the EXOR operation 21 in the horizontal direction, and the result is overwritten on the hard disk HD0.

FIG. 5 shows the RAID groups in the RAID group management table 14 before and after the integration. 5A shows the status before the integration, and FIG. 5B shows the RAID group management table 14 after the integration. That is, the number M of hard disks in the RAID group RG0 is integrated to become "7", and the number M of hard disks in the RAID group RG1 disappears due to integration to become "0".

The storage data thus destroyed can be restored, and further, the storage data of the normal hard disks HD4, HD5, HD7 of the RAID group RG1 can be made redundant again to restore the reliability.

Next, the second method will be described with reference to FIG. In FIG. 6, the RAID group RG0 is composed of hard disks HD0, 1, 2, 3 and
D group RG1 has hard disks HD4, 5, 6,
7 and the RAID group RG2 is composed of hard disks HD8, 9, 10, and 11.

As shown in FIG. 6, RAID group RG
When a problem occurs in the hard disk HD2 of 0 and the stored data is destroyed, for example, RAID group RG
Consider using a free area of RG2 (area in which stored data is not written).

First, a free area corresponding to the data capacity of the destroyed effective data is searched. Here, the free space S0 and the RA of the hard disk HD4 of the RAID group RG1
The free space S3 of the hard disk HD8 of the ID group RG2 is used. Here, it is assumed that S0 is a capacity for storing the destroyed data D5 and S3 is a capacity for storing D4.

If the defective hard disk HD2 is not a parity device, D0, D2, D6 to D
4 is written in S3, and D1, D3, D7 and D
Write 5 to S0. At this time, of course, D5, D
Parity matching must be taken between 8, D9, P1 and between D4, D10, D11, P2. Where D4, D
Data to be restored corresponding to 5 need only be data that is effective as stored data, and it is not necessary to perform this series of processing for all data for one hard disk.

In addition, the hard disk HD which has a problem
If 2 is a parity device, first RAID group R
The effective storage data of one of the normal hard disks in G0, for example, the data D2 and D3 of the hard disk HD1 here are moved to S3 and S0, respectively. At this time, of course, the parity matching is re-established. Then, the normal hard disks HD0 and HD3 of the RAID group RG0 are added to the hard disk HD1 that has become an empty device.
Write parity to and hard disk HD1 RA
It is assumed to be the parity device of the ID group RG0.

As described above, the destroyed storage data can be restored, and the data in the normal hard disks HD0, HD1, HD3 of the RAID group RG0 can be made redundant again to recover the reliability.

Next, details of a series of processes of the present invention in this information storage device will be described with reference to the flowchart of FIG. First, n representing the number of requests to add an empty storage device is cleared to 0 (ST1). In the normal state, initially n =
Since it is 0, YES is selected in step ST2. Further, the storage device defect detection unit 6 checks the storage device for defects (ST6).
Will be selected, and it will loop here and continue to loop until a problem occurs.

When a problem occurs, first, the empty storage device free area check processing unit 12 as a checking means checks whether or not an empty storage device is mounted in the system (ST7), and if there is, destroys it. The stored data is restored by the storage data RAID processing unit 5 (S
T8).

If it does not exist, the storage data recovery method determination unit 11 first checks the RAID group management table 14 and finds a group having the smallest number M of hard disks among the other normal RAID groups (ST).
9). _Let M of this group be M _min . Then, the number of hard disks constituting the RAID group to which the defective hard disk belongs is set to M ₀ and it is checked whether the following expression is satisfied. Here, “B” is the minimum number of constituent hard disks that guarantee certain reliability.

M ₀ + M _min ≦ B As a result of the examination, if the above expression is satisfied, the storage data restoration method determination unit 11 selects the method of integrating the RAID groups (ST15). If the above equation cannot be satisfied, the empty storage device free area check processing unit 12 first checks and secures a free area for the restored data (ST11).
A method is selected by the storage data RAID processing unit 5
Data is restored to this free area by (ST1
4).

When the empty area cannot be secured, the empty storage device addition request determination unit 13 recognizes this and requests the addition of an empty storage device to the outside through the empty storage device information transmission / reception unit 3 "n". Is added to "1" (ST13). In this case, the storage data recovery method determination unit 11 causes the RAI
A method of integrating the D groups is selected (ST15), and it is possible to prevent the reliability of the storage data of the RAID group to which the defective hard disk belongs from from extremely decreasing.

Further, when n = 0 is not satisfied and the addition of an empty storage device is requested to the outside through the empty storage device information transmission / reception unit 3, the additional information of the empty storage device from the empty storage device information transmission / reception unit 3 is checked. (ST3). When the RAID group is added, the RAID group management table 14 is checked and the RAID group M> B is M
From the largest to the largest by the storage data RAID processing unit 5
ID group division processing is performed (ST4), and the number of added empty devices is set to n ₀ and n is set to n−n ₀ (ST).
5).

As described above, when a failure occurs in the storage device and the stored data is destroyed, the reliability of the storage device should not fall below a certain value even if there is no empty device in the information storage device. Yes, while maintaining a certain degree of reliability even when it drops below a certain value,
It is possible to notify the outside to the effect.

FIG. 8 shows a schematic structure of an information storage device according to the second embodiment of the present invention. That is, the information storage device includes a plurality of optical disc drives 31 to 35, a plurality of optical discs 36 to 41, and the optical disc drive 31.
-35 and optical disks 36-41, the optical disk drive can be automatically loaded or unloaded with the optical disk autochanger 42, the data transmitter / receiver 22, the empty storage device information transmitter / receiver 23, the storage data manager 24, the storage data. It is composed of an optimum placement processing unit 25 and a storage data RAID processing unit 26, which are connected by a common bus 44.

The external system 21 includes a data transmission / reception section 22.
The read data and the write data are transferred to the optical disks 36 to 41 via the. Empty storage device information transmitter / receiver 2
3 is the storage data optimum when the empty device addition request notification from the empty storage device addition request determination unit 27 in the storage data management unit 24 is transmitted to the external system or the information that the empty device is attached is received. It serves to inform the placement processing unit 25.

In the storage data management unit 24, the optical disk 3
Various management of the data stored in 6 to 41 is performed. In the storage data management unit 24, an empty storage device investigation processing unit 27, a RAID group new establishment determination unit 2
8, a RAID group division determination unit 29, a data storage location and a RAID group management table 30.

The stored data optimum arrangement processing section 25 is an optimum arrangement means for optimizing the arrangement of the stored data between the optical disks in order to improve the access response speed of the stored data.

The storage data RAID processing unit 26 performs a normal redundant distributed storage processing of write data and a recovery processing of storage data when a failure occurs. In addition, it is a processing unit that performs a RAID group division process based on the RAID group management table 30 in the storage data management unit 24 according to an instruction from the RAID group division determination unit 29.

Here, the optimum arrangement of stored data will be described. The time required for loading / unloading the optical disks 36-41 to / from the optical disk drives 31-35 in the optical disk autochanger 42 adversely affects the response speed of access to stored data.

In order to improve the access response speed, it is necessary to devise a method of loading and unloading the optical disk as much as possible. Therefore, consider arranging the storage data with high access frequency as much as possible in the optical disk loaded in the optical disk drive, and expelling the storage data with low access frequency to the removed optical disk, and consider the parity matching process during the write process. And the same RAI as possible
RAI as if the D group optical disc is loaded
Consider creating a group with high access frequency and a group with low access frequency even at the D group level. This is the storage data optimum allocation processing, which is performed by the storage data optimum allocation processing unit 25.

Further, although this is accompanied by the movement of the stored data, the movement of the data is performed in the RAID group by RA.
The amount of processing is considerably smaller than that performed between ID groups. RA
The optimum arrangement of the storage data in the ID group can be used because it is not necessary to re-match the parity when exchanging data between the same blocks of the optical disks in the RAID group. This is because the parity matching process is always accompanied when performing between RAID groups.

Next, the determination of RAID group division and the change of the status of the RAID group in the information storage device will be described. 9A to 9H show how the states of the RAID group and the optical disks forming the RAID group in the information storage device are in the process of increasing the storage data while the information storage device is being used. The change is shown in steps 1 to 8. FIGS. 10A to 10H show the transition of the contents of the RAID group management table 30 at this time similarly in steps 1 to 8.

As the stored data increases, as shown in FIGS. 9 and 10, step 1 → step 2 → step 3 → step 4 and the number M of optical disks constituting the RAID group becomes 2 → 3 → 5 → 6. You can see that it will increase. The number of optical disks forming a RAID group that guarantees the reliability of the storage data in this information storage device is B =
5, the RAID group 0 at the time of step 4
The number of optical disk components M> B.

Therefore, the RAID group is divided into two,
As a result, M of the two groups is 2 and 5, respectively, and both are smaller than B = 5. As a result, it is possible to prevent the stored data from becoming lower than the certain reliability. The same applies to RAID group 0 in step 7, and is divided as in step 8.

Here, the processing for dividing the RAID group into two will be described. 11 and 12
9A and 9B show access frequency distributions of storage data of optical disks in RAID group 0 in steps 4 and 7 shown in FIGS. 9 and 10, respectively. Data with high access frequency of stored data is concentrated on some optical disks, and the data access frequency of the optical disks in the RAID group is biased. In this way, the storage data optimum allocation process is performed in advance within the RAID group.

When the RAID group is divided, for example, in the case of step 4, it is divided into an optical disk 1 in which storage data having a high access frequency is concentrated and optical disks 2 to 5 in which low storage data is concentrated, and parity matching is performed for each. To do. The same applies to the case of step 7, and the optical discs 1 and 7 on which stored data with high access frequency are concentrated,
It is divided into optical disks 8 to 10 in which low storage data are concentrated, and parity matching is performed for each.

As a result, a group having data having a high access frequency in a concentrated manner and a group having a data having a low access frequency are generated for each RAID group, and a large amount of data between the RAID groups generated during the data optimum allocation processing is generated. The movement can be minimized.

As described above, when the information storage device is used and the number of optical disks forming the RAID group increases in the process of gradually increasing the storage data, the reliability of the storage data of the optical disk becomes lower than a certain value. Can be prevented by dividing the RAID group,
In addition, it is possible to notify a request for a blank optical disk required at this time. Further, at this time, the RAID group is divided while considering the optimum arrangement of the data in the storage device, so that the RAID group can be processed during the optimum data arrangement processing.
Occurrence of data movement between D groups is minimized, and therefore, efficient data optimal placement is performed and performance is improved.

As described above, according to the above-described embodiment of the present invention, a failure occurs in a storage device under the control of a storage system having redundancy between a plurality of media, and when the storage data is desired to be restored, a failure occurs. If there is no new storage device that is the same as the generated device, data recovery is performed by reconfiguring the redundantly distributed storage data group to secure a free storage device.
Alternatively, the stored data can be restored to a redundant storage form by restoring the data to the free area of the storage device secured as the data restoration destination.

Further, by managing the reliability of the stored data and selecting and deciding the data recovery method based on this, it is possible to prevent the reliability from lowering below a certain standard, and nevertheless the reliability is restored. If insufficient, it is possible to notify the outside of the addition and replacement of a new storage device, and when the storage device is added, the reliability can be restored and restored to a sufficient state.

Further, as the storage data increases, the number of devices constituting the redundant distributed storage data group increases, so that the reliability of the storage data decreases. When it is mounted by requesting addition, it is possible to recover the lowered reliability by dividing the redundant data group. Further, at this time, by dividing the RAID group while considering the optimal placement of the data in the storage device, the occurrence of data movement between the RAID groups is minimized during the data optimal placement processing, and efficient data optimization is achieved. Placement is done to improve performance.

Further, here, the RAI is performed for each storage device.
Although the storage location (parity device) of the parity data of the D group is fixed, the present invention is not limited to this, and the same can be applied to a block unit of the storage device or the like.

Although a hard disk or an optical disk has been described as an example of the storage device here,
The present invention is not limited to this, and can be applied to all information storage media capable of storing data.

[0085]

As described in detail above, according to the present invention,
(EN) Provided is an information storage device which does not need to mount a new empty storage device when a storage device is destroyed and maintains the redundancy of other storage devices in a RAID group without lowering reliability. You can

In the process of increasing the stored data, R
Prevents deterioration of reliability in the AID group and
It is possible to provide an information storage device that can prevent the deterioration of performance by reducing the occurrence of data movement between RAID groups due to the establishment of a new group.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of an information storage device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a configuration example of a RAID group.

FIG. 3 is a diagram showing a configuration example of a RAID group management table.

FIG. 4 is a diagram for explaining a first method in recovery of destroyed storage data.

FIG. 5 is a diagram for explaining before and after RAID group integration.

FIG. 6 is a diagram for explaining a second method in recovering destroyed storage data.

FIG. 7 is a flowchart for explaining a series of processing operations in the information storage device.

FIG. 8 is a block diagram showing a schematic configuration of an information storage device according to a second embodiment of the present invention.

FIG. 9 is a diagram for explaining changes in the status of a RAID group in an information storage device and the optical disks forming the RAID group.

FIG. 10 is a diagram for explaining changes in the contents of a RAID group management table.

FIG. 11 is a diagram showing an access frequency distribution of stored data in step 4.

FIG. 12 is a diagram showing an access frequency distribution of stored data in step 7.

FIG. 13 is a diagram showing a configuration example of a general information storage device.

FIG. 14 is a diagram showing an example of a management table relating to data storage locations.

FIG. 15 is a diagram for explaining the relationship between a hard disk and block numbers.

FIG. 16 is a diagram showing a configuration example of a RAID group.

FIG. 17 is a diagram for explaining a management table showing the configuration of each RAID group.

FIG. 18 is a diagram for explaining restoration of stored data.

[Explanation of symbols]

 1 ... External system 2 ... Data transmitting / receiving unit 3 ... Empty storage device information transmitting / receiving unit 4 ... Stored data management unit 5 ... Stored data RAID processing unit 6 ... Stored device failure detection unit 7,8,9,10 ... Hard disk 11 ... Stored data Restoration method determination unit 12 ... Empty storage device empty area investigation processing unit 13 ... Empty storage device addition request determination unit 14 ... RAID group management table

Claims (10)

[Claims] 1. An information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group, wherein the group and a plurality of information forming the group. A storage information management means for redundantly storing and managing information in a storage medium, a detection means for detecting the occurrence of a failure of the information storage medium, and a detection means for detecting a failure of the information storage medium, An investigating means for investigating whether or not there is an empty information storage medium in which no information is stored, and a storage of the information storage medium detected by the detecting means when there is no empty information storage medium by this investigating means. And a processing unit that performs a recovery process using the above-mentioned stored information management unit according to the recovery method determined by this determination unit. Information storage device. 2. The information storage device according to claim 1, wherein the determining means determines based on the reliability of the group stored in a redundantly distributed manner. 3. An information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group, and occurrence of a defect in the information storage medium is detected. Detecting means, an investigating means for investigating whether or not there is an empty information storage medium in which no information is stored when the detecting means detects a defect in the information storage medium, and When the information storage medium does not exist, the above-mentioned trouble has occurred depending on the determination means for selectively determining the restoration method of the stored information of the information storage medium detected by the detection means, and the restoration method determined by this determination means. An information storage device comprising: a processing unit that performs a process of restoring information stored in an information storage medium and a writing process of redundantly distributing stored information. 4. An information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group, and occurrence of a defect in the information storage medium is detected. Detecting means, an investigating means for investigating whether or not there is an empty information storage medium in which no information is stored when the detecting means detects a defect in the information storage medium, and the investigating means When there is no information storage medium, there is provided an integration means for integrating the information storage medium belonging to the group of the information storage media in which the defect has been detected by the detection means, into another group, apparatus. 5. An information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored in each group, and occurrence of a defect in the information storage medium is detected. Detecting means, an investigating means for investigating whether or not there is an empty information storage medium in which no information is stored when the detecting means detects a defect in the information storage medium, and When there is no information storage medium, a securing means for investigating and securing a free area of the plurality of information storage media corresponding to the amount of stored information stored in the information storage medium in which the defect detected by the detection means has occurred And storage means for restoring and storing the storage information stored in the information storage medium in which the defect detected by the detection means has occurred in the free area secured by the securing means. Information storage device, wherein the door. 6. An information storage device in which a plurality of information storage media are connected in parallel to form one group, and information is redundantly distributed and stored for each group, and occurrence of a defect in the information storage medium is detected. Detecting means, an investigating means for investigating whether or not there is an empty information storage medium in which no information is stored when the detecting means detects a defect in the information storage medium, and When there is no information storage medium, a securing means for investigating and securing a free area of the plurality of information storage media corresponding to the amount of stored information stored in the information storage medium in which the defect detected by the detection means has occurred And a deciding means for deciding whether or not a new information storage medium is required from the reliability of the group stored in a redundantly distributed manner when the vacant area is not secured by this securing means, Information storage device comprising a notification unit for notifying to the outside, by comprising a when the it is determined to require a new information storage medium judgment means. 7. Information in which a plurality of portable information storage media are loaded / unloaded to read / write information, a group is composed of the plurality of portable information storage media, and information is redundantly distributed and stored for each group. In the storage device, the management means for managing the access frequency for reading and writing for each portable information storage medium forming the group, and the access frequency for each portable information storage medium forming the group managed by this management means are set. Accordingly, the information storage device is provided with: an optimum arrangement means for optimally arranging information stored in a plurality of portable information storage media forming a group. 8. An information storage device, wherein a group is constituted by a plurality of portable information recording media, and the information stored in the portable information storage media constituting the group is redundantly distributed and stored, and the group is constituted. The portable information storage medium comprises a judgment means for judging the division of the group when the number exceeds a predetermined number, and a processing means for dividing the group when the judgment means judges the division of the group. An information storage device characterized by the above. 9. Information in which a plurality of portable information storage media are loaded / unloaded to read / write information, a group is composed of the plurality of portable information storage media, and information is redundantly distributed and stored for each group. In the storage device, a management unit that manages an access frequency for reading and writing each portable information storage medium that constitutes the group, and a group when the number of portable information storage media that constitutes the group exceeds a predetermined number. And a determination means for determining the division of the group, and when the determination means determines the division of the group, the group is divided based on the access frequency of each portable information storage medium forming the group managed by the management means. An information storage device comprising: a processing unit. 10. Information in which a plurality of portable information storage media are loaded / unloaded to read / write information, a group is composed of a plurality of portable information storage media, and information is redundantly distributed and stored for each group. In the storage device, a management unit that manages an access frequency for reading and writing each portable information storage medium that constitutes the group, and a group when the number of portable information storage media that constitutes the group exceeds a predetermined number. And a determination means for determining the division of the group, and when the determination means determines the division of the group, the group is divided based on the access frequency of each portable information storage medium forming the group managed by the management means. Optimal for optimally arranging the processing means and the information stored in the plurality of portable information storage media forming the groups divided by the processing means Information storage device being characterized in that comprising a location means.