JPH10320132A - Information storage system control method and information storage system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the fault-tolerant performance of an information storage system such as a RAID system composed of a storage device group having a redundant configuration provided with a portable medium. SOLUTION: Input data 1 on a buffer 3 is divided into data D1 to D4 by an array control unit 2, a parity DP is generated by a parity generation / check circuit 4, and a plurality of drives 7a to 7e are generated by a SCSI control unit 5. In a configuration in which the drive is distributed and stored in the portable media 8a to 8e, the spare drive 9, the drives 7a to 7e, the spare drive 9, and the portable media 8a to 8e are stored.
A drive mapping switch unit 6 for managing the correspondence of e while updating the mapping table 6a is provided. At the time of the first failure of the drives 7a to 7e, the portable medium of the failed drive is moved to the spare drive 9 to continue use. When a further failed drive occurs, the portable device of the failed drive is used while replacing it with the portable media of another healthy drive, and the operation is continued until the last healthy drive becomes one.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information storage technique, and more particularly to a data transfer rate by dispersing and storing arbitrary data and redundant data generated from the data in a plurality of storage devices or media. (Redundant Arrays of Inexpensive D)
isks) Regarding technology that is effective when applied to systems and the like.

[0002]

2. Description of the Related Art In the field of information processing technology, as a data storage system, data and redundant data generated from the data are distributed and stored in a plurality of storage devices, so that an input / output speed by distributed parallel transfer of data is obtained. RAID systems that realize higher quality and higher data reliability have become widespread.

[0003] By the way, such a RAID system is also constituted by a plurality of portable media. For example, in a technology described in Japanese Patent Application Laid-Open No. 08-263226, in a RAID system in which a plurality of portable media are connected in parallel and used as a disk array,
Data is distributed and stored in a plurality of media, and parity is generated to provide redundancy, thereby realizing reliability of data management. According to this technology, when an inaccessible storage device occurs, a failure caused by a medium is premised, and the data corresponding to the storage device is unconditionally restored to another medium so that the operation can be continued, and the inaccessibility factor can be improved. However, the factors such as whether the failure is due to the failure of the portable medium itself or the temporary removal of the portable medium are separated, and accurate media management that avoids duplication of the portable medium where the restoration data is stored is avoided. Trying to realize.

[0004]

In the prior art, when a failure occurs in one storage device in a RAID system composed of a plurality of media, the reliability is reduced without stopping the system by degenerating the storage system. Is kept. But,
If another storage device fails (two or more)
Has a technical problem that the stored data cannot be read or written, and the system stops.

Further, in the above-mentioned prior art, when the inaccessibility is detected, the data restoration processing to the new medium is executed unconditionally, so that the overhead in the event of a failure becomes unnecessarily large and the data on the portable medium is unnecessarily large. There is also a technical problem that it cannot be used effectively. That is, non-contact type portable media such as optical discs have a low failure rate due to the medium itself and have a large number of failures in the storage device, so the data on the portable media is alive. Although there is a high possibility, in the case of the above-described conventional technique, a portable medium having a high possibility of being sound and having been loaded into the storage device at the time of detecting a failure and data stored in the portable medium cannot be effectively used.

An object of the present invention is to provide a control technique of an information storage system capable of improving fault-tolerant performance in an information storage system including a storage device having a redundant configuration provided with a portable medium.

Another object of the present invention is to provide a control technique of an information storage system capable of shortening the time required for failure recovery in an information storage system including a storage device having a redundant configuration with a portable medium. It is in.

Another object of the present invention is to provide an information storage system capable of realizing effective use of a portable medium and data on the portable medium in an information storage system including a storage device having a redundant configuration provided with the portable medium. An object of the present invention is to provide system control technology.

[0009]

According to the present invention, a plurality of data units and at least one redundant data generated from the data units are distributed and stored in a plurality of storage devices provided with exchangeable portable media. In an information storage system, when a failure occurs in an arbitrary storage device, the operation is continued by moving a portable medium of the failed storage device to another sound storage device and using it.

According to the present invention, at least one spare storage device is arranged as necessary, and in the event of the first storage device failure,
The portable media in the failed storage device is moved to the spare storage device for use, and more storage devices than the number of spare storage devices have failed, and the number of storage devices is less than the number of portable media. In this case, the operation is continued until the number of healthy storage devices becomes one while replacing the portable medium with each storage device. If no spare storage device is provided, the operation is continued while replacing the portable medium with each storage device after the first failure occurs. When the operation is performed while replacing the portable medium, the storage device for replacing the portable medium is sequentially changed so that the load is not concentrated on a specific storage device.

In order to realize the above control, as an example, a plurality of storage devices (drives) in which data and redundant data are stored in a distributed manner and provided with a portable medium, and a plurality of portable devices are provided by these storage devices. An operation of distributing and storing data with redundancy in a medium, an operation of detecting a failure in a storage device, associating a drive number with an identification number unique to a portable medium in the drive, and A control logic for executing an operation for controlling the mapping of the identification number and an operation for arbitrarily rewriting and controlling the drive number at the time of failure can be provided.

[0012]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a conceptual diagram showing an example of a configuration of a RAID system including a portable medium as an embodiment of an information storage system in which a control method of the information storage system according to the present invention is implemented. FIG. 3 is a conceptual diagram showing an example of a control method of the information storage system of the present embodiment and control information used in the information storage system. FIG. 3 is a diagram showing a control method of the information storage system of the present embodiment and an operation of the information storage system. It is a flowchart which shows an example.

In the following description, as an example, an arbitrary portable medium such as an optical disk, a magneto-optical disk, a compact disk, a mini disk, and a floppy disk is provided so as to be interchangeable, and a drive having a SCSI interface is provided as a storage device. The RAID system will be described.

As exemplified in FIG. 1, the RAID system according to the present embodiment records / reproduces data on / from each of a plurality of removable media 8a, 8b, 8c, 8d and 8e which are detachably loaded. Drives 7a, 7b, 7c, 7d, 7 to be performed
e, an array control unit 2 that controls the entire system and stores the input data 1 arriving from a higher-level device in a buffer 3 by dividing the input data 1 into redundancy so as to have redundancy.
Parity DP from data D1 to D4 divided above
Parity generation / check circuit 4 for generating or checking the data D1 to D4 and restoring the data D1 to D4, and for driving the divided data D1 to D4 and the parity DP to the drive 7
a to 7e, a SCSI control unit 5 for determining a drive failure and detecting a drive failure, a drive mapping switch unit 6, for controlling mapping of the plurality of drives 7a to 7e and SCSI_ID, and portable media 8a to 8e.
And a spare drive 9 for switching when a failure occurs in the drives 7a to 7e, and movement of the portable media 8a to 8e with respect to the drives 7a to 7e and the spare drive 9 based on a mapping table 6a as illustrated in FIG. It is composed of a medium transport unit 10 for performing operations and changing operations.

The array control unit 2 is composed of, for example, a microprocessor (not shown) and a program (control logic) for controlling the microprocessor.
A series of control operations as illustrated in the flowchart of FIG. 3 described below are performed in cooperation with the SI control unit 5.

The plurality of drives 7a to 7e, the spare drive 9, the medium transporting unit 10, and the SCSI control unit 5
They are connected by a bus 11. The SCSI control unit 5 and the array control unit 2 are connected by an interface 12, and notify the occurrence of a failure in the drives 7a to 7e, the spare drive 9, and the like.

In the case of the present embodiment, the input data 1 is divided into four data D1 to D4 on the buffer 3 by the array control unit 2 so as to have redundancy, and divided by the parity generation / check circuit 4. Parity DP of data D1 to D4
Is generated and checked, and the five data of the data D1 to D4 and the parity DP divided by the SCSI control unit 5 are distributed, and the five drives 7a to 7e (that is, five portable media 8a to 8e) Control to which of the above. For example, during the input / output processing, the data D1 to D1
When a failure occurs in any of D4, an operation of restoring the failed data from the remaining healthy data and the parity DP is performed.

Since the plurality of portable media 8a to 8e can be loaded into any of the drives 7a to 7e and the spare drive 9 (replacement by the medium transport unit 10), the plurality of portable media 8a to 8e and It is necessary to dynamically manage the loading state (correspondence relationship) of the portable media 8a to 8e in the present embodiment, and in the present embodiment, the drive mapping switch unit 6 has a mapping table 6a illustrated in FIG. By doing this management.

FIG. 2 shows an example of the mapping table 6a provided in the drive mapping switch section 6 in the present embodiment. The plurality of drives 7a to 7e and the spare drive 9 have a unique SCSI_ID.
That is, each of the drives 7a to 7e has a SCSI_
ID: “0” to “4” are assigned to the spare drive 9 by the SCS
I_ID: “5” is assigned.

The portable media 8a, 8b, 8c, 8d,
8e has a medium ID unique to the medium in advance.
(“A” to “E”) are recorded, and the portable media 8a to
8e, and the drives 7a to 7e used (loaded) by each of the portable media 8a to 8e,
SCSI_ID of the spare drive 9 (“0” to “4”,
"5") is associated with the mapping table 6a. The data D1 to D4 and the parity DP on the divided buffer 3 are respectively transferred to the target portable media 8a to 8e based on the medium ID. At this time, the SCSI controller 5 stores the data in the mapping table 6a. Drives 7a to 7e loaded with the target portable media 8a to 8e based on the
Alternatively, the data D1 to D4 and the parity DP are transferred to the spare drive 9.

Hereinafter, an example of the operation of the RAID system and the control method thereof according to the present embodiment will be described with reference to a flowchart illustrated in FIG. 3 and a transition diagram of the degeneration level illustrated in FIG.

In the initial state, the portable media 8a to 8e ("A" to "E") are stored in the drives 7a to 7e (SCSI_ID: "0" to "4") as shown in "normal state" in FIG. Are operating in a loaded state.

In this state, first, an operation when a failure occurs in the first drive will be described. When the first failure occurs in the drive 7c (SCSI_ID = 2) in which the portable medium 8c (medium ID “C”) is inserted, the information is notified to the SCSI control unit 5 through the SCSI bus 11. Based on this, the SCSI control unit 5 causes the medium transport unit 10 to replace the portable medium 8c (medium ID “C”) with the spare drive 9 and continues the processing. At this time, if the data can be normally transferred to the portable medium 8c, the SCSI controller 5 rewrites the mapping table 6a as the state numbers in FIG.
b, 7d, 7e, and the spare drive 9 to continue the process, notify the array controller 2 that the drive 7c (SCSI_ID = 2) has a failure through the interface 12, and set the degeneration level to "1". Set the flag.

As described above, the portable medium 8c of the failed drive is moved to the spare drive 9, and the portable medium 8c
By switching the data input / output destination to the spare drive 9, the processing is continued without requiring a large overhead processing such as data restoration or the like, and even if one drive fails, the conventional drive can be used. There is no need to perform degeneration as in a RAID system composed of fixed disks.

Next, an example of a case where a failure occurs in the second drive (for example, the drive 7a) while the spare drive 9 is used (degeneration level "1") will be described. If a failure occurs in the state of the degeneration level "1", the mapping table 6a is rewritten as the state. Here, the portable medium 8 moved from the failed drive 7a
a and the existing portable medium 8b (medium ID “A”, “B”) share the same drive 7b (SCSI_ID = 1), so that the existing portable medium 8b (medium ID
After the process for “B”) is completed, the process for the failed portable medium 8a (medium ID “A”) is performed. Here, before performing processing on the portable medium 8a (medium ID “A”) moved from the failed drive 7a, it is determined whether or not normal access is possible. If the processing of the portable medium 8a (medium ID “A”) cannot be continued even at the destination drive 7b, it is determined that the portable medium 8a (medium ID “A”) has some defect, The access to the portable medium 8a is blocked, for example, the control is switched to the degeneration control of the conventional RAID system composed of fixed disks, and the degeneration operation (degeneration level “Z”) is performed on four media. Although data can be input and output at the degeneration level "Z", data cannot be restored when a failure occurs.

The flag of the degeneration level "Z" is a flag independent of the degeneration levels 1 to 4. Even at the degeneration level "Z", the degeneration operation of the drive by switching the portable medium is continued. .

When processing is continued without a medium failure,
Before proceeding to the processing of the next data, it is confirmed whether or not the restoration processing of the failed drive has been performed, and then proceed to the next processing. When a failure is recovered by replacing a failed drive or the like, a recovery flag is raised for this drive, and the mapping table 6a is returned to a normal state before data processing is performed in accordance with the flag, and the degeneration level is set to "0" to perform a normal operation. Run.

Next, in the mapping table 6a of FIG. 2, the number n of drives becomes smaller than the number m of portable media due to the failure of the drive, so that a plurality of portable media can be switched to share an arbitrary drive. In a state where the operation is continued while the operation is being performed, the portable medium discharged for the replacement is rewritten so as to be mapped to the SCSI_ID of another drive. By changing the SCSI_ID of the overlapping drives sequentially (from the state of the mapping table 6a), it is possible to prevent the load at the time of switching (loading) of the portable medium from being concentrated on one specific drive. Can be.

If a further failure occurs in a plurality of drives at the degeneration level "1", the process of switching is executed in the same manner, and the process can be continued until the number of normal drives becomes one. When the last one drive fails, the next data is not accepted until the recovery of the drive is completed and the recovery flag is set.

In the case of the degeneration level "Z", a failure of the portable medium has occurred, and the degeneration operation is performed (in the case of this embodiment,
In the operation state using four portable media), if the medium abnormality is determined again, the next data will not be accepted until the data recovery processing with the new medium is performed.

The transition of the degeneration level in the above degeneration processing and the like is illustrated in FIG. 4, and the above-described operation will be comprehensively described with reference to the flowchart of FIG. Become like

First, it is checked whether the current number m of media is larger than the number n of healthy drives (step 101). If it is, the mapping table 6a is rewritten (step 10).
2) and while performing the operation of replacing the portable medium according to the mapping table 6a (step 103), and determining whether or not the recovery flag is set (step 10).
4), execute data input / output, etc. (step 10)
5) is repeated while monitoring for the occurrence of a failure (step 106).

When it is detected in step 104 that the recovery flag is set, the mapping table 6a is rewritten (initialized) according to the recovery state of the medium failure or the drive failure.
(Step 113), and also initializes the degeneration level, the number m of media, the number n of healthy drives, the recovery flag, and the like (step 114).

On the other hand, when a drive failure is detected in step 106, the failed drive is closed and the number n of healthy drives is updated (decremented) (step 10).
7) Then, it is determined whether or not the number n of healthy drives is more than one (step 108). If it is determined in this determination that the number n of healthy drives is 0, the operation stop is reported to the upper-level device, and reception of data input / output is stopped.
A drive recovery process such as replacement of a failed drive and a recovery flag set are executed (step 118), and the process returns to step 104 to execute steps 113 and 114, thereby resuming the reception of data input / output.

If the number n of healthy drives is one or more in step 108, the operation can be continued by replacing the portable medium. Therefore, first, the mapping table 6a is rewritten corresponding to the blockage of the failed drive (step 10
9) The portable medium is moved from the failed drive to the spare drive (at the time of the first failure) or another drive (at the time of the second or subsequent failure) (step 110).

Then, the destination drive verifies the portable medium moved from the failed drive (step 1).
11) If it is determined that the state is sound, the degeneration level is incremented (step 112), and the process returns to step 101.

On the other hand, if an abnormality is detected in the portable medium moved from the failed drive in step 111,
It is checked whether or not the current degeneration level is “Z” (that is, whether or not there is a failure of the second or subsequent portable medium). If the degeneration level is not “Z”, “Z” is set as the degeneration level. Is set and the number m of media is decremented (step 11).
6) Returning to step 101, the operation is continued at the degeneration level = “Z”. In this state of the degeneration level = “Z”, all four data D1 to D4 and the parity DP cannot be held, so that normal data input / output is possible, but data recovery in the event of a failure cannot be performed.

Although not particularly shown, step 111
It is needless to say that the data restoration using the new medium may be executed immediately when the failure of the portable medium is found. In this case, overhead for restoring data to a new medium is generated, but data redundancy is maintained.

At step 115, if the current degeneration level is "Z", it is determined that the degeneration operation, which is the input / output of only normal data, is impossible because it is the second failure of the portable medium. Then, while reporting the suspension of data input / output of the higher-level device, executing data restoration using a new medium and setting a recovery flag (step 117),
Returning to step 04, the above-mentioned steps 113 and 114 are executed, and at the same time, the host device is notified of the resumption of data input / output and the operation is resumed.

As described above, according to the information storage system of the present embodiment, in a RAID system constituted by a storage device having a redundant configuration provided with a portable medium,
Even if a failure occurs in two or more storage devices exceeding the redundancy, for example, the failed storage device is closed and the portable medium loaded in the failed storage device is replaced with another sound storage device. In this case, it is possible to continue processing without stopping the system, and the fault-tolerant performance of the RAID system using a portable medium is improved.

Further, since the portable medium of the failed drive is moved to another drive to continue the use, the data can be compared with a case where the data of the failed drive is uniformly restored to a new medium. As a result, the time required for failure recovery is reduced, and the availability of the portable medium and the data stored in the portable medium is improved.

When the number of portable media is larger than the number of drives, the operation of continuing the operation until the last drive becomes the last one by changing the portable media is performed, thereby enabling the system to operate. Fault-tolerant performance can be further improved. In this case, by changing the drive in which the transfer of the portable medium is performed sequentially, it is possible to prevent the load caused by the transfer of the portable medium from being concentrated on a specific drive.

Although the invention made by the present inventor has been specifically described based on the embodiments, the present invention is not limited to the above-described embodiments and can be variously modified without departing from the gist thereof. Needless to say, there is.

For example, in the above-described embodiment, the case where the number of the spare storage devices is one is exemplified, but the number may be two or more. Further, although the number of parity data, which is redundant data, is one, two or more redundant data may be generated.

[0046]

According to the information storage system control method of the present invention, it is possible to improve the fault-tolerant performance of an information storage system including a redundant storage device having a portable medium.

Further, according to the information storage system control method of the present invention, it is possible to reduce the time required for failure recovery in an information storage system including a storage device having a redundant configuration with a portable medium. Is obtained.

Further, according to the information storage system control method of the present invention, a portable medium and effective use of data on the portable medium in an information storage system including a redundantly configured storage device having a portable medium are realized. Can be obtained.

Further, according to the information storage system of the present invention, it is possible to improve the fault-tolerant performance of an information storage system including a storage device having a redundant configuration with a portable medium.

Further, according to the information storage system of the present invention, it is possible to reduce the time required for recovery from a failure in an information storage system including a storage device having a redundant configuration with a portable medium. .

Further, according to the information storage system of the present invention, it is possible to realize effective use of the portable medium and the data on the portable medium in the information storage system including the storage device having the redundant configuration with the portable medium. Can be obtained.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing an example of a configuration of an information storage system in which a method for controlling an information storage system according to an embodiment of the present invention is implemented.

FIG. 2 is a conceptual diagram illustrating an example of a control method of an information storage system according to an embodiment of the present invention and control information used in the information storage system.

FIG. 3 is a flowchart illustrating an example of an operation of the information storage system according to the embodiment of the present invention;

FIG. 4 is a state transition diagram illustrating an example of an operation of the information storage system control method and the information storage system according to an embodiment of the present invention.

[Explanation of symbols]

1 ... input data, 2 ... array control unit, 3 ... buffer, 4
... Parity generation / check circuit, 5 ... SCSI control unit,
Reference numeral 6: drive mapping switch unit, 6a: mapping table, 7a to 7e: drive, 8a to 8e: portable medium, 9: spare drive, 10: medium transport unit, 11: SC
SI bus, 12 interface, D1 to D4 data, DP parity.

Claims (5)

[Claims] 1. A method for controlling an information storage system, wherein a plurality of data units and at least one redundant data generated from the data units are distributed and stored in a plurality of storage devices provided with interchangeable portable media. When a failure occurs in any of the storage devices, the operation is continued by moving the portable medium of the failed storage device to another storage device and using it. A method for controlling an information storage system. 2. The method for controlling an information storage system according to claim 1, wherein at least one spare storage device is provided.
When a failure occurs in any of the storage devices, the operation is continued by moving the portable medium of the storage device in which the failure has occurred to the spare storage device. When a failure occurs in the device or the spare storage device, the portable medium of the failed storage device or the spare storage device is replaced with the remaining healthy storage device or the portable medium of the spare storage device. A method for controlling an information storage system, characterized in that an operation of continuing operation while replacing is performed until the number of sound storage devices or spare storage devices becomes the last one. 3. The information storage system control method according to claim 2, wherein the total number of the healthy storage devices and the spare storage devices is smaller than the number of the portable media, and the device operates while replacing the portable media. The method of controlling an information storage system according to claim 1, further comprising: changing the storage device that replaces the portable medium in succession when continuing. 4. Redundant data having a function of generating at least one redundant data from a plurality of data units received from the outside and a function of restoring a lost data unit from the remaining data units and the redundant data. Control means, each comprising a replaceable portable medium, a plurality of storage devices in which the data unit and the redundant data are stored in a distributed manner, and the movement of the portable medium among the plurality of storage devices and A medium transporting means for performing a switching operation; and when a failure occurs in any of the storage devices, the operation is continued by moving the portable medium of the failed storage device to another storage device and using it. An information storage system, comprising: control logic for causing the information storage system to: 5. The information storage system according to claim 4, further comprising a spare storage device in addition to said storage device, wherein said control logic is configured to: The operation is continued by moving the portable medium of the device to the spare storage device, and when a failure has occurred in the storage device or the spare storage device that is equal to or more than the number of the spare storage devices, the failure has occurred. The operation of continuing the operation while replacing the portable medium of the storage device or the spare storage device with the remaining healthy portable device of the storage device or the spare storage device is performed on the sound storage device or the spare storage device. A first operation to be performed until the device becomes the last one, wherein the total number of the sound storage devices and the spare storage devices is less than the number of the portable media, and When continuing the operation while recombinant, information storage system comprising the said friendly replacement second operation of changing the memory device in sequence for performing the portable medium, a function of performing at least one of.