JP3063666B2 - Array disk controller - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an array disk control device for controlling an array disk in which data is distributed and written to a plurality of disks, and more particularly to an array for separating disks when a failure occurs in a disk or a disk interface. It relates to a disk control device.

[0002]

2. Description of the Related Art Conventionally, as disclosed in Japanese Patent Application Laid-Open No. 4-67476, an array disk control for monitoring the occurrence of a read error or seek error of a magnetic disk device and separating a magnetic disk having a medium error. A device has been proposed. In this method, the state of the medium is detected based on the number of occurrences of the error, and the disk is forcibly disconnected when the number of occurrences of the error exceeds a predetermined value.

[0003]

However, the conventional technique disclosed in Japanese Patent Application Laid-Open No. 4-67476 is
In order to know the occurrence frequency of the medium abnormality, it is necessary to store a report about the position of each medium defect, and there is a disadvantage that an extremely large storage capacity is required.

Further, in the conventional technique, even if there is an abnormality in the disk interface control unit, the disk connected to the disk interface control unit is disconnected because it has an abnormality. .

More specifically, in recent years, the quality of a magnetic disk has been improved, and the failure due to a failure in a recording medium is relatively small. Conversely, with the increase in the recording density of the disk, the recording medium of the disk may be attracted to the recording surface of the head, a failure of the spindle motor, or a failure of the interface between the array disk controller and the disk. A failure in which the entire disk becomes inaccessible is more likely to occur than a part of the failure, and the damage caused by the failure is greater because the entire disk becomes inaccessible.

For the failure of the disk itself such as the suction of the head on the recording surface or the failure of the spindle motor, the reliability can be improved by separating the disk by extension of the prior art. However, it is difficult to determine whether a disk interface failure is a failure of the disk itself or a failure of the interface unit of the array disk controller. Disconnecting can have adverse effects.

For example, in an array disk device as shown in FIG. 6 in which each of the disks 71 to 74 can be accessed from two array disk controllers, the disk interface control circuit 42A in the array disk controller 10A fails. In this case, the processing can be continued by disconnecting the disk 52A, but in this case, a normal disk is disconnected.

Further, in the conventional example, when a failure occurs in a partial area of one of a plurality of disks,
If a failure occurs in another disk and the disk is disconnected, there is a disadvantage that data of another portion of the one disk without failure cannot be read. That is, if the disconnection process is simply performed by the number of times of occurrence of a failure, there is a disadvantage that even recoverable data is lost.

First, in a redundant array disk, if writing cannot be performed due to a medium failure, the failed disk is separated in order to ensure consistency between the data disk for storing data and the redundant disk for storing parity data. Must be processed.

However, in the case of reading, since the data can be recovered and transferred to the host computer by using the redundant disk, the processing is continued only by temporary disconnection without disconnecting the entire disk, which has a greater effect. You may try. Alternatively, in a situation where a relatively minor failure has occurred in the medium but the medium can be remedied by retrying, the disk may not be disconnected.

Although such a partial medium failure has occurred, the disk may not be disconnected. If a severe failure occurs in another disk in such a state, if the disk is detached only by counting the frequency of failure as in the prior art, the disk is detached and processing continues. Would be.

When the disk is disconnected in this state, the processing can be continued. However, since the redundancy is lost, the part where a partial medium failure has occurred cannot be read out thereafter. . This is a bigger problem than the inability to read data that has already been written, that is, the loss of data, that is, the fact that data cannot be written and processing is interrupted.

[0013]

SUMMARY OF THE INVENTION The object of the present invention is to provide an array disk control device which can solve the disadvantages of the prior art and, in particular, can accurately separate disks with little failure history information and improve reliability. That is its purpose.

[0014]

Therefore, according to the present invention, there is provided a host interface control unit for receiving data transmitted from a host device, and an array for storing data received by the host interface in a redundant configuration on a plurality of disks. Control means. Further, a disk failure management table in which a predetermined value is stored according to a disk failure is provided in the array control means. The array control means includes a disk failure value addition function for increasing the value of the disk failure management table when a disk failure occurs, and a disk failure value when data is normally stored or read from or to the disk. A disk failure value subtraction function for decreasing the value of the management table, and a function of disconnecting the interface when the disk failure value of the disk failure management table exceeds a predetermined threshold. I am taking it .

The array control means performs a disk disconnection process according to the disk failure state. At this time, the present invention uses a disk failure management table in which a predetermined value is stored according to a disk failure. The array control means increases the value of the disk failure management table when a disk failure occurs by using the disk failure value addition function. On the other hand, in order to prevent disconnection of the disk due to an intermittent failure, the value of the disk failure management table is reduced by the disk failure value subtraction function when data is normally stored or read from the disk. . Then, the array control unit disconnects the interface when the disk failure value in the disk failure management table exceeds a predetermined threshold value by the disk disconnection function . The disk failure value subtraction function, the occurrence of unrecoverable media failure, is reflected in precisely disk failure management table and generation of the intermittent read / write errors. For this reason, there is no need to perform disconnection processing when disconnection is not required due to a minor failure.

[0016] In the present invention, in particular, the array control means is used.
Step further divides the plurality of disks into a plurality of areas.
And the disk failure management table
It has a table item that manages the history of failures for each area,
The disk failure value addition function failed the disk.
Sometimes has a function to add a constant value a,
The threshold value of the management table is set to n times the constant value a.
At the same time, the value of n is
Set the number of retries automatically and reduce the disk failure value.
Calculation function calculates a constant value a according to the number of divided areas of the disk.
With a function to subtract values smaller than
I am taking it. This will try to achieve the above objectives
Things.

[0017]

Embodiments of the present invention will be described below. The array disk control device according to the present embodiment includes a host interface control unit that receives data transmitted from a higher-level device, and an array control unit that stores data received by the host interface in a redundant configuration on a plurality of disks. ing. The array control means is provided with a disk (media) fault management table in which a predetermined value according to a disk fault is stored.

Moreover, the array control means has a disk failure value addition function for increasing the value of the disk failure management table when a disk failure occurs, and data has been normally stored or read from the disk. Sometimes, a disk failure value subtraction function for decreasing the value of the disk failure management table and a disk disconnection function for disconnecting the interface when the disk failure value of the disk failure management table exceeds a predetermined threshold value are provided. I have.

In this configuration, the disk failure value subtraction function subtracts the disk failure value added by the disk failure value addition function when data is normally stored or read from the disk. For this reason, in the case of intermittent disk failure, even if the disk failure value is added by the disk failure value addition function, the disk failure value does not reach the threshold value due to continued normal operation. There is no need to perform disconnection processing when there is no need to disconnect.

Further, in the embodiment , the array control means further divides the plurality of disk devices into a plurality of areas, and the disk fault management table has a table item for managing a fault history for each of the divided areas. . Further, the disk failure value addition function has a function of adding a constant value a when a disk failure occurs, and sets the threshold value of the disk failure management table to n times the constant value a, and Is set to the number of times the higher-level device automatically retries the same data. The disk failure value subtraction function has a function of subtracting a value smaller than the fixed value a according to the number of divided areas of the disk device.

In the present embodiment for dividing the area, the difference between the size of the divided area and the location where a failure occurs in the disk, and the number of times the higher-level device automatically retries the same data are taken into account. If the host device repeats writing the same data three times, and if the writing cannot be repeated three times, the position of the disk where the writing or reading is performed is the same, and the data is written to the sector where the data is to be written. It can be determined that a media failure has occurred. Therefore, if the threshold value is set to n times a, and this n is set to the number of times the upper device retries the same data, if the retry of this higher device fails, the failure always occurs, It is the target of disconnection processing.

On the other hand, when a plurality of different access commands are output from the host device, and when a partially faulty divided area is accessed by a different command, the actual access location differs. In this case, since a normal part of the divided area may be accessed, the disk failure value is subtracted by accessing the normal part. However, the value of the subtraction is set to a value sufficiently smaller than the set value a of the addition, so that a history that a disk has failed is left in the disk failure table. In other words, when the divided area is small, even if it can be determined that the location where the failure has occurred once becomes normal, the divided area is large and there is no faulty part in the divided area. When there is a part, data reading and writing can be performed normally even in a divided area where a failure has occurred. Therefore, the value to be subtracted is set to a value smaller than the fixed value a according to the number of divided areas of the disk device. When the subtraction with this small value continues, that is, when normal read / write is repeated, the value of the disk failure table becomes 0. On the other hand, when the retry instruction is repeated from the host device, a large value a is added, and the threshold value is reached.
As described above, since the relationship between the divided area, the threshold value, and the addition / subtraction value is set, it is possible to accurately perform the disk separation processing.

In another embodiment, when the array control means performs control for disconnecting a disk by the disk disconnection function, if the other disk has a failure according to the disk failure management table, the disconnection control is stopped. It has a disconnection stop function. The disconnection stop function automatically performs the disconnection processing only when a failure has occurred in only one of the plurality of disks. Therefore, even if the disconnection processing is performed, the data is redundantly configured. Can be recovered. On the other hand, even if it is determined that a failure has occurred in one of the disks and the disconnection process is to be performed, if another disk also has a failure, data recovery for the failed portion of the other disk becomes impossible. , Without reporting to the host device. Therefore, it is possible to reliably prevent a situation in which data cannot be read due to the separation processing.

Further, in an array disk control device for monitoring a failure of a device interface connecting the array control means and the disk, a disconnection process is performed according to the failure state of the device interface. In this embodiment, a device interface fault management table in which a predetermined value is stored according to a device interface fault is provided in the disk fault management table. Furthermore, the array control means increases the value of the device interface fault management table when a device interface fault occurs, and the data storage or reading is normally performed via the device interface. An interface failure value subtraction function for decreasing the value of the device interface failure management table and a device interface disconnection function for disconnecting the interface when the interface failure value of the device interface failure management table exceeds a predetermined threshold value And

Even in the case of performing the disconnection processing of the disk according to the state of the device interface, the failure value is subtracted if the data can be normally transferred after the occurrence of the failure. Thus, the separation process can be performed only when the separation process is necessary.

In the embodiment for managing the device interface, if there are two or more array control means and each array control means is connected to the same disk, the disconnection processing is performed, and then the other array control means is used. The disk is accessed by the control means. In other words, in this embodiment, when the device interface is disconnected by the device interface disconnection function, a retry from the higher-level device is performed by sequentially selecting a plurality of array control means. A trial control function is also provided. This makes it possible to automatically perform the disk interface disconnection processing and the recovery processing of connection by other array control means.

FIG. 1 is a block diagram showing a configuration example of this embodiment. As shown in FIG. 1, in the present embodiment, a media failure management table (disk failure management table) 1 for storing a history of detecting a media failure for each connected disk, and a connected disk interface Each device 41 to 44 has a device interface failure management table 2 for storing a history of an interface failure detected and reported to the host computer.

In the example shown in FIG. 1, the microprogram processing unit 5 adds or subtracts a fault value to each of the fault management tables 1 and 2. The microprogram processing unit 5 includes a CPU, a RAM, and the like, and realizes various functions of an array control unit by executing a program describing a procedure for adding or subtracting a fault value.

When a failure in the disk or device interface is detected or reported, the microprogram processing unit 5 adds a preset value to the corresponding media failure management table 1 or device interface failure management table 2.

On the other hand, when no failure occurs even when the disk is used, or when the disk is separated and processed to temporarily avoid the failure, the media failure management table of the disk is used. 1, and a preset value is subtracted from the contents of the device interface fault management table 2.

Further, the microprogram processing unit 5
Before accessing the disk, the contents of the media failure management table 1 and the device interface failure management table 2 are checked. If the contents exceed a specified value, the disk is regarded as a failure and is subject to disconnection. The above-described array control means includes the microprogram processing unit 5 and the array control unit 3.

The microprogram processing unit 5 checks the contents of the media failure management table 1 and the device interface failure management table 2 and, if a disk to be separated is detected, detects the media failure history of another disk. If there is, there is a function to stop disconnecting the disk.

In the embodiment shown in FIG. 1, the information necessary for managing the failure history is one for disc management of the history of disk interface faults per disk, and the recording of one disc for the management of fault history on a medium. All that is needed is information that corresponds to the number of divided areas. For example, if the recording area of one disk is divided into 32, it is sufficient if there are 32 storage areas. Disk interface, or
When a media failure is detected, a preset value is added to the corresponding failure history management table.

When the value of these failure histories reaches a threshold value, the process of disconnecting the disk has been conventionally performed. According to the present invention, when no abnormality occurs in accessing the disk, a preset value is subtracted from the contents of the media failure management table 1 and the device interface failure management table 2 corresponding to the storage area. . As a result, it is possible to prevent the disk from being disconnected due to accumulation of minor failures that may occur intermittently.

When a failure occurs in the disk interface, a failure report is always sent to the host computer. Device interface obstruction filling table 2
Is processed at the time of reporting a failure to the host computer, and can be treated as the number of times the host computer has recognized the occurrence of the failure. In a system in which the same disk can be accessed from a plurality of array disk controllers, the host computer recognizes the occurrence of the failure, and can retry via another array disk controller.

If a failure occurs in the disk interface control circuit in the array disk controller, the processing can be continued using another normal array disk controller without disconnecting the disk by this retry. If the cause of the disk interface failure is on the disk side, it will not recover even if the host computer retries via another array disk controller, and it will be restarted using the array disk controller that first reported the failure. Take back the trial. As a result, when the value of the device interface fault management table 2 is added and exceeds the threshold value, the disk is disconnected.

Further, before the disk is actually separated, the contents of the media failure management table 1 of the other disk are examined before the actual separation, and if a failure history remains, the disk is recovered by disconnecting the disk. Stop disconnection because there is a possibility that unreadable data may occur. This prevents data that cannot be read due to disconnection of the disk from occurring.

[0038]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing one configuration example of the array disk control device of the present invention. The array disk controller has a host interface controller 6 for controlling data transfer between the host computer and the array disk controller, distributing data sent from the host computer to a plurality of disks, and generating redundant data by RAID. Or multiple disks (direct access storage, D
ASD) to combine the original data (as sent from the host computer) from the data read from the
An array control unit 3 for checking consistency with redundant data, restoring data with redundant data, etc .; disks 71, 72, 73, 74 for storing data;
Disk interface controllers 41, 42, 43, 44 for controlling the interface between the array controllers 3, 72, 73, 74 are provided.

Further, in the example shown in FIG. 1, a media fault management table 1 is provided. In the present embodiment, the media fault management table 1 includes the disks 71, 72, 73,
This is a set of 1-byte counters for dividing the recording area of 74 into eight and counting the error occurrence history for each area. Further, the array disk control device includes a device interface fault management table 2. The device interface fault management table 2 is a set of 1-byte counters for counting the error occurrence history for each of the disk interface control units 41, 42, 43, and 44.

The microprogram processing section 5 is a section for controlling the operation of each section such as the host interface control section 6 and the array control section 3. In the present embodiment, the microprogram processing table 1 and the device interface fault management table 2 Updating is also realized by the microprogram by the microprogram processing unit 5.

Next, the operation will be described based on a flowchart. In this example, the added value of the disk failure is (2
0) ₁₆ , the threshold value is (60) ₁₆ , and the subtraction value is (01) ₁₆ . Also, the disc is divided into eight areas. The added value of the failure of the device interface is cleared to (01) ₁₆ , the threshold value to (03), and the subtraction to the initial value (00). The number of retries of the operating system of the host device is three. With this setting, the determination as to whether or not the disconnection process is appropriate is made satisfactorily.

FIG. 2 is a flowchart of a process for receiving a command from the host computer and checking for the presence or absence of a disk to be disconnected.

When the array disk controller receives a command from the host computer via the host interface controller 6 (step S201), the disks 71 to 74 which are accessed from the read or write address specified by the command, and record data. It is determined which area (hereinafter, referred to as a zone) is obtained by dividing the entire recording area of the disc into eight (step S202).

Next, the contents of the counter corresponding to the disk and zone determined in step S202 are read from the media failure management table 1 (step S203). If the data spans a plurality of disks or zones, all the counter values are compared with a preset threshold value (step S204).

Here, it is assumed that (60) ₁₆ is set as the threshold value. If there is no counter value exceeding the threshold value, the failure history counts of the disk interface control units 41 to 44 corresponding to the used disks determined in step S202 are read from the device interface failure management table 2. (Step S205).

If the data spans a plurality of disks, all the counter values are compared with a preset threshold value (step S206). Here, (03) ₁₆ is set as the threshold value. Therefore, when a failure of a device interface is reported three times, the interface is disconnected. If the determination in either of steps S204 and S206 does not exceed the threshold value, a necessary disk is selected and the disk is accessed (step S208).

In step S208, if a necessary disk has already been disconnected, it is also determined whether or not to use a redundant disk so that data can be recovered. It is determined whether the content of the media failure management table 1 does not exceed the threshold value (step S204), and whether the content of the device interface failure management table 2 does not exceed the threshold value (step S204). S20
If a counter exceeding the threshold value is detected in 6), it is determined whether or not the disk can be detached with this disk as an object to be detached (step S207).

If the contents of the media failure management table 1 or the device interface failure management table 2 exceed the threshold value and can be separated, in step S207, the disk is separated and a necessary disk is selected. Then, access such as reading and writing is performed (step S208).

FIG. 3 is a flowchart showing the operation of the media fault management table 1 and the device interface fault management table 2 after accessing the disk. After accessing the disk, it is first determined whether there is a failure related to the disk interface (step S).
301).

When an abnormality is detected in the disk interface, (01) is added to the contents of the counter of the device interface fault management table 2 corresponding to the disk in which the error has been detected (step S302), and the host computer is notified. The status reporting the occurrence of the error is reported (step S303).

If no abnormality is detected in the disk interface, the contents of the device interface fault management table 2 corresponding to all used disks are cleared to (00) (step S304).

If no abnormality is detected in the disk interface, it is checked whether there is an error in the disk (step S305). If no error has occurred in the disk, a status indicating a normal end is reported to the host computer (step S309), and based on the disk to be accessed and the contents of the corresponding media fault management table 1 of the zone ( 01) is reduced (step S310).

In the process of step S310, the value of the corresponding counter of the media fault management table 1 is set to (00).
In the case of, that is, when there is no history of occurrence of a media failure, the process is skipped and nothing is performed.

An error has occurred in the disc in the judgment of step S305.
If over it occurs to determine the zone corresponding to the failed disk, and generating position error (step S3
06). The counter corresponding to the zone determined in step S306 is read from the media failure management table 1, and (20) ₁₆ is added to the counter value (step S30).
7) The status indicating the occurrence of the error is reported to the host computer (step S303).

As in the case where no error has occurred in the disk, the counter of the media fault management table 1 corresponding to the disk or zone to which access has been made and for which no error has occurred, subtracts (01) (step). S
310).

Also in this case, if the value of the corresponding counter in the media failure management table 1 is (00), that is, if there is no history of occurrence of the media failure, the process of step S310 is skipped and nothing is performed.

As shown in FIG. 3, even when a failure occurs due to disk access, the disk is not disconnected at this time only by operating the counter values of the media failure management table 1 and the device interface failure management table 2. .

When disconnecting a disk, an error occurs.
An error is reported to the host computer, and a finite number of retries are performed at the discretion of the host computer. As a result, the count value of the media failure management table 1 or the device interface failure management table 2 increases, and the disk is disconnected. Reach.

Whether the disk can be disconnected when the contents of the media failure management table 1 or the device interface failure management table 2 shown in step S207 of FIG. 3 exceed the threshold value is shown in the flowchart of FIG. If the threshold value is exceeded, first, the current state of the array disk is checked (step S401).

It is assumed that the array disk of this embodiment is composed of three disks and one redundant disk. Therefore, if one disk has already been disconnected, the redundancy has been lost, and it is not possible to disconnect any more disks. Therefore, it is checked whether there is any disk that has already been disconnected (step S401).

If there is a disk that has already been separated, the determination as to whether or not the disk can be separated is terminated without doing anything. When the disconnected disk does not exist and the redundancy can be maintained, the failure history of the media regarding the disks other than the disk to be disconnected is read from the media failure management table 1. The read counter reads all zones (step S402).

The read counter value is checked to determine whether there is a failure history in a disk other than the disk to be disconnected (step S403). If there is something other than (00), the separation is stopped because there is a possibility that data that cannot be read later may occur due to the separation.

If there is no failure history in the disks other than the disk to be separated, the failure history of the disk interface corresponding to the disk other than the disk to be separated is read from the device interface failure management table 2 (step S404).

The failure history is checked in a disk interface corresponding to a disk other than the disk to be separated (step S403). If there is a disk other than (00), data that cannot be read later due to separation may occur. Cancel disconnection because there is. If there is no failure history in the other disk interface control unit, the disk to be separated is separated (step S
406).

Next, a description will be given of a second embodiment in which the same disks 71 to 74 can be accessed from two array disk controllers 10A and 10B as shown in FIG. In this case, the operation of the array disk control device is exactly the same as the above-described processing.

When the host computer retrys,
By alternately retrying the two directors, if the failure is caused by the disk interface control units 41A to 44A in the array disk controller 10A, the retry is performed from the array disk controller 10B side. The output command succeeds and processing can continue without disconnecting the disk.

If the cause of the failure is in the disks 71 to 74, the two array disk controllers 10A, 10A
B, an error occurs through any of the media failure management tables 1 each time an error is reported to the host computer.
Or the contents of the device interface fault management table 2 increase, and the host computer unconditionally repeats the retry, thereby leading to the disconnection of the disk.

As described above, according to this embodiment, the following effects can be obtained.

The first effect is that counters corresponding to the number of disk interface control units and counters of the number obtained by dividing the recording area of each disk for the purpose of media error management can be prepared. Can be managed separately.

The second effect is that even if the cause is on the array disk controller and the cause is on the disk, such as a failure in the disk interface, it is not easy to distinguish the failure, it is possible to accurately disconnect the failed disk. Control can be performed.

In particular, in an array disk device composed of two or more array disk controllers, the retry from the host computer is executed alternately through each array disk controller, so that there are cases where there is a cause on the array disk controller side, The disk side can be properly separated from the cause, and if there is a failure in the array disk controller, processing can be continued without disconnecting the normal disk.If the disk side has the cause, the disk can be disconnected. Continuation of the process can be made possible.

This is based on the retry from the host computer. However, the retry from the host computer may simply be performed by alternately switching the array disk controller that performs the retry. Since there is no necessity, the load of software for using the array disk device does not increase.

The third effect is that, as described in the section of the problem to be solved by the present invention, in an array disk having redundancy, if data cannot be written due to a medium failure, data to be stored is stored. In order to maintain consistency between the disk and the redundant disk that stores parity data, it is necessary to separate the failed disk for processing, but in the case of reading, it is possible to recover the data using the redundant disk and transfer it to the host computer. Because it is possible, there is a case where an attempt is made to continue processing only by temporary detachment without detaching the entire disk, which has a greater effect.

Alternatively, in a situation where a relatively small failure has occurred in the medium but the medium can be remedied by retrying, the disk may not be disconnected in some cases.

Even if such a partial media failure has occurred, but the disk has not been disconnected, the media failure management table or the device interface can be used to determine whether or not the disk should be disconnected. Since the failure management table can be checked and the presence or absence of a failure in another disk can be easily determined, it is possible to avoid data that cannot be read later due to easy disconnection of the disk.

[0077]

According to the present invention, the array control means increases the value of the disk failure management table when a disk failure occurs. When data storage or reading is normally performed, the value of the disk failure management table is reduced, and when the disk failure value of the disk failure management table exceeds a predetermined threshold, the interface is disabled. Due to the disconnection, the occurrence of an unrecoverable medium failure and the occurrence of an intermittent read / write error can be accurately reflected in the disk failure management table. Without performing disconnection processing,
By setting the addition value, the subtraction value, and the threshold value according to the number of retries of the higher-level device and the number of divided areas in the disk management, even if a failure occurs in a part of the disk, the state can be improved. Can be reflected in the disk failure management table.Furthermore, by superimposing and managing the states of the device interfaces, disconnection processing can be performed only when necessary, and when performing disconnection processing,
Use the disk failure management table to check the status of other disks, and if there is a failure in another disk, abort the disconnection process, so that it is possible to satisfactorily avoid the situation where recovery cannot be performed based on the redundant configuration Can be. Thus, a small fault history information, accurately perform disconnection of the disk, Ki is possible to improve the reliability, further, divided
Set the relationship between the area, threshold value, and addition / subtraction value
In this way, it is possible to perform more accurate disk disconnection processing.
Kill, it is possible to provide an excellent array disk controller unprecedented called.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of an array disk control device according to the present invention.

FIG. 2 is a flowchart showing a disconnection determination process performed before a disk access is started by the array disk control device of the present invention.

FIG. 3 is a flowchart showing processing of error history management after disk access of the array disk control device of the present invention.

FIG. 4 is a flowchart showing a disk disconnection determination process in the array disk control device of the present invention.

FIG. 5 is a block diagram showing one configuration example of an array disk using two array disk control devices of the present invention.

FIG. 6 is a block diagram showing a configuration example of an array disk including two control devices according to the related art.

[Explanation of symbols]

1, 1A, 1B Media fault management table 2, 2A, 2B Device interface fault management table 3, 3A, 3B Array control unit 5, 5A, 5B Micro program processing unit 6, 6A, 6B Host interface control unit 10, 10A , 10B Array disk controller 41, 42, 43, 44 Disk interface controller 41A, 42A, 43A, 44A Disk interface controller 41B, 42B, 43B, 44B Disk interface controller 71, 72, 73, 74 Disk

Claims (4)

(57) [Claims] 1. A method for receiving data transmitted from a host device.
Host interface control unit and the host interface
The data received by the
Array control means to store data in the disk
Disk failure management in which predetermined values are stored according to
An array disk control device comprising a table and the
When increasing the value of the disk failure management table
Disk failure value addition function and normal for the disk
When data is stored or read in the
Disk failure that decreases the value in the disk failure management table
Value subtraction function and the disk failure management table
When the fault value exceeds a predetermined threshold
Disk detach function that detaches the interface in question
The array control means further divides the plurality of disks into a plurality of areas, and the disk failure management table has a table item for managing a failure history for each of the divided areas, The disk failure value addition function has a function of adding a constant value a when a disk failure occurs. The threshold value of the disk failure management table is set to n times the constant value a and this n Is set to the number of times that the higher-level device automatically retries the same data. An array disk control device, comprising: 2. A device interface failure management table in which a predetermined value is stored in accordance with a device interface failure in the disk failure management table, wherein the array control means generates a failure of the device interface. An interface failure value addition function that increases the value of the device interface failure management table when the data is stored, and decreases the value of the device interface failure management table when data is normally stored or read through the device interface. An interface failure value subtraction function to cause the device interface failure management table to disconnect when the interface failure value of the device interface failure management table exceeds a predetermined threshold value. Array disk controller according to claim 1, wherein the was e. 3. A host interface control unit for receiving data transmitted from a higher-level device, and array control means for storing data received by the host interface control unit in a redundant configuration and storing the data on a plurality of disks via a device interface. An array disk control device comprising: a device interface fault management table in which a predetermined value is stored in accordance with a fault in the device interface in the array control means; An interface failure value addition function for increasing the value of the device interface failure management table when an interface failure occurs, and the device interface when data is normally stored or read through the device interface. An interface failure value subtraction function for reducing the value of the failure management table and a device interface separation function for disconnecting the interface when the interface failure value of the device interface failure management table exceeds a predetermined threshold value. An array disk control device, characterized in that: 4. A host interface control unit for receiving data transmitted from a higher-level device, and a plurality of disks connected to a plurality of disks for storing data received by the host interface control unit and connected to the plurality of disks via a device interface. An array disk control device comprising a plurality of array control means for storing, wherein said array control means is provided with a device interface fault management table in which a predetermined value is stored in accordance with a fault of said device interface, Each of the array control means has a table value adding function for increasing the value of the device interface fault management table when a fault occurs in the device interface, and data has been normally stored or read through the device interface. A table value subtraction function for decreasing the value of the device interface failure management table, and a device interface separation function for disconnecting the interface when the value of the device interface failure management table exceeds a predetermined threshold. When the device interface is disconnected by the device interface disconnection function, the device interface disconnection function is controlled to sequentially select the plurality of array control means to retry from the higher-level device. An array disk control device having a retry control function.