JP2910847B2 - Array type 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 input / output control device for controlling an input / output device based on a command from a higher-level device. Monitoring control including control of a backup input / output device. With the recent demand for higher reliability in computer systems, higher reliability is required for input / output subsystems. For this reason, an input / output subsystem having a spare input / output device has been provided. In this input / output subsystem, the spare input / output device is not used in normal times, but is used only when an abnormality such as a reading error or a failure occurs in another current input / output device.

At this time, in order for the spare I / O device to operate normally and fulfill the role of the spare I / O device, the normal / abnormal state of the spare I / O device is monitored in advance, and hardware errors and read errors are monitored. It is necessary for the input / output subsystem to have a means for providing preventive maintenance information such as the above to the host device.

[0003]

2. Description of the Related Art As an input / output control device for providing preventive maintenance information to a host device using a spare input / output device instead of a faulty input / output device, an array type magnetic disk control device, which is an array type storage device, has conventionally been used. Are known. This array type magnetic disk controller has an interface controller for an upper device, an interface controller for a lower device, a processor, and various registers. As the lower device, a plurality of data disk devices and one parity disk device are provided. And one spare disk device are connected and synchronously controlled.

The array type magnetic disk controller reads data and parity from each of a plurality of data disk devices and parity disk devices in parallel, and checks for data abnormality. Normally, when there is no abnormality in the data, the data is transferred to the host device as it is. However, when it is detected that the data has an abnormality, the array-type magnetic disk drive reads data from the remaining normal data disk device and the parity disk device other than the one disk device that has read the abnormal data. Based on the data and parity, correct data is restored and transferred to the host device.

If an error such as a reading error or a failure occurs for a certain number of times in one of the data disk devices or the parity disk device which is the current input / output device, the remaining normal data disk is not restored. From the device and parity disk unit, all data written on the abnormal data disk unit or abnormal parity disk unit is automatically restored to the spare disk unit. By using it in place of a disk device, required processing operations can be continued,
System down can be prevented.

[0006]

However, even if there is a hard error or a read error in the spare disk device itself, the error is not known to the upper-level device at all, and the spare disk device is replaced by the spare disk device for the first time. When using as an input / output device, a hardware error, a reading error, or the like of the spare disk device itself may be found, and the spare disk device may not function as a spare disk device.

Therefore, in a conventional apparatus having a normal system including a plurality of units and a spare unit of the normal system unit, the same input signal as that of an arbitrary unit of the normal system is given to the spare unit, and the spare unit and the above optional unit are supplied. A method of monitoring a spare unit by comparing both output signals of the two units is known (JP-A-56-7235).
No. 9). However, this monitoring method requires that the standby unit always be operated at the same time as the normal working unit. Therefore, the input / output control device described above is equivalent to substantially increasing the number of input / output devices to be controlled by one. The load on the device is large.

Further, a method of monitoring an abnormality of the input / output control device according to a time monitoring time table (Japanese Patent Laid-Open No. 62-21)
No. 2856) is also known, but this does not monitor an abnormality in the spare input / output device, nor does it use a process idle time. Further, in a conventional system in which a central processing unit (CPU) is configured in a duplex configuration for working and standby, a monitoring system in which a standby CPU runs software and diagnoses in a manner similar to actual operation (Japanese Patent Laid-Open No. 2-93953). Publication) and a standby CPU
Monitoring system for diagnosing the data at regular intervals (Japanese Patent Laid-Open No. 62-90)
No. 068) has been conventionally known. But,
None of the above-mentioned monitoring methods of the apparatus having a duplicated CPU monitor the spare CPU using the idle time of the processing.

SUMMARY OF THE INVENTION The present invention has been made in view of the above, and when monitoring the status of a spare input / output device and / or a current input / output device, a process that does not execute a command from a host device. Even when the spare disk device (input / output device) is not used, the power is always turned on and the spare disk device is always ready, and the spare disk device is included. An object of the present invention is to provide an array-type disk control device capable of determining abnormality of each device.

[0010]

According to the first aspect of the present invention, there are provided a plurality of disk devices (30) for storing data transferred to and from a higher-level device and redundant data created from the data, respectively. And a lower-level device interface controller (26) to which a logical device including a spare disk device (31) is connected. The lower-level device interface controller connects to a higher-level device (10). And the plurality of disk devices (30) and the spare disk device (3) issued from the higher-level device (10).
A channel interface controller (21) for receiving an operation command to 1), and read / write operations of a plurality of disk devices for storing the data based on a command input from a higher-level device via the channel interface controller. When the plurality of disk devices included in the logical device storing the data are normal, the plurality of disk devices storing the data are controlled and included in the logical device storing the data. An operation control unit that controls a normal disk unit and a spare disk unit when an error is found in any of the plurality of disk units; and an operation control unit that controls the plurality of disk units and the spare disk unit in the logical device. Power supply means for supplying power to the logical device. The device, always be held in the execution state capable of read and write commands and seek command is an array type disk control apparatus according to claim.

According to the above configuration, not only is it possible to supply power to all the disk devices, but also it is possible to execute a ready state in which the disk device can operate immediately in response to an access from a host device, that is, to execute a read / write command and a seek command. By keeping it in a consistent state,
That is, it is possible to realize an array-type disk controller that always keeps the spare disk device in a normal state. In this regard,
This is different from the known hot start method in which the standby apparatus is always supplied with power only and the standby apparatus is not kept ready. That is, the standby device cannot operate simply by being energized, but must be in a ready state to operate normally. The present invention makes it possible to judge these abnormalities by putting each device including the spare disk device in a ready state.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in more detail with reference to the accompanying drawings. FIG. 1 shows a principle configuration diagram of the present invention. In FIG. 1, the input / output control device 20 includes a higher-level device interface control unit 21, an inspection unit 22, a control unit 23, a determination unit 24, a preventive maintenance information storage register 25, a lower-level device interface control unit 26, and a patrol timer register 27. Have. The host device interface control unit 21 receives a command issued from the host device 10. The lower-level device interface control unit 26 is connected to a plurality of input / output devices 30 and one spare input / output device 31, respectively. The checking means 22 reads the data and the error detection code from the input / output device 30 in parallel, performs an error check using the read data and the error detection code, and checks the normality of the data of the input / output device 30.

The control means 23 selects only the input / output device 30 when the inspection result of the inspection means 22 indicates that the read data of all the input / output devices 30 is normal, and According to the instruction, the input / output control device 3
0 operation is controlled. Therefore, at this time, the spare input / output device 31 is not used. On the other hand, if the inspection result of the inspection means 22 indicates that the read data of one input / output device among the plurality of input / output devices 30 has an abnormality, and this number exceeds a specified number, the lower-level device interface control unit 26 has generated an abnormality. From the data read from the remaining normal input / output devices other than the input / output device and the error detection code, all data written to the abnormal input / output device is restored, and the restored data is transferred to the spare input / output device 31. Write. Therefore, the spare input / output device 31 is used as a working input / output device instead of the abnormal input / output device.

The judging means 24 inputs, via the lower-level device interface control unit 26, an idle time during which the control unit 23 does not execute the processing of the command input from the higher-level device 10 via the higher-level device interface control unit 21. Output device and spare I / O device 31 or spare I / O device 3
1 is selected, and it is determined whether or not the selected I / O device 30 and backup I / O device 31 or backup I / O device 31 operate normally.

When it is determined by the determining means 24 that an abnormality has occurred, information relating to the abnormality is stored in the preventive maintenance information storage register 25. This is to notify the host device 10 later as preventive maintenance information. The idle time for which the determination means 24 makes a determination is determined from the count value of the patrol timer register 27. The above determination means 24
Is determined to be abnormal when a signal sent from the spare input / output device 31 after writing data or seeking operation to the spare input / output device 31 indicates a hard error. Further, the determination means 24 includes a spare input / output device 31.
Is read, and if the read data has a read error, it is determined to be abnormal.

According to the input / output control device 20 having such a configuration, even when the standby input / output device 31 is not used, it is possible to detect the presence or absence of the abnormality of the standby input / output device 31 by utilizing the idle time of processing. . Further, when there is an abnormality in the spare input / output device 31, the preventive maintenance information from the preventive maintenance information storage register 25 is notified to the host device 10. A hardware error or a reading error of the device 31 itself can be confirmed.

According to the input / output control device 20, not only the spare input / output device 31 but also the presence / absence of an abnormality of the input / output device 30 can be detected. In this case, the determination unit 24 performs a seek operation on the input / output
Is determined to be abnormal when a hard error signal is input from the CPU. The output control unit 20 in FIG. 2, for example 20 _1, 2
0 ₂ provided two sets as shown, to share the high-level equipment 10, and may share a plurality of logical devices 40 ₀ to 40 _N. Here, consisting of the logical device 40 _0-40 each of _N and a plurality of input-output device 30 one spare output device 31. of.

According to this configuration, the input / output control device 20
₁ and 20 _2, respectively different logical devices on the basis of the instructions of the from the host device 10 can be simultaneously selected and controlled. Next, one embodiment of FIG. 3 will be described. In the figure, an array type magnetic disk controller 60 corresponds to the input / output controller 20. The array type magnetic disk controller 60 includes a channel interface controller 61 corresponding to the upper device interface controller 21, an operation controller 62a, a determination controller 62b, a lower device interface controller 63 corresponding to the lower device interface controller 26, and a prevention device. Maintenance information storage register 64, patrol timer register 65
And a rotation synchronization control unit 66.

The operation control unit 62a is provided with the inspection unit 2 described above.
2. A circuit part constituting the control means 23, and a judgment control part 62
Reference numeral b denotes a circuit unit constituting the above-described determination means 24, each of which includes a processor 621, a control storage 622,
System reset timer register 623, operating time monitoring timer register 624, seek address register 62
5, a logical device address register 626, an error detection flag register 627, and a data transfer control unit 628.

The processor 621 controls the entire array type magnetic disk controller 60 of the microprogram control system. The control storage 622 stores the instruction code of the processor 621 and is also used as a firmware control table. The data transfer control unit 628 controls data transfer between the channel interface control unit 61 and the lower device interface control unit 63, generates a parity of the transfer data, and transfers the parity to the lower device interface control unit 63.

The lower-level device interface controller 63 outputs $ 0
To # 9 disk control units 630 to 639. The channel interface control section 61 has a control register group, and performs bidirectional data communication with the host device 50. The host device 50 is a central processing unit (CPU)
It is composed of Further, the rotation synchronization control unit 66 synchronously rotates the magnetic disks of ten magnetic disk devices in the logical devices 70 ₀ to 70 _N to be described later.

On the other hand, the disk controllers 630 to 639
+1 logical device 70₀~ 70 _NConnected to each other
I have. Logical device 70₀~ 70_NHave the same configuration,
It comprises ten magnetic disk devices 700 to 709. This
Of these, the magnetic disk devices 700 to 707
Data is written to the magnetic disk in advance
And can be written and read arbitrarily.
ing.

A magnetic disk device 708 is a disk device for parity, and is a magnetic disk for recording and reproducing (hereinafter, referred to as a magnetic disk).
The parity bits of the data written to the data disk devices 700 to 707 are written in a parity disk. Further, the magnetic disk device 709 is a spare disk device, and data or parity is not written in a magnetic disk for recording / reproducing (hereinafter referred to as a spare disk) at the initial stage. Logical device 70 ₀
To 70 _N is controlled by the array-type magnetic disk controller 60 only logical device one at the same time.

The array type magnetic disk drive shown in FIG.
Two control devices 60 are provided in parallel as shown in FIG.
And the logical device 70₀
~ 70 _NAre connected so that each can be controlled. this
The array-type magnetic disk control device 60 having such a configuration is
8 bytes of data based on the instruction from the position device 50
At the same time, write to data disk units 700 to 707
And the parity of the data is
Write to and read them from disk device 708
Data transfer speed to one magnetic disk
The feature is that it can be operated at a higher speed than when a disk device is used.
There is a head.

Further, when it is found that a read error or a failure has occurred in any of the nine disk devices 700 to 708 during execution of an instruction from the host device 50, the disk device having the failure is identified. The contents are automatically restored in the magnetic disk subsystem from the data of the remaining normal eight disk devices, and the spare disk device 7 is restored.
Since the restoration data can be written in the data 09, there is a feature that the reliability is high.

For example, the data written on the eight magnetic disks are D _{0 to} D ₇ , the parity written on the parity disk of one parity disk unit 708 is P _0, and the data disk unit If abnormalities in 702 of the magnetic disk occurs, the data D ₂ unreadable, other data D ₀ which can read normally, D
₁ , D ₃ -D ₇ and parity P ₀ . Then, when an abnormality occurs more than the specified number of times in the data disk device 702, all the data written in the data disk device 702 are automatically extracted from the remaining eight normal disk devices in the magnetic disk subsystem. And write it to the spare disk.
A spare disk device 709 is used in place of 02.

In the input / output subsystem using the array type magnetic disk controller 60, the operation of each embodiment mainly performed when an instruction from the host device 50 is not executed will be described. First, the operation of the first embodiment of the present invention will be described with reference to FIG. In this embodiment, the system reset timer register 623, the operating time monitoring timer register 624, the seek address register 625,
The logical device address register 626 and the error detection flag register 627 are not used, and the spare disk device 7
Only the patrol operation of 09 is executed. In FIG.
First, the processor 621 determines whether there is an instruction from the host device 50 (step 81). Here, when the host device 50 including the central processing unit (CPU) issues a data write / read command to one logical device via the array type magnetic disk controller 60, the array type magnetic disk controller 60 Disk control units 630-6
Data and parity are simultaneously written and read to and from nine magnetic disks of the disk devices 700 to 708 via the disk 39.

That is, the processor 621 checks whether or not there is an instruction from the host device 50 every time one instruction is executed as shown in FIG. 4 (step 81). It is determined whether the command is for the logical device that has detected the error (step 82). Normally, this is not the case, so the command from the host device 50 is executed (step 83), and the process returns to step 81.

Here, assuming that the data is, for example, 8 bytes, at the time of writing, 1-byte data is transferred to the data disk device 7 via the disk control units 630 to 637.
The parity is output in parallel to 00 to 707, and the parity generated by the data transfer control unit 628 from the 8-byte data is output to the parity disk device 708 via the disk control unit 638. At the time of reading, the data and parity reproduced by the disk devices 700 to 708 are transferred to the data control transfer unit 6 via the disk control units 630 to 638.
28.

On the other hand, when it is determined in step 81 that there is no command from the host device 50, it is determined whether or not an interrupt process is necessary for the host device 50 (step 84). For example, for a command chain, it is determined whether or not an interrupt process such as an interrupt process for the host device 50 based on a device end notification from the device after the seek is completed is necessary. If an interrupt is required, an interrupt process is executed for the host device 50 (step 85). If an interrupt is not required, reading (step 86), incrementing (step 87), and storing (step 88) of the patrol timer register 65 are sequentially performed to determine whether the value of the patrol timer register 65 after the increment exceeds a predetermined value. It is determined whether or not the value has exceeded the value corresponding to the specified time (step 89). If it is determined in step 89 that the value does not exceed the predetermined value, the process returns to step 81 again.

On the other hand, if it is determined in step 89 that the value exceeds the predetermined value, the flow advances to step 90 to execute a patrol operation on the spare disk device 709 (step 90). Specifically, the processor 621 generates data of a specific pattern predetermined by itself, inputs the data to the disk control unit 639 via the data transfer control unit 628, and converts the error detection code into data of the specific pattern. It is generated on the basis of the above and added to the data of the specific pattern. The data of the specific pattern and the error detection code are stored in the spare disk device 639 in the logical device designated by the logical device address register 626.
, And read it out.

The spare disk unit 639 sends a hard error signal to the disk control unit 639 when the above patrol operation (read / write operation) is not performed normally. The data and error detection code read and output from the spare disk by the read operation are input to the data transfer control unit 628 via the disk control unit 639. The processor 621 determines whether or not the patrol operation has been normally completed based on whether or not the hard error signal has been input and whether or not there has been a read error of the read data or the like (step 91). If there is no error, the processor 621 determines that the operation has been completed normally, and resets an error detection flag for the patrol operation in the control storage 622 (step 92).

On the other hand, if it is determined in step 91 that an error has occurred, the processor 621 collects detailed information of a hard error such as a failure of the seek to end normally, or an error state such as a reading error, and stores the information for storing preventive maintenance information. The data is stored in the register 64 (step 93), and an error detection flag for the patrol operation in the control storage 622 is set (step 94).

When the processing in step 92 or 94 is completed, the flow returns to step 81. Thereafter, in the same manner as described above, during the processing idle time when there is no command from the host device 50 (step 8).
1) At a specified time determined by the patrol timer register 65 (steps 86 to 89), the data of the specific pattern and the error detection for one track of the spare disk of the spare disk unit 709 of each logical device are detected. Code writing and reading, error detection, and the like are performed (step 90).

The logical device 70 When the above writing and reading of _N to the first track is finished, then the logic device 70 ₀ the specific pattern for the next track of the spare disk of the spare disk device 709 The writing and reading of the data and the error detection code, the error detection, and the like are performed. Hereinafter, the same operation as described above is repeated.

Thereafter, when an instruction is issued from the higher-level device 50 to the logical device to which the spare disk in which the error has been detected by the patrol operation belongs (step 8).
1, 82), the processor 621 reads the preventive maintenance information stored from the preventive maintenance information storage register 64, sends it out to the upper-level device 50 via the channel interface control unit 61, and outputs the error of the spare disk device. After the information is notified (step 95), the process returns to step 81.

Thus, according to the present embodiment, if there is a hard error, a reading error or the like in the spare disk device 709 itself, the array type magnetic disk controller 6
Based on the preventive maintenance information notified from 0, repair or replacement of the spare disk unit 709 itself for a hard error, a reading error, or the like can be performed in advance. This allows the spare disk device 709 to be used immediately when the spare disk device 709 is used as a substitute for the disk device in which an abnormality has occurred.

Next, the operation of the second embodiment of the present invention will be described.
This will be described with reference to the flowcharts of FIGS. In the present embodiment, the patrol operation is performed not only on the spare disk device 709 but also on the data disk devices 700 to 707 and the parity disk device 708 which are the current input / output devices. .

In FIG. 5, when power is first supplied to the array type magnetic disk controller 60, step 1 in FIG.
At 01, a power-on sequence is executed. By this power-on sequence, the various registers 64, 6
5, 623 to 627 are initialized. Also,
By turning on the power, the logical devices 70 ₀ to 70 _N are turned on, and the spare disk unit 709 is also turned on similarly to the other disk units 700 to 708.

Subsequently, the processor 621 is connected to the host device 50.
(Step 10)
2). Here, when the host device 50 including the central processing unit (CPU) issues a data write / read command to one logical device via the array type magnetic disk controller 60, the array type magnetic disk controller 60 The disk device 70 via the disk control units 630-639
Data and parity are simultaneously written and read from nine magnetic disks 0 to 708.

That is, the processor 621 checks whether or not there is an instruction from the host device 50 every time one instruction is executed as shown in FIG. 5 (step 102). It is determined whether the command is for the logical device that has detected the error (step 103). Normally, this is not the case, so that an instruction from the host device 50 is executed (step 104), and the operation time monitoring timer register 624 is initialized (step 10).
5) Return to step 102.

On the other hand, if it is determined in step 102 that there is no command from the host device 50, it is determined whether or not an interrupt process such as a control unit end (CU End) or a device end is required for the host device 50. (Step 106). If an interrupt is required, host device 50
(Step 107). If no interrupt is required, patrol timer register 65
(Step 108), the system reset timer register 623 is incremented (step 1).
09) is sequentially performed to determine whether the value of the system reset timer register 623 has exceeded a predetermined value corresponding to 10 minutes (step 110), and whether the value of the patrol timer register 65 has exceeded the predetermined value (that is, the specified value). (Step 111) is sequentially determined. If neither of Steps 110 and 111 exceeds the predetermined value, Step 10 is performed again.
Return to 2. One second determined by the patrol timer register 65 is a logical device selection cycle, as described later, and is a time determined by experience so as not to hinder the instruction processing from the host device 50. is there.

On the other hand, if it is determined that the value exceeds the predetermined value in any of the determinations in steps 110 and 111, the process proceeds to step 112, in which the setting of the hardware switches provided on the maintenance panel or the Is determined whether or not the patrol operation is prohibited. Step 1 if patrol is prohibited
02, on the other hand, in the normal case where the patrol operation is not prohibited, the mode for returning the control unit busy (Cu Busy) is set for the command from the host device 50 (step 113). The contents of the logical device address register 626 are read (step 114), and the logical device is selected (step 115).

Subsequently, it is determined whether the logical device is usable (step 116). This is because, even if the power of the logical device is not turned on, the logical device is selected in step 115, and the processor 621 checks the status from the logical device to check whether the logical device is usable. is there. If the logical device is available, it is determined whether the logical device is being used by another system (step 11).
7). That is, when two or more array-type magnetic disk controllers 60 are used simultaneously as shown in FIG. 2, the logical device is already in use by another array-type magnetic disk controller. This is because sometimes it cannot be used.

When the logical device is not in a usable state (step 116) or when the logical device is being used by another system (step 117), the selected state of the logical device is released (step 118) and patrol is performed. The operation execution logical device address register 626 is incremented and stored (step 119). That is, the selection of the logical device is abandoned, and the next logical device is selected. Thereafter, the patrol timer register 65 and the operation time monitoring timer register 62
4 are initialized (Step 120), and Step 10
Return to 2.

On the other hand, when the logical device is in a usable state and is not being used by another system (step 11).
6, 117) proceeds to step 121 in FIG. 6, and proceeds to the physical device (ie, the data disk devices 700 to 707 and the parity disk device 708) in the logical device.
Then, the state of the spare disk device 709) is read (step 121).

Subsequently, it is determined whether a command has been executed for the logical device within 20 seconds (step 122). Whether or not “20 seconds” is determined based on whether or not the value of the operation time monitoring timer register 624 has reached a value (number of times) corresponding to 20 seconds. The value of “20 seconds” itself is an empirical value, and is not necessarily limited to this time.

If no command is executed within 20 seconds, the patrol operation mode is set to the logical device mode (step 123). This logical device mode uses the current data disk devices 700 to 707, the parity disk device 708, and the spare disk device 709. When the mode is switched to this mode, the seek address register 625 for executing the patrol operation is incremented next (step 124), and then the respective heads of the above logical devices (that is, the disk devices 700 to 708) are once fixed from the current position. A seek operation is performed to move the distance and return to the original position again (step 12).
5). This seek operation is a patrol seek operation, and if the head is kept at a fixed position for a long time, dust adheres to the head and adversely affects the recording / reproducing operation.
This is performed to prevent dust from adhering to the head.

If the patrol seek operation does not end normally, a hard error signal is sent from the disk device that does not end normally to the corresponding disk control unit. Therefore, the processor 621 detects whether an error has occurred from the presence or absence of the input of the hard error signal (step 12).
6) When an error is detected, the process of step 138 described below is executed. When no error is detected, the process of step 140 described below is executed.

On the other hand, if it is determined in step 122 that the command has been executed within 20 seconds, it is checked whether the spare disk unit 709 is in an operable state (standby state) (step 127), and the patrol operation mode is set. Is set to the physical device mode (step 128). This physical device mode is a mode in which only the spare disk device 709 is used. When the mode is switched to this mode, the seek address register 625 for executing the patrol operation is incremented next (step 129), and then the spare disk device 7
A seek operation is executed for the step 09 (step 130).

This seek operation is a seek operation that is performed in advance at the time of writing, and the seek operation in which the spare disk device 709 of the present embodiment moves the head by a predetermined distance prior to each of writing and reading. This is because it is structurally necessary. However, this seek operation prevents dust from adhering to the head as in the case of the patrol seek. If the seek operation has not been completed normally, the spare disk device
A hard error signal is sent to 39. Therefore, in the next step 131, when this hard error signal is input, it is determined that an error has occurred, and when the hard error signal is not input, it is determined that it is normal.

If the data is normal, the write operation of the data of the specific pattern and the error detection code is executed for the spare disk device 709 (step 132). When the above-mentioned write operation is not performed normally, the spare disk device 639 sends a hard error signal to the disk control unit 639 similarly to the seek operation.
21 checks whether or not an error has occurred based on the presence or absence of the input of the hard error signal (step 133). continue,
If there is no error, a seek operation prior to the read operation is executed for the spare disk device 134 (step 13).
4), 1 where the above error detection code and data are written
After moving the head to the head position of the track, it is checked whether or not an error has occurred based on the presence or absence of a hard error signal (step 135). If there is no error, a read operation is executed (step 136), and the above data and error are detected. Play the detection code.

If there is a hard error at the end of the read operation, a hard error signal is supplied to the disk controller 639. The data and the error detection code read out from the spare disk by the read operation are input to the data transfer control unit 628 via the disk control unit 639. The processor 621 determines whether or not the hard error signal has been input and whether or not there has been a read error (step 137). If there is no error, the processor 621 determines that the processing has been completed normally, and resets the error detection flag of the error detection flag register 627 (step 140).

On the other hand, steps 126, 131, 133,
If it is determined at 135 or 137 that an error has occurred, the processor 621 collects detailed information on an error state such as a hard error such as a case where the seek does not end normally or a reading error, and stores the preventive maintenance information storage register 64.
(Step 138), and sets the error detection flag of the error detection flag register 627 (step 1).
39).

When the processing in step 139 or 140 is completed, or in step 127, the spare disk unit 709
If it is determined that the state is not in a state where patrol operation is possible, the Cu Busy state is released (step 14).
1), after canceling the current logical device selection state in steps 118 to 120 of FIG. 4, the patrol operation execution completion logical device address register 626 is incremented to designate the next logical device, and the patrol timer register 65 And operating time monitoring timer register 1
After the initialization of each of the variables 20, the process returns to step 102.

Thereafter, in the same manner as described above, during the processing idle time when there is no command from the host device 50 (step 102), every prescribed time (normally 1 second) determined by the patrol timer register 65 (step 111), Writing and reading of the data of the specific pattern and the error detection code, error detection, and the like are performed on one track of the spare disk of the spare disk unit 709 of each logical device (steps 129 to 137).

[0057] When the writing and reading to the first track of the logical device 70 _N of the spare disk device 709 is completed, the next following track of the spare disk of logical devices 70 ₀ of the spare disk device 709 On the other hand, writing and reading of the data of the specific pattern and the error detection code, error detection, and the like are performed. Hereinafter, the same operation as described above is repeated.

Here, when a logical device is selected one second later, the track on which data is written and read from the spare disk is not a track adjacent to the previous track but a track separated by a certain number of tracks. That is, the recording surface of the spare disk is divided into a plurality of zones, and the number of increments of the seek address register 625 in step 129 is determined in advance so that data is written to the representative track of each divided zone. Thus, the normality of the spare disk can be determined in a short time.

Thereafter, when an instruction is issued from the higher-level device 50 to the logical device to which the spare disk in which the error has been detected by the patrol operation belongs (step 10).
2, 103), the processor 621 reads the stored preventive maintenance information from the preventive maintenance information storage register 64, sends the readout preventive maintenance information to the upper-level device 50 via the channel interface control unit 61, and outputs the error of the spare disk device. After notifying the information (step 142), the error detection flag by the patrol operation in the error detection flag register 627 is reset (step 143), and step 102
Return to

Thus, according to the present embodiment, not only when the spare disk unit 709 itself has a hard error or a read error, but also when the current data disk units 700 to 707 and the parity disk unit are used. The array-type magnetic disk controller 6 is used even when an error occurs.
Based on the preventive maintenance information notified from 0, repair or replacement for a hardware error, a reading error, or the like can be performed in advance.

The present invention is not limited to the above-described embodiment. For example, the patrol timer register 65
The register 64 for storing preventive maintenance information may not be hardware, and the processor 621 may execute a timer function by firmware, or may store the preventive maintenance information in the control storage 622.

[0062]

As described above, the array-type disk controller according to the present invention is connected to a logical device consisting of a plurality of input / output devices and one spare input / output device. It is suitable for high-speed data transfer because data is written and read from the host device in parallel by controlling at the same time as the input / output device, and it is used in place of one input / output device with an error. The standby I / O device is always monitored when the power is turned on, and its preventive maintenance information is provided to the host device. Therefore, the I / O device can be applied to computer systems that require high reliability and high reliability. It is suitable.

Further, it is possible to provide an array type storage system having the above effects.

[Brief description of the drawings]

FIG. 1 is a diagram showing a basic principle configuration of the present invention.

FIG. 2 illustrates a host device and a plurality of logical devices that are shared.
FIG. 3 is a configuration diagram of one input / output control device.

FIG. 3 is a configuration diagram of an embodiment of an input / output control device according to the present invention.

FIG. 4 is a flowchart for explaining the operation of the first embodiment of the present invention.

FIG. 5 is a flowchart (part 1) for explaining the operation of the second embodiment of the present invention.

FIG. 6 is a flowchart (part 2) for explaining the operation of the second embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 High-order apparatus 20 I / O control apparatus 21 High-order apparatus interface control part 22 Inspection means 23 Control means 24 Judgment means 25 Preventive maintenance information storage register 26 Subordinate equipment interface control part 27 Patrol timer register 30 I / O device 31 Reserved I / O apparatus

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G06F 3/06

Claims (1)

(57) [Claims] 1. A plurality of disk devices (30) each storing data transferred to and from a higher-level device and redundant data created from the data, and one spare disk device (31). A lower-level device interface control unit (26) to which a logical device comprising: is connected, and the lower-level device interface control unit is connected to a higher-level device (10) and is issued from the higher-level device (10). A channel interface controller (21) for receiving operation commands to the plurality of disk devices (30) and the spare disk device (31); and a command input from a higher-level device via the channel interface controller. A logical device that controls read and write operations of a plurality of disk devices that store the data, and that stores the data. If the plurality of disk devices included in the device are normal, the plurality of disk devices that store the data are controlled, and an error is found in one of the plurality of disk devices included in the logical device that stores the data. An operation control unit that controls a normal disk device and a spare disk device in a case where it is performed; and a power supply unit that supplies power to each of the plurality of disk devices and the spare disk device in the logical device. An array-type disk controller, comprising: always keeping all disk devices in the logical device in a state in which a read / write command and a seek command can be executed.