JPH04149719A - Magnetic disk controller - Google Patents

Abstract

PURPOSE:To judge real trouble by storing address information which specifies a written place when fault occurrence information is received from one of two central processors and also storing address information on writing to a magnetic disk device, and comparing the both and performing processing. CONSTITUTION:The address information which specifies the written place is stored in a 1st storage means 260 when the trouble occurrence information is received from one of the two central processors 101 and 102 and the address information on the writing to the magnetic disk device 300 is written in a 2nd storage means 270. Then when the contents of the 1st and 2nd storage means match each other, the contents of the 1st storage means 260 are erased. If an error occurs during the read operation of the magnetic disk device 300, a read error caused by the trouble interruption of the central processor is reported if the address information in said reading matches the contents of the 1st storage means 260, but a normal read error is decided if not. Consequently, the cause of trouble is made clear.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a magnetic disk control device, and particularly to a magnetic disk control device provided between a central processing unit and a magnetic disk device to control the magnetic disk device. 6. Related to Devices, Prior Art 1 Conventionally, in this type of magnetic disk control device, if a failure occurs in the central processing unit during the writing operation of data transferred from the central processing unit to the magnetic disk device, , the content of the transferred data is not guaranteed, so writing to the magnetic disk device is interrupted.

After the central processing unit system has recovered from the failure, if the record that was being written to at the time of the magnetic disk unit failure is accessed, the data written to this record, that is, the contents, is not guaranteed. The data was read and a read error was generated to prevent the data from being used as is for processing by the programmer, and a write operation was being performed when the central processing unit system failed, so a read due to a write interruption occurred. Report an error.

Alternatively, if a failure similar to that described in F occurs, data equal to the number of data that was not transferred is transferred from the magnetic disk control device to the magnetic disk device, and the write operation is continued.

After the central processing unit system has recovered from the failure, if the record that was being written to at the time of the failure is accessed, the data written to this record is not guaranteed. The central processing unit performs correction to prevent missing data from being used as is in the processing of the program 10, and prevents read errors from occurring so that processing to the magnetic disk device can continue to be executed.

[Problem to be solved by the invention]

In the conventional magnetic disk control device described above, for example, in an online system that requires high reliability, a read error occurs during a read operation to the magnetic disk device performed during recovery processing of a failed central processing unit system. When a read error occurs, there is no way to determine whether the read error is simply due to a defect in the medium or a read error due to a failure in the central processing unit system that occurred during a write operation.

In addition, even if a device has a means for determining, if it is connected to multiple central processing units, the functions of the central processing unit will prevent reading errors from occurring due to failures in the central processing unit system that occur during write operations. erroneously reports a read error regardless of the

Even in devices that have a function to remember that there was a failure in the central processing unit during a write operation, the code may be rewritten due to a failure in the central processing unit that occurred during a write operation. Even later, not only will the history of the above failure remain, but even if a read error occurs due to a defective medium in the record addressed at the time of the failure, it will be incorrectly reported as a failure in the central processing unit system. However, it has the disadvantage that it takes a lot of time to determine the true fault.

[Means to solve the problem]

It is connected to multiple central processing units, and when a liI file occurs in the first central processing unit, the write operation from the first central processing unit to the magnetic disk device is interrupted, and when a failure occurs in the first central processing unit, the second central processing unit In a magnetic disk controller, the number of data specified by the second central processing unit is written without interrupting the write operation from the central processing unit to the magnetic disk unit. 2. A first storage means for storing information on a write location for specifying a writing location when failure occurrence information is received from a user; , erasing means for erasing the contents of the first storage means when the contents of the first storage means match the contents of the second storage means; and when a read operation is performed in the magnetic disk device and a read error occurs. , when the address information at the time of reading matches the contents of the first storage means, it is reported as a reading error caused by a failure or interruption of the central processing unit, and when they do not match, it is reported as a normal reading error. means, and a first means for making all the means effective when a failure occurs in the first central processing unit, depending on the type of write operation to the magnetic disk device when the failure occurs in the first and second central processing units; The second central processing unit includes a second means for disabling the first storage means and the determination means among the respective means when a failure occurs in the second central processing unit.

〔Example〕

Examples of the present invention will be described based on the V-plane.

FIG. 1 is a block diagram of a system using a magnetic disk control device according to an embodiment of the present invention.

In FIG. 1, this system is composed of central processing units 101 and 102, a magnetic disk control device 200, and a magnetic disk device 300.

The magnetic disk control device 200 includes a channel interface control section 210 for connecting to the central processing units 101 and 102, and a device interface control section 220 for controlling the connection interface to the magnetic disk device 300.
, a microprocessor unit 230 that controls the overall operation of the magnetic disk control device 200, and an operational
Detects that 0ut) has become inactive○p
o detection circuit 240, a write interruption control circuit 250 that interrupts the write operation of the device interface control unit 220 when the output signal a of the opo detection circuit 240 becomes active, and a magnetic disk device number that interrupted the write operation. It is composed of a first storage section 260 that stores address information for specifying a write location, and a second storage section 270 that stores a disk device number and address information of a write location when a write operation is executed.

Furthermore, the channel interface control unit 210
○PO detection output signal control circuit 2 capable of setting the output signal a to active or inactive when the O detection circuit 240 detects that ○PO becomes inactive
11 and 212.

In the above configuration, the output signal a is output by the OP○ detection output signal control circuit 211 connected to the central processing unit 101.
is activated and connected to the central processing unit 102.
The magnetic disk when the output signal a is operated as inactive by the PO detection output signal control circuit 212 and a failure occurs in the central processing unit 101 or 102 and the ○PO signal in the channel interface signal line becomes inactive. The operation of the control device 200 will be explained.

When the magnetic disk control device 200 receives a write command from the central processing unit 101, the microprocessor unit 230
starts writing the data transferred from the central processing unit 101 to the magnetic disk device 300. If a serious failure or power failure occurs in the central processing unit 101 during this write operation, the OP○ signal in the channel interface signal line becomes inactive. Further, the operation of the central processing unit 102 is also similar.

The OPO detection circuit 240 detects that the OPO signal becomes inactive via the channel interface control unit 210, and if the central processing unit 101 is operating, the OP○ detection output signal control circuit 211 becomes active. Therefore, the output signal a is made active. When the write interruption control circuit 250 detects the pressure signal a, it causes the device interface control section 220 to interrupt the write operation to the magnetic disk device 300.

On the other hand, the microprocessor section 230 outputs the output signals a and O.
The output signal from the PO detection circuit 240 is detected, and the magnetic disk device 3 is activated due to a failure in the central processing unit 101.
It is determined that the write operation for 00 has been interrupted, and the magnetic disk device number at that time and address information specifying the write location are stored in the first storage unit 260.

Also, if it is being operated from the central processing unit 102, ○
Since the PO detection output signal control circuit 212 is inactive, the pressure signal a remains inactive. Then, the write interruption control circuit 250 does not operate, and the device interface control unit 220 requests that the magnetic disk device 3
The write operation for 00 ends without interruption.

On the other hand, the microprocessor unit 230 detects a state where the output signal a is active and the output signal b is inactive,
Although the occurrence of an IF error in the central processing unit 102 is detected, it is determined that the write operation to the magnetic disk device 300 has been completed without interruption, and the first storage of address information that specifies the magnetic disk device number and writing location at that time is performed. 260 is bypassed.

After the failure of the central processing unit 101 is recovered and the system is started up, when a write operation is performed on the magnetic disk device, the magnetic disk control device 200 writes the magnetic disk device number and the address information that specifies the write location. are stored in the second storage unit 2701\. At the end of a series of commands including this write command, the microprocessor unit 230 stores the magnetic disk device numbers of the first storage unit 260 and the second storage unit 270 and address information that specifies the write location in the first storage unit 260. It compares the write location information with the ATS information in the second storage section 270, and if they match, the first storage section 260
Delete the magnetic disk device number and address information that specifies the writing location.

Further, the operation is the same even when the failure of the central processing unit 102 is recovered.

By the way, when a write operation is not executed at the address that specifies the write location mentioned above, and a read operation is performed at the address that specifies the write location of this magnetic disk device number, the address that specifies this write location is previously used. Since the write operation from the central processing unit 101 has been interrupted and the read error check code has not been updated, a read error inevitably occurs during the read operation from the central processing unit code 01 or 102 to the magnetic disk device 300. If the read error is a report to the central processing unit 101, the central processing unit 101 is connected to the magnetic disk control unit W2O0 through the OPO detection power signal control circuit 211 which is in an active state. The current magnetic disk device number and the address information specifying the write location are compared with the magnetic disk device number stored in the first storage unit 260 and the address information specifying the write location.

If this comparison matches, it is reported to the central processing unit 101 that the read error occurred because the write operation was interrupted due to a failure in the central processing unit 101, and if this comparison does not match, Magnetic disk device 30
It is determined that the error is caused by a medium failure of 0, and the read error is reported to the central processing unit 101.

If the read error is a report to the central processing unit 102, the central processing unit ff 102 transmits the information to the OPO test output signal control circuit 212, which is in an inactive state, to the magnetic disk control unit F202. Therefore, even if a read error occurs, the microprocessor unit 230 only reports the read error without performing the comparison operation described in Section 4, which would be performed if it were connected to the central processing unit 101.

In this way, when a failure occurs in the central processing unit 101 and the writing operation to the magnetic data storage device 300 is interrupted, the magnetic disk device number and the Attos number that specifies the writing location are stored, and this memory is used. Contents and central processing unit] 0
Compare the magnetic disk device number during the write operation after recovery from failure 1 or 102 with the address information of the write location, and when the comparison information matches, erase the address information that specifies the write location in the first storage unit 260. By doing this, the magnetic disk number and read attos information of the read error that occurred after the failure recovery in the central processing unit 101 are converted into the magnetic disk device number and address information that specifies the write location at the time when the write operation was interrupted. When the above comparison results match, the central processing eliminates the misrecognition that the read error is due to a write interruption even though it is caused by a media defect. Device 101 /\Report.

Furthermore, in the case of a co-selling error that occurs in the central processing unit 102 after recovery from a failure, the central processing unit 102 eliminates a false report that it is a read error due to a writing interruption in the central processing unit π system.

〔Effect of the invention〕

As explained above, according to the present invention, one central processing unit is connected to a plurality of central processing units that require different processing when a central processing unit fails or after recovery. When writing to the magnetic disk device is interrupted due to a device failure, it stores the attribution information that specifies the writing location and connects it. When rewriting a record, by erasing the address information that specifies the interrupted writing location, 1.
When comparing the address information that identifies the read location of a read error that occurred after failure recovery of two central processing units with the address information that identifies the write location when the write operation is not interrupted, there is an error that the read error is due to a media defect. It has the effect of eliminating recognition.

In addition, in the case of a read error report to one other central processing unit, this read error is not caused by a failure of this other central processing unit, and the OP○ with the write interruption function enabled Since this occurred due to a write interruption due to a failure in one central processing unit connected to the detection output signal control circuit, false reporting is eliminated by bypassing the reporting of a read error due to a writing interruption to the other central processing unit. It has the effect of

Furthermore, when the comparison results show a match, it is possible to determine whether this read error is due to a simple media defect by reporting to one central processing unit that this read error is due to a write interruption. Since it is possible to easily determine whether the cause of the problem is caused, the effect is that the recovery time can be shortened.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a system using a magnetic disk control device according to an embodiment of the present invention. 101, 102...Central processing unit, 200...Magnetic disk control device, 210...Channel interface control unit, 211, 212...OPO test output signal control circuit 20...Device interface control Department,
230...Microprocessor section, 240...OP
O detection circuit, 250...Writing interrupt circuit, 260...
- First storage section, 270...Second storage section, 300...
Magnetic disk device.

Claims (1)

[Scope of Claims] The second central processing unit is connected to a plurality of central processing units, and when a failure occurs in the first central processing unit, the writing operation from the first central processing unit to the magnetic disk device is interrupted; In a magnetic disk control device that writes as much data as instructed by the second central processing unit without interrupting the write operation from the second central processing unit to the magnetic disk device when a failure occurs in the magnetic disk device, When failure occurrence information is received from one of the first and second central processing units, a first storage means stores address information for specifying a write location, and a second storage means stores address information for executing writing in the magnetic disk device. a second storage means; an erasing means for erasing the contents of the first storage means when the contents of the first storage means and the contents of the second storage means match; and a read operation in the magnetic disk device. When a reading error occurs during reading, if the address information at the time of reading matches the contents of the first storage means, it is reported as a reading error caused by a failure or interruption of the central processing unit, and if they do not match. a determination means that reports a normal read error when a failure occurs in the first and second central processing units; and a determination unit that determines whether the first central processing unit A first means that enables all of the means when a failure occurs, and a second means that disables the first storage means and the determination means among the aforementioned means when a failure occurs in the second central processing unit. 1. A magnetic disk control device comprising: means.