JP2500440B2 - Fault handling method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a failure processing method for an information processing apparatus which performs real time processing.

[0002]

2. Description of the Related Art In a conventional failure processing method in an information processing apparatus, since the state of the internal memory at that time is logged out when a failure is detected, information in the internal memory is stored in the internal memory for the purpose of freezing the internal memory state by a microprogram. After saving to the logout area, normal operation recovery processing is performed.

[0003]

Generally, since a large amount of data is saved in the internal memory, a transition from a failure detection to a normal operation recovery process (a process for making a request such as a real-time process available for reception) is performed. Takes time, and even in the case of a single continuous failure, the real-time processing request is kept waiting, which has a drawback that it hinders business.

[0004]

A failure processing method according to the present invention is a failure processing method in an information processing apparatus that performs real-time processing, and includes a first storage section having a plurality of blocks, and the first storage section. A second storage unit having a plurality of blocks corresponding to addresses of the plurality of blocks for storing a log, and the same data in the second storage unit when writing data to the first storage unit. Storage unit control for performing a data log operation of writing to the second storage unit and controlling inhibition / start of each data storage block in each block of the second storage unit and reading / writing of data to the second storage unit Means, storage unit write detection means for detecting the presence or absence of a write operation to the first storage unit for each block, and initializing each block individually based on the detection result, and failure detection of the information processing device. A normal operation recovery means to said storage unit controlling means can be accepted instructions normally process the suppression of data logging operations to all the blocks of the second storage unit, the
Logout means for logging out the data in the second storage unit after processing by the normal operation recovery means, and copy processing of all data from the first storage section to the second storage section after processing by the logout means When there is another processing request during the copy process, the copy process is interrupted, the storage unit control unit is instructed to restart the data log operation up to the block where the copy is completed, and the storage unit write detection unit is executed. And a storage unit data restoring means for instructing initialization of the block in which the copy processing is interrupted.

Further, the storage section data restoring means is based on the detection result by the storage section write detecting means, when the writing operation to the block in which the copy processing is interrupted is not performed during the copy processing interruption. The copy process is restarted from the interrupted position, and if the write operation is being performed, the copy process is restarted from the beginning of the block in which the copy process is interrupted, and when the copy process of all data is completed, the second storage unit is restarted. It is characterized by restarting the data log operation of all blocks that are suppressed.

[0006]

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

FIG. 1 is a block diagram showing the configuration of an embodiment for realizing the present invention. As shown in FIG. 1, the failure processing method of the present invention includes a processor 1 and a control memory (hereinafter, referred to as a control memory) that stores a program for operating the processor 1.
2) and internal memory (hereinafter referred to as LM) 3
And a log internal memory (hereinafter referred to as a log LM) 4
And a log LM controller 5 for controlling the operation of the log LM 4, and an LM write detection circuit 6. In addition, a normal operation recovery processing program 7, a logout program 8 and a log LM data restoration program 9 are stored in the CS 2.

The processor 1 operates according to the instructions of the program stored in the CS 2 and performs processing by accessing the LM 3 and an I / O (input / output unit) not shown. The log LM4 is arranged at the same address as the LM3. Further, when the processor writes data to the LM3, the write LM controller 5 writes the same data to the log LM4 so that the log LM4 has the same image as the LM3. Hereinafter, the data log operation) is controlled.

Further, the LM 3 and the log LM 4 are divided into a plurality of blocks with the same structure, and in this embodiment, each is divided into 4 blocks.
The blocks of LM3 are LM3a to 3d, and LM4 for log is used.
Blocks are designated as log LMs 4a to 4d.

The log LM controller 5 has a function of starting and suppressing a data log operation for each block according to an instruction from the processor 1, a function of reading data in the log LM 4 according to an instruction from the processor 1, and a log. For LM4
It has a function of writing data to the LM3 separately.

The LM write detection circuit 6 has write detection F / Fs 6a to 6d composed of F / Fs (flip-flops) for detecting the presence / absence of writing corresponding to each block of the LM3, and the LM3 of the processor 1 is connected to the LM3. Monitors access. Then, at the time of write operation, the corresponding block is determined by the address to be accessed, and the corresponding write detection F
Set one of / F6a-6d. The write detection F / Fs 6a to 6d can be read by the processor 1, and can be reset for each write detection F / F 6a to 6d.

Here, the correspondence between the blocks of the LM3 and the log LM4 and the write detection F / Fs 6a to 6d are assumed to correspond to those having the same subscript in the figure.
Block 3a of LM3, block 4a of LM4 for log,
The write detection F / F 6a of the LM write detection circuit 6 corresponds thereto.

Next, the operation when a failure of the information processing device is detected will be described together with the operations of the normal operation recovery processing program 7, the logout program 8 and the log LM data restoration program 9 provided in the CS2.

First, when a failure of the information processing device is detected and the normal operation recovery program 7 is started by the processor 1, the normal operation recovery processing program 7 causes the log LM controller 5 to perform a data log operation to the log LM 4. Immediately after giving an instruction to suppress, normal processing such as real-time processing by interruption is ready to be accepted.

After the normal operation recovery processing program 7 is operated, the processor 1 is ready to execute the normal processing and operates the logout program 8 by utilizing the idle time of the normal processing. However, if a failure occurs again at this point, LM logout corresponding to this failure cannot be performed.

Next, the logout program 8 uses the maintenance diagnostic processor (S) (not shown) for the data in the log LM 4.
VP) and then operates the log LM data restoration program 9. However, if another normal processing request occurs during execution of the logout program 8, the execution of the logout program 8 is temporarily suspended, and the requested normal processing is preferentially executed. Then, when the normal process ends, the logout program 8 restarts the process from the position where it was interrupted.

Further, the log LM data restoration program 9 performs a copy process of all data from the LM 3 to the log LM 4 by transferring the program. However, if a normal processing request occurs even during the operation of the log LM data restoration program 9, the normal processing request is preferentially executed, so the log LM data restoration program 9 suspends its operation. Sometimes.

Now, from the block 3a of the LM3 to the log L
The copy processing from the block 4a of M4 and the block 3b of the LM3 to the block 4b of the log LM4 is completed, and then the copy processing from the block 3c of the LM3 to the block 4c of the log LM4 is completed. In the meantime, when the processing by the log LM data restoration program 9 is interrupted due to a request for normal processing, the log LM controller 5
To the block 3c of the LM3 in which the copy process is interrupted, as well as instructing to start the data log operation of the blocks 4a to 4b of the log LM4 for which the copy process is completed The write detection F / F 6c is reset, and the requested normal processing is performed.

When the processing by the log LM data restoration program 9 is restarted after the end of the normal processing, the write detection F / F 6c is checked. And write detection F /
If F6c is not set, the copy process is restarted from the position where the copy process was interrupted. If F6c is set, the copy process is restarted by returning to the head of the block 4c of the log LM4.

When the copy processing from the LM3 to the log LM4 is completed, the log LM controller 5
LM4 for log that has suppressed the data log operation at that time
To start the data log operation for all blocks of This enables the LM3 to log out even if a failure occurs again.

[0021]

As described above, according to the present invention, if there is a failure that does not occur consecutively, even if the internal memory state is frozen for logout, the failure detection can be immediately followed by the normal operation recovery processing. can do.

Further, there is an effect that it is not necessary to wait for a real-time processing request because the internal memory state is frozen.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment for realizing the present invention.

[Explanation of symbols]

 1 Processor 2 CS (Control Memory) 3 LM (Internal Memory) 3a to 3d Each Block of LM3 4 Log LM (Internal Memory for Log) 4a to 4d Each Block of LM4 for Log 5 LM Controller for Log 6 LM Write Detection Circuit 6a to 6d Write detection F / F (flip-flop) 7 Normal operation recovery processing program 8 Logout program 9 LM data restoration program for log

Claims (2)

(57) [Claims] 1. A failure processing method for an information processing apparatus that performs real-time processing, comprising: a first storage unit having a plurality of blocks; and a plurality of storage units corresponding to addresses of the plurality of blocks in the first storage unit. Second with a block for storing logs
And a data log operation for writing the same data to the second memory unit when writing data to the first memory unit, and this data log operation is performed for each block of the second memory unit. Storage unit control means for controlling inhibition / starting to the second storage unit and reading / writing of data from / to the second storage unit, and presence / absence of write operation to / from the first storage unit is detected for each block and detected. A storage unit write detection unit that initializes each of the blocks individually based on the result, and a data log operation to all blocks of the second storage unit for the storage unit control unit when a failure of the information processing device is detected. And a logout means for logging out the data in the second storage unit after the processing by the normal operation recovery means, and the logout means. Run the copy processing of all data from management after the first storage unit to the second storage unit,
When there is another processing request during the copy process, the copy process is interrupted, the storage unit control unit is instructed to restart the data log operation up to the block where the copy is completed, and the storage unit write detection unit is connected to the copy process. And a storage unit data restoring means for instructing initialization of the block in which the interruption is performed. 2. The storage unit data restoration means, based on the detection result of the storage unit write detection means, when the writing operation to the block in which the copy processing is interrupted is not performed during the interruption of the copy processing, The copy process is restarted from the interrupted position, and if the write operation is being performed, the copy process is restarted from the beginning of the block in which the copy process is interrupted, and when the copy process of all data is completed, the second storage unit is restarted. 2. The failure processing method according to claim 1, wherein the data log operation of all the blocks which are suppressed is restarted.