JPH09293026A - Redundant conversion method of device storage part and its device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To protect rising data in a device and important data such as state setting information, which is kept in an in-device storage part. SOLUTION: The storage part is physically made into a redundance and, for example, a zero-system storage part 31 and a one-system storage part 32 are provided. When a device control part writes prescribed sized data in the storage part, data required to be stored is divided into plural blocks. Then, data is alternately written in the storage parts (the zero-system storage part 31 and the one-system storage part 32) which are made into the redundance by block unit at the time of writing data in the storage part. Thus, writing in the both-system storage parts 31 and 32 by a same timing is excluded and the distruction of existing normal data by abnormal data writing is prevented by the device control part.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device storage unit redundancy system, and more particularly to a storage unit redundancy method capable of preventing destruction of existing data on the storage unit when an abnormality occurs on the control side for writing. And equipment.

[0002]

2. Description of the Related Art Conventionally, this type of redundancy system has been used for the purpose of protecting information stored in a device from a hardware failure of the device in a device requiring high reliability.

As a conventional technique of a redundancy method, for example, Japanese Patent Laid-Open No. 57-137948 discloses that two data storage units are provided, and a parity check circuit is provided in each duplicated data storage unit. A method has been proposed in which, when writing, the data is written in the storage means of both systems, and when the stored data is read, the data of both systems is compared, the normality of the data is checked from the parity result, and normal data is selected and read. .

Further, in Japanese Patent Laid-Open No. 58-001898, the storage units are duplicated, and a multi-bit error detection / correction function (ECC) is provided in each storage unit, so that the normality of the data when the storage unit is refreshed. If an error is detected by checking the above, by rewriting (correcting) the incorrect data section with the normal data from the other system storage section at that time,
A method has been proposed in which the storage units of both systems are always kept in a normal state.

[0005]

However, in the above-mentioned conventional technique, when the control unit for writing data in the storage unit becomes an obstacle, the storage unit is made redundant, but There is a problem that existing data (normal data) on the storage unit may be destroyed at the same time in both systems.

The reason for this is that, in the above-mentioned prior art, when writing data, the same data is simultaneously written in the duplicated storage section. Therefore, in the control section or the peripheral circuit of the control section for writing the data. This is because when abnormal data is written due to the occurrence of a failure or the like, the storage unit data of both systems are destroyed at the same time.

The storage unit in the device stores device start information,
It stores important information necessary for the device to operate, such as device management information. If this information is destroyed, it will be impossible to start up the device, or a resetting / re-entering work process by the operator will be required.

Protecting the stored information from abnormalities such as device failure is extremely important for maintenance of networks / devices.

Therefore, the present invention has been made in view of the above circumstances, and an object of the present invention is to protect important data stored in an internal storage unit such as startup data and status setting information in the device from the device failure. By doing so, it is an object of the present invention to provide a device storage unit redundancy method that improves the reliability of the device.

Another object of the present invention is to provide a storage section redundancy method and apparatus that reduce the load on the control section in the guarantee processing for protecting the storage section data from the apparatus failure.

[0011]

To achieve the above object, according to the present invention, when the storage unit is duplicated and the device control unit writes to the duplicated storage unit, the write data is divided into blocks, In a block unit, first, the data is written in the redundant 0-system storage unit, then the written data is copied from the 0-system storage unit to the 1-system storage unit, and the data of the next block is written in the 1-system storage unit. By duplicating the process of writing to the system memory unit, copying the data from the system 1 memory unit to the system 0 memory unit, and then writing the next block to the system 0 memory unit, duplication is achieved. Provided is a method of making a device storage unit redundant, which is characterized by writing information in the storage unit.

In the present invention, when the information is read from the duplicated storage unit, the validity of the data in each of the 0/1 system is determined from the management information and the written data in block units in the 0/1 system storage unit. Is confirmed, and information from the effective system is used as read data.

Further, in the present invention, each of the duplicated storage units is mapped to the same address as seen from the device control unit, the information in the storage unit of a certain system is read, and the read data is rewritten to the same address. It is characterized in that copying is performed between the duplicated storage units by performing writing.

Further, according to the present invention, in a redundant device in which a storage unit is duplicated into a 0-system storage unit and a 1-system storage unit, when data is written from the device control unit to the duplicated storage unit, the data is blocked. And writes the nth block in the 0-system storage unit, copies the nth block written in the 0-system storage unit from the 0-system storage unit to the 1-system storage unit, and then copies the n + 1th block. After writing to the system 1 storage unit, the (n + 1) th block written to the system 1 storage unit is copied from the system 1 storage unit to the system 0 storage unit. (EN) Provided is a device storage unit redundancy device, characterized in that the device storage unit is provided with means for controlling switching so as to alternately write data in block units at different times.

In the apparatus storage unit redundancy apparatus according to the present invention, a system selection signal output from the apparatus control unit is used to select a system of the 0-system storage unit and the 1-system storage unit, and the system selection is performed. The read operation and the write operation of the system designated by the logical value of the signal are configured to be opposite to each other in the 0-system storage unit and the 1-system storage unit, and the 0-system storage unit and the 1-system storage unit are configured as follows. The address controller is configured to be mapped to the same address space as seen from the device control unit, and after reading from one system at a certain address, writing to the other system at the same address as the address, It is characterized in that it is configured to copy data between the system and the system of the 1-system storage unit.

[0016]

Preferred embodiments of the present invention will be described below with reference to the drawings. First, the principle of the storage unit redundancy system according to the embodiment of the present invention will be described below.

In a preferred embodiment of the invention,
The storage unit is physically made redundant (the 0-system storage unit 3 in FIGS. 1 and 3).
When writing data from the 1st and 1st system storage section 32) and the device control section (1 in FIG. 3) to the storage section, the write data is divided into blocks and written alternately to both system storage sections.

When reading data from the storage unit, the validity of the data in each of the 0/1 type storage units is determined from the management information and the written data in block units in the 0/1 type storage units 31 and 32. Information from the confirmed and validated system is used as read data.

Further, at the time of writing to and reading from the redundant storage unit, the system control signal (42 in FIG. 3) is used to select the system of the 0-system storage unit 31 and the 1-system storage unit 32. Specify from.

Further, as shown in FIG. 4, the system selection signal 4
In the system designated by 2, the hardware is configured such that the read and write operations are opposite to each other in the 0/1 system storage units 31 and 32, and the 0 system / 1 system storage unit is the device control unit. It is configured to be mapped to the same address space as seen from the above.

As described above, according to the embodiment of the present invention, the data is alternately written in the redundant storage section,
This makes it possible to eliminate the timing of simultaneously writing data to both storage units.

The read / write operation of the storage section by the system selection signal 42 is reversed between the 0 system storage section and the 1 system storage section, and both system storage sections 31 and 32 are mapped in the same memory space. Therefore, by reading from a certain address and then writing to the same address, it is possible to perform intersystem copying of the redundant storage units 31 and 32. Therefore, in the embodiment of the present invention, duplicate address management and operation of the system selection signal for each copy operation unit are unnecessary.

Further, at the time of reading data, the normality of the data can be confirmed by checking the error detection information in block units.

The above-described embodiment of the present invention will be described in more detail with reference to the drawings.

FIG. 1 schematically shows a processing method when writing data from a device control unit (not shown) to a duplicated storage unit in the embodiment of the present invention.

When the device control unit writes data of a predetermined size in the storage unit, first, the data that needs to be stored is divided into a plurality of blocks.

When data is written in the storage unit, the storage unit (0-system storage unit 3 is made redundant in block units.
By alternately writing data to the 1st and 1st system storage units 32), it is possible to eliminate writing to both system storage units 31 and 32 at the same timing, and to write the abnormal data by the device control unit. To prevent the destruction of normal data.

It is conceivable that a failure that occurs in the control unit can be guaranteed by making the control unit redundant, but in the apparatus, the control unit often writes data to the storage unit, and in particular, continuous data In many cases, the write operation is performed.

Conventionally, in order to detect a failure in the control unit (CPU), for example, a detection method using a watchdog timer or the like has been used. However, in the present technology, an abnormality is instantly detected and the control unit is switched or The reality is that forced stop is impossible.

Therefore, conventionally, when a failure occurs in the control unit during continuous writing in the storage unit, although it depends on the content of the failure, illegal data is written in both system storage units and existing normal data is written in both. There was a great possibility that the systems would be destroyed at the same time.

On the other hand, in the embodiment of the present invention, various methods of dividing the stored information are conceivable.
If the function is divided in consideration of the content of information, it becomes possible to manage efficiently when a device failure occurs.

On the other hand, when considering only the redundancy processing algorithm, the method of dividing into the same size becomes easy regardless of the content of the data. Further, the smaller the size of the division is, the less damage is caused when a failure occurs, but the processing time is lengthened because the processing amount of the control unit is increased.

In the embodiment of the present invention, information for data management is preferably created in each divided block. As the data management information, as shown in FIG. 1, error detection information 51 (CRC or the like) for confirming the normality of the data in the block, and because the data is divided into blocks, the data area is taken into consideration. Block number 52, area usage size (block size) 53 that represents the number of data in the divided blocks, and if writing is interrupted by a failure while the control unit is writing data, the block in which the writing is interrupted is invalid data. The write normal state 54 for identifying this invalid block at the time of data reading. Actual user data 55 is written in the remaining area in the block.

The procedure when the control unit writes data will be described below.

First, the write state 54 in the first block of the 0-system storage unit 31 is changed to "writing".

After that, user data is written, error detection information is calculated, and the error detection information in block 51 is written.

Further, the data management block number 52,
The area use size 53 is changed.

When all of these processes are completed, the write normal state 54 is changed to "write completed", and the writing in the 0-system storage unit 31 is completed.

Next, the selection system is switched, the 0-system storage unit 31 is changed to the read mode, and after confirming that the write normal state 54 of the 0-system storage unit 31 is "write completed", 0
Using the error detection information 51 read from the system storage unit 31, the normality of the data block of the 0-system storage unit 31 that was just written is confirmed (the error detection information calculated for the data block and the read error). It is confirmed whether or not the detection information 51 matches).

After confirming that the data block of the 0-related storage unit 31 is normal, the data of the 0-based storage unit 31 is copied to the 1-based storage unit 32. Regarding the procedure of copying, change the writing normal state 54 of the 1-system storage unit 32 to “writing” before starting writing to the 1-system storage unit 32.
After that, the data in the 0-system storage unit 31 is copied to the 1-system storage unit 32.

When copying is completed for the number of bytes written in the 1-system area use size 53, data is then written to the 1-system storage section 32.

After the writing to the 1-system storage section 32 is completed, the error detection information 51, the block number 52, and the area use size 53 are updated in the same manner as when the 0-system storage section 31 is written. Finally, the write state 54 is changed to "write complete".

Writing of a block to the storage unit of one system, copying of the write block of the storage unit of one system to the storage unit of the other system, writing of the next block to the other system, the other The operation of copying one of the write blocks of the storage unit of the system to the storage unit of one system is repeated for all blocks until the write data is completed.

FIG. 2 shows a method of accessing the storage unit at the time of reading data in the embodiment of the present invention.

First, the area (block) to be read is specified (step S20).

After that, the write state 5 in the corresponding block
Read 4 In the determination processing of step S21, if the content of the write state (54 in FIG. 1) is “writing”, it is assumed that the device control unit has stopped due to a failure or the like while writing to the storage unit. Therefore, the data of the block is not used.

Next, the selection of the duplicated storage section is reversed, and the same processing is performed in the opposite system (step S
24).

When the contents of the write state 54 in the storage units of both systems are "writing", the data of the corresponding block is read out.

On the other hand, when the content of the write state 54 is "write completed" in any of the duplicated systems, error detection information is calculated from the user data in that block (step S22).

The result calculated from the user data in the block is compared with the error detection information written in the error detection information 51 in the storage block at the time of writing the data (step S23). The selection of the redundant storage unit (redundant system) is reversed (step S2
4) The same process is performed in the opposite system.

If the error detection information does not match in the storage units of both systems, it is considered that the data has been destroyed due to a hardware failure in the storage units of both systems. Performs unreadable processing.

If the value recorded in the error detection information 51 in the storage unit and the error detection information calculated from the user data 55 in the block match in any of the duplicated systems, the user data in the block is matched. 55 is used as valid data.

The above-described embodiments of the present invention will be described in more detail with reference to Examples of the present invention. FIG. 3 is a diagram showing an embodiment of the redundancy method according to the present invention.

Referring to FIG. 3, in the present embodiment, the storage unit for storing the in-apparatus important data is a 0-system storage unit 31,
It is composed of a 1-system storage unit 32. That is, the storage unit is physically separated and completely duplicated.

The device control unit 1 manages not only data management but also the entire device. In FIG. 3, only the interface unit with the storage unit is shown for easy understanding.

As an interface with the storage unit, the device control unit 1 uses the address bus (n bits) 5 for specifying the positions of the 0-system storage unit 31 and the 1-system storage unit 32, and the device control unit 1. When writing data to the 0-system storage unit 31 and 1-system storage unit 32, write information is specified, and when the data is read from the storage unit, the 0-system storage unit 31 and the 1-system storage unit 3 are specified.
A data bus (n 'bit) 4 for reading the information from 2 and a write control signal (Write signal) for giving a write instruction from the device control unit 1 to the 0 system storage unit 31 and the 1 system storage unit 32. 41 and device controller 1 to 0
A read control signal (Read signal) 43 that gives a read instruction to the system storage unit 31 and the system 1 storage unit 32, and the system control unit 31 and the system 1 storage unit 3 in which the device control unit 1 is duplicated.
System selection signal 42 that specifies which of the two is to be accessed
And is composed of

The address bus 45 and the data bus 44 are connected in parallel from the device control unit 1 to the 0 system storage unit 31 and the 1 system storage unit 32.

A write control signal (Write signal)
41 and a read control signal (Read signal) 43 are the first and second switches 2 controlled by the system selection signal 42.
It is connected to the storage units 31 and 32 via 1 and 22.

FIG. 4 shows an example of the access relationship between the system selection signal 42 and the redundant storage sections 31 and 32.

System selection signal 4 designated by the device controller 1
When the device control unit 1 performs a write operation in the state where the logical value of 2 is “0”, data is written in the 0-system storage unit 31 and a read operation occurs in this state. Data is read from the 1-system storage unit 32,
When the device control unit 1 performs a write operation while the logical value of the system selection signal 42 designated by the device control unit 1 is "1", data is written in the 1-system storage unit 32, and When the reading operation is performed in this state, the first and second data are read so that the data is read from the 0-series storage unit 31.
The switches 21 and 22 are configured.

Further, the first and second switches 21 and 22
Is configured so that the operation to the storage unit becomes inactive when the output to the storage unit side becomes open.

The writing and reading operations of the device controller 1 will be described below with reference to FIGS. 3 and 4.

When the device control unit 1 writes data in the storage unit, first, the system selection signal 42 is controlled to specify the target storage unit system. For example, when it is desired to write to the 0-system storage section 31, the system selection signal 42 is set to "0".

After that, the address information indicating the position of the storage portion is specified by the address bus 45, the data to be written is specified by the data bus 44, and the write control signal (Wri) is specified.
te signal) 41 to instruct writing.

When the device control unit 1 reads the data in the storage unit, first, the system selection signal 42 is controlled to specify the target storage unit system. For example, when it is desired to read out to the 0-system storage section 31, the system selection signal 42 is set to "1".

After that, the address information indicating the position of the storage section is specified by the address bus 45, the read instruction is given by the Read signal 43, and the data on the data bus 44 is read.

In the embodiment of the present invention, as described above, the copying operation is required between the duplicated storage sections.

As shown in FIG. 3, in this embodiment, the system 0 storage unit 31 and the system 1 storage unit 32 are replaced by the device control unit 1.
Since they are mapped on the same address, the system selection signal 42 from the device control unit 1 causes the system 0 storage unit 3 to
A read / write operation is instructed to the 1st or 1st system storage unit 32.

Since the operations to the 0-system storage section 31 and the 1-system storage section 32 are configured to be opposite to each other, the device control section 1 can operate without changing the logical value of the system selection signal 42. After the read operation is performed, the read data is rewritten to the same address, thereby copying between the duplicated storage units. Further, by inverting the system selection signal 42 and performing the same operation, it is possible to reverse the direction of copying between the storage units.

[0070]

As described above, the present invention has the following effects.

According to the present invention, the device control unit itself or the device control unit while writing information in the redundant storage unit, or
Even if an abnormal operation occurs due to a hardware failure between the device control unit and the storage unit, it is possible to prevent the destruction of existing data in at least one storage unit of the redundant storage unit and guarantee the important data in the device. It has the effect of being able to.
Therefore, according to the present invention, it is possible to prevent a device and communication failure due to abnormal data reading at the time of switching the redundant system or restarting the device.

Further, according to the present invention, it is possible to minimize the destruction of the data stored in the storage unit at the time of the device failure, which has the effect of improving the reliability of the device.

This is because in the present invention, when information is written from the device control unit to the duplicated storage unit, simultaneous writing is not performed to the storage units of both systems, but the writing is blocked, and redundancy is performed in units of blocks. This is because writing is alternately performed in the converted storage unit.

Further, another advantage of the present invention is that the device can read out normal information even when the redundant one-side storage unit itself fails, and in the case of both-system failure, The device control section has the effect of being able to recognize that the data in the storage sections of both systems is abnormal.

This is because, in the present invention, the device control section can recognize the system of the duplicated storage section and read / write independently for the 0 system and the 1 system, and at the time of reading information, By checking the normality of the data in the storage unit from the data itself written in the storage unit and the error detection information, it is possible to detect the data abnormality even with respect to the failure of the storage unit itself.

Further, according to the present invention, there is an effect that copying from a storage unit of one system to another system storage unit is easy. This means that in the present invention, a copying operation is indispensable between the redundant storage units, but the storage units of both systems are mapped to the same address as seen from the control unit, and the read operation and the write operation are 0 system. With the configuration in which the storage unit and the 1-system storage unit are inverted, a copy operation can be performed by reading and rewriting to the same address, and address duplication management and copy operation (byte) unit This is because the control of the system selection signal for each is unnecessary.

[Brief description of drawings]

FIG. 1 is a diagram for explaining an embodiment of the present invention, and is a diagram schematically showing a method of writing data to a redundant storage unit.

FIG. 2 is a diagram for explaining the embodiment of the present invention and is a flowchart for explaining a method of reading data from a redundant storage unit.

FIG. 3 is a diagram showing a configuration of an exemplary embodiment of the present invention, and is a diagram showing a functional configuration of a device regarding a storage unit.

FIG. 4 is a diagram for explaining one embodiment of the present invention;
It is a figure which shows the logic of a system selection signal, and the relationship of the selection system of memory | storage part access at the time of data writing and data reading.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Device control unit 21 Write signal switching unit 22 Read signal switching unit 31 0 system storage unit 32 1 system storage unit 41 Write instruction signal 42 Storage unit system selection instruction signal 43 Read instruction signal 44 Data bus signal 45 Address bus signal 51 Error detection Information 52 Block number 53 Area usage size 54 Writing status 55 User data

Claims (7)

[Claims] 1. When a storage unit is duplicated, and when writing is performed from the device control unit to the duplicated storage unit, write data is divided into blocks, and in each of the blocks, the 0-system storage unit is made redundant first. To the first system storage unit, and the data of the next block is written to the first system storage unit. Then, the data of the next block is written to the first system storage unit. After copying the data to the 0-system storage section, further writing the next block to the 0-system storage section is repeated, thereby writing information in the duplicated storage section. Redundancy method. 2. When reading information from the duplicated storage unit, the validity of the data in each of the 0/1 system is confirmed from the management information and the written data in block units in the 0/1 system storage unit. 2. The method for making a device storage unit redundant according to claim 1, wherein information from a valid system is used as read data. 3. By mapping each of the duplicated storage units to the same address as seen from the device control unit, reading the information in the storage unit of a certain system, and rewriting the read data to the same address. 2. The apparatus storage unit redundancy method according to claim 1, wherein copying between the duplicated storage units is performed. 4. The block includes, in addition to user data, information for managing the block, and error detection information, further information indicating whether or not writing of the block is completed. 3. The apparatus storage unit redundancy method according to any one of 1 to 3. 5. A redundant device in which a storage unit is duplicated into a 0-system storage unit and a 1-system storage unit, and when the data is written from the device control unit to the duplicated storage unit, the data is divided into blocks, and After writing the block of n into the 0-system storage unit,
The nth block written in the system memory is copied from the 0 system memory to the system 1 memory, and then the n + 1th block is written in the system 1 memory and then written in the system 1 memory. Means for controlling switching so that the (n + 1) th block is copied from the 1-system storage section to the 0-system storage section such that the 0-system storage section and the 1-system storage section are alternately written in block units at different times. And a device storage unit redundancy device. 6. A system selection signal output from the device control unit is used to select a system of the 0-system storage unit and the 1-system storage unit,
The 0-system storage section and the 1-system storage section are configured so that the read operation and the write operation of the system specified by the logical value of the system selection signal are opposite to each other in the 0-system storage section and the 1-system storage section. The storage unit is configured to be mapped in the same address space as seen from the device control unit, and after reading from one system at a certain address, writing to the other system at the same address as the address is performed. 6. The apparatus storage unit redundancy device according to claim 5, wherein data is copied between the 0-system and 1-system storage units. 7. When the information is read from the duplicated storage unit, the normality of the data stored for each system is confirmed, and when the information of one selected system is invalid, the other system is checked. 7. The device storage unit redundancy device according to claim 5, wherein the information in the storage unit is confirmed, and the information from the valid system is used as read data.