JP2001318801A - Duplex computer system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quickly perform the restoration processing of a memory file, and to quickly transit a standby computer to an operating state in a duplex computer system accompanied with processing to transit a standby system computer to an operating state. SOLUTION: An active computer 101 transmits the contents of a memory file 203 being data to be shared and log data 205 of the memory file 203 as restoration data, and when the restoration data are the contents of the memory file, a standby computer 102 stores the data in a memory file 302 as they are, and when the restoration data are log data, reads a preservation file 501, and adds update with the log data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for transitioning the operation state of a standby computer in a dual computer system.

[0002]

2. Description of the Related Art Conventionally, in a system which requires continuous operation for 24 hours, in order to improve the reliability of the system, an execution system computer for actually performing processing and an execution system computer when a failure occurs in the execution system computer. In many cases, a dual computer system including a standby computer that performs the above alternative processing is employed.

In a dual computer system, the state in which a standby computer is stopped due to a failure, system maintenance, or the like must be kept as short as possible. However, in order to transition the standby computer from the stop state to the execution state, it is necessary to make the operating computer and the operating environment the same, and speeding up the processing is a problem.

For example, in Japanese Patent Application Laid-Open No. Hei 8-202571, "Data Matching Method and Computer System for Shared Multiplexing Auxiliary Apparatus", a redundant auxiliary storage device is shared between an active computer and a standby computer. While writing a log about the location of the area where the writing has occurred in the shared auxiliary storage device, the writing is performed to both of the duplicated auxiliary storage devices. If a change occurs between the active computer and the standby computer while the write is being performed, the new active computer checks the area where the write has not been completed based on the log from the active computer before the replacement. Then, data is matched at high speed by copying only that area.

Japanese Patent Application Laid-Open No. H10-49418 discloses a method and apparatus for reflecting journal data and a redundantly configured computer system as a method for matching data between an active computer and a standby computer when an auxiliary storage device is not shared. "
There is. In this method, the log data acquired by the active computer is reflected by sequentially transferring the log data to the standby computer using a communication medium.

[0006]

According to the conventional method,
When the active computer and the standby computer do not share the auxiliary storage device, the write information of the data on the active computer while the standby computer is stopped is accumulated as log data. In order for the standby computer to recover from the stopped state to the normal standby state, it is necessary to reflect the log data during the stopped state. When the log data in the active computer becomes enormous, it takes time to reflect the data, and during that time, the standby computer cannot transition to the normal standby state.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to perform high-speed file and memory file recovery processing, and to quickly transition the operation state of a standby computer. .

[0008]

A dual computer system according to the present invention is a dual computer system comprising an active computer and a standby computer which are connected to each other and store synchronized memory files respectively. The execution system computer includes an execution system computer side memory file storage unit that stores a memory file used by the execution system computer, a data update unit that updates a memory file of the execution system computer side memory file storage unit, and an update of the memory file. A log acquisition unit that acquires the log of the log, a log data storage unit that stores the acquired log, and a memory file that is stored in the memory file storage unit of the active computer when restoring the memory file of the standby computer. Alternatively, a log stored in the log storage unit is selected as the recovery data, and the recovery data for transmitting the selected recovery data is selected. A standby computer, wherein the standby computer includes a standby computer-side memory file storage unit that stores a memory file used by the standby computer, and a file storage unit that duplicates the memory file of the standby computer-side memory file storage unit. And a storage file storage unit for storing the copied memory file as a storage file, and when restoring the memory file of the standby computer, receiving restoration data, and when the received restoration data is a memory file,
The recovery data is stored in the memory file storage unit as a memory file in the memory file storage unit, and when the received recovery data is a log, the recovery log is stored in the recovery log data storage unit. When the recovery data is a log, the storage file reading unit reads the storage file in the storage file storage unit as a memory file into the memory file storage unit on the standby computer, and the storage file reading unit reads the storage file into the memory file storage unit on the standby computer. And a memory file recovery unit that updates the memory file based on the log in the recovery log data storage unit.

The storage file reading unit determines, based on the log, a portion of the memory file that is not updated by the memory file recovery unit, and selects a portion of the storage file in the storage file storage unit that is not updated as the memory file. And reading.

[0010] The log obtaining unit obtains a log having a log entry corresponding to each update, the log entry being able to grasp before and after the update timing, in response to a plurality of updates to the memory file by the data update unit. The memory file restoring unit updates the memory file based on the new log entry at the sequential update timing, excluding the already updated portion.

The standby computer further has a write request receiving unit for receiving a write request to the memory file of the standby computer, and the memory file recovery unit transmits the received write request to the log file of the memory file. If the write request is for a portion to be updated based on the log, the memory file is updated based on the log.

[0012] In addition, when the data update unit has performed a previous update and a subsequent update to a common part of the memory file by the data update unit, the log acquisition unit deletes the data recorded as the memory file at the time of the previous update. It is obtained as a part of the log at the time of later update.

In addition, the log acquisition unit acquires a log having a log entry corresponding to each update and having a log entry that can grasp before and after the update timing in response to a plurality of updates to the memory file by the data update unit. In addition, the execution system computer further supports the previous update when the acquired log includes a log entry corresponding to the previous update to the common part of the memory file and a log entry corresponding to the later update. A log compression unit that deletes, from the log entry, data included in the log entry to be recorded and recorded in the common part of the memory file at the time of the previous update.

Further, the storage file reading section is characterized in that a portion of the storage file in the storage file storage section that has not been read by the storage file reading section before is selected and read.

Further, the log acquisition unit is characterized in that when the size of the log to be acquired exceeds a certain size, the acquisition of the log is stopped.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiment 1 In the present embodiment, in a dual computer system including an active computer and a standby computer, while the standby computer is in a stopped state,
The active computer has a mechanism to save the writing process to the file in the log, and a mechanism to determine whether to send the file itself as recovery data from the log content or to send the log when the standby computer returns A high-speed transition method for quickly transitioning the operation state of the standby computer, which is characterized by having the feature, is described.

FIG. 1 shows a dual computer system,
FIG. 12 is a diagram illustrating a state transition from when an abnormality occurs in one computer system to when both computers transit to a normal state.

Computers 101 and 10 constituting a dual system
2, when the computer 101 is operating normally as the active computer and the computer 102 is operating normally as the standby computer (state 1), the computer 101 is disabled for some reason.
Occurs, the computer 102, which has been operating as a standby system, operates as an active system, and the computer 101 transitions to a stopped state (state 2).

The computer 101 operates normally as a standby computer by removing the cause of the abnormal state of the computer 101 and performing a repair operation, thereby operating as a normal dual computer system (state 4).

The computer that becomes abnormal even when the state transitions from state 1 to state 5 to state 6 to state 7 is the computer 102, and the executing computer operates as it is.

The present invention is a method for reducing the time required for the standby computer to transition from the stopped state to the normal (standby) state (reducing the time in state 3).

In the data write process, the data flow differs between state 1 and state 2. FIG. 2 is a diagram illustrating a data flow in a data write process of the active computer and the standby computer when the dual system is in the state 1.

When the application 201 on the computer 101, which is the execution computer, writes data to the memory file 203, the data update unit 202 is notified of the address, size, and update data in the write processing.

The data updating unit 202 transmits the notified contents to the data updating unit 301 of the computer 102 which is the standby computer.
Notify.

In the subsequent processing, the same processing is performed by the computer 101 and the computer 102. Data update unit 202,
301 notifies the log acquisition units 204 and 303 of the address, size, and update data.

The log acquisition units 204 and 303 store the address, size, and update data notified to the log data 205 and 304.

Thereafter, the data updating units 202 and 301
The memory files 203 and 302 are updated.

The storage files 401 and 501 on the auxiliary storage devices 103 and 104 are stored in the memory files 203 and 3 respectively.
02, the log data 205,
304 is cleared.

The timing at which the memory files 203 and 302 are stored in the storage files 401 and 501 on the auxiliary storage devices 205 and 304 is arbitrary. For example, the following timing can be considered. 1. 1. When the log data size exceeds a certain size Timing set in the system

FIG. 3 is a diagram showing a data flow in writing data in the active computer and the standby computer when the dual system is in the state 2. The processing when the dual system is in state 2 differs from the processing in state 1 in the following points. 1. The data updating unit 202 does not transmit the address / size / update data to the data updating unit 301 of the standby computer. 2. If the log data is cleared when the memory file 203 is saved, the log cannot be transmitted when the standby computer starts up, so that the clearing process is not performed.

As a result, the write processing until the standby system becomes normal can be stored as log data. When the size of the log data becomes large, the collection of the log data is stopped.

FIG. 4 is a diagram showing a write processing flow of the execution computer.

FIG. 5 is a diagram showing the flow of processing for saving a memory file.

FIG. 6 is a diagram showing a write processing flow of the standby computer.

Next, the standby system is changed from the stopped state to a normal state (standby).
Describe the recovery process required to transition to the state. FIG.
FIG. 9 is a diagram showing a data flow in a recovery process necessary for the standby system to transition from the stop state to the normal (standby) state.

The standby computer first requests data from the restored data transmitting unit 207 so that the restored data receiving unit 306 restores the memory file.

The recovery data transmission unit 207 transmits the log data 2
Based on the size and contents of the log data in 05, it is determined whether to transmit the contents of the memory file as it is or to transmit all the addresses, sizes, and update data in the log data on the active computer.

Thereafter, the determined data is transmitted to the recovery data receiving unit 306.

The recovery data receiving unit 306 performs processing according to the content of the received data. When the contents of the memory file are received, the contents are written into the memory file 302, and the process is terminated.

If the received data is log data, the contents are stored in the recovery log data 307. After that, the storage file reading unit 308 reads the restoration data from the storage file 501 on the auxiliary storage device 104 and reads the restored data into the memory file 302.

The memory file recovery unit 309 updates the memory file 302 by referring to the address, size, and update data of each log entry in the recovery log data 307 in order from the top. By performing the update process using all the log entries, the contents of the memory file of the standby computer are made the same as the contents of the memory file of the active computer.

FIG. 8 is a diagram showing a processing flow of the executing computer at the time of data recovery.

FIG. 9 is a diagram showing a flow of a recovery process in the standby computer.

The time required for restoring the memory file is "time (time 1) for transmitting the entire memory file from the active computer to the standby computer" when transferring the entire memory file. On the other hand, when transferring log data, “time for transmitting log data from the active computer to the standby computer (time 2)”, “time for reading the log file from the auxiliary storage device to the memory file (time 3)”,
It is the sum of "time to reflect log data to memory file (time 4)".

Although the time 1 and the time 3 increase depending on the size of the memory file, when compared with the processing time of the same data amount, the time from the auxiliary storage device to the memory is generally shorter than the time for transmitting and receiving data using the network. Since the time for reading into the file is shorter, the time 3 is shorter than the time 1. Since the time 2 and the time 4 depend on the size of the log data, if the amount of the log data is small, it will be faster than the copying process of the entire memory file. By sending the entire memory file when the log size grows,
At least the conventional processing time is sufficient.

As described above, when the log data is small, the reflection processing of the log data is performed, and when the log data is large, the memory file can be quickly identified by copying the file itself. The operating state can be changed at high speed.

Further, since there is an effect that the amount of data transmitted using the network is reduced, the load on the network can be reduced.

Embodiment 2 In the present embodiment, a description will be given of a high-speed transition method characterized by including a mechanism for selecting data to be read from an auxiliary storage device on a standby computer based on the content of recovery data transmitted from an active computer. .

In the present embodiment, a part of the recovery processing method performed when the standby computer transitions from the stopped state to the normal (standby) state is different. Specifically, the processing when the data received by the recovery data receiving unit of the standby computer is log data is different from the above-described embodiment.

FIG. 10 is a diagram showing a recovery data flow of the standby computer in the second embodiment.

When the recovery data receiving unit 306 receives the log data, the contents are stored in the recovery log data 307.

Thereafter, the storage file reading unit 308 refers to the recovery log data 307, checks the address and size of the log data, and obtains a portion where the log is changed to the memory file. Since the other parts are the parts that are not changed by the recovery log data, only the parts that are not changed are read from the storage file 501 as the recovery data, and the memory file 302 is rewritten.

Thereafter, based on the address, size and update data of the recovery log data, the memory file recovery unit 30
9 updates the memory file 302. By performing this process, the contents of the memory file on the standby computer are made the same as the contents of the memory file on the active computer.

FIG. 11 is a diagram showing a recovery processing flow in the standby computer in the second embodiment.

Based on the recovery log data 307, the following processing is performed to determine the position where the data needs to be read from the storage file 501 into the memory file 302. 1. The log entries of the update log data are rearranged using the position as a key (S11004). 2. From the positions and sizes of the rearranged entries, the start position and the end position of each entry are obtained (S11005). 3. Combine overlapping entries into one entry. 4. In order to invert the log data, a new entry is created with the end position as the start position and the start position of the next entry as the end position for each entry (S11006).

The data is read from the storage file 501 into the memory file 302 based on the data thus generated.

FIG. 12 is a diagram showing an example of obtaining data that requires reading a storage file from the update log data.

As a result, the data size read from the storage file 501 into the memory file 302 is reduced, so that the memory files can be identified at a higher speed, and as a result, the operation state of the standby computer can be changed at a high speed. .

Embodiment 3 In the present embodiment, when synchronizing files on the standby computer, a mechanism is provided that reflects only the latest data for writing to the same location based on the contents of the recovery data, The transition method will be described.

In the present embodiment, a part of the restoration processing is different. When the memory file recovery unit 309 determines that the memory file 30
In applying the log entry to the second embodiment, in the first and second embodiments, the following processing is performed on the log entry in order from the top. FIG. 13 is a diagram illustrating a flow of a log entry reflection process according to the first and second embodiments.

It is checked whether there is any unprocessed log entry in the recovery log data 307 (S 13001). If there is, the first log entry is taken out (S 13002), and the memory is determined based on the position, size, and changed data. File 3
02 (S13003). Since the log entries are arranged in chronological order, processing is performed in the order from the old log to the new log.

On the other hand, in the present embodiment, the log entry of the recovery log data 307 is stored in the first and second embodiments.
Conversely (from the new log entry to the old log entry), the following processing is performed.

FIG. 14 is a diagram showing a log entry reflection processing flow in the third embodiment. Checking whether there is a duplicate of the address of the log entry that has already been reflected or whether there is a duplicate part is performed by making a list structure while sequentially comparing the completed entries with a double-linked list structure. Performed (S14003).

FIG. 15 is a conceptual diagram of the link list processing in the third embodiment. For example, as a specific example, when there is a link list of completed entries as shown in the figure, a process for processing a log entry of position 1536 and size 1024 will be described.

It is determined whether there is an entry including the position 1536 in the double link list until an entry having a position exceeding the position 1536 from the head appears. as a result,
It can be seen that the middle entry is location 1024, size 1024, and includes location 1536.

Next, the end position is obtained by adding the size to the position (1536 + 1024 = 2560). In order to search for an entry including the end position, the process is repeated until an entry having a position exceeding the end position appears. In this case, since the next entry is 2560 or more, it is understood that there is no entry including the end position.

Next, a position not yet written in the memory file is determined between the entries including the position 1536 and the end position 2560. Then, it can be seen that 512 bytes from 2048 (the end position of the middle entry) to the end position 2560 are not written.

Thereafter, the entry under processing and the overlapping entry are combined into one, and a process of linking to the corresponding position is performed.

Here, position 1024, size 1024
And the entry of the position 1536 and the size 1024 are combined into one to obtain the entry of the position 1024 and the size 1536.

According to this processing, only the update processing by the new log entry is performed on the memory file.

In other words, the log entry is not redundantly reflected in the memory file, and the operating state of the standby computer can be changed more quickly.

Embodiment 4 In this embodiment, when the standby computer receives the recovery data, the standby computer is shifted to the operating state without performing the actual file synchronization processing, the file synchronization processing is distributed, and the file A description will be given of a high-speed transition method characterized in that reflection to a file is performed when a new write process arrives.

This embodiment is a modification of the first and second embodiments, and differs from the first embodiment in the processing of the standby computer and the recovery processing.

First, different points in the restoration processing will be described. FIG. 16 is a diagram depicting a standby computer recovery processing flow in the fourth embodiment; When the restoration data receiving unit 306 receives the data, if the contents are the contents of the memory file, the same processing as in the first embodiment is performed (S16).
002).

If the received data is log data, the received contents are stored in the recovery log data (S16003). After that, the stored file reading unit 308 reads data from the auxiliary storage device 104 according to the methods of the first and second embodiments (S16004). After that, the recovery process ends without performing the process of reflecting the recovery log data 307 in the memory file 302.

Next, the processing of the standby computer during the data write processing will be described. FIG. 17 is a diagram showing a write processing data flow in the fourth embodiment. FIG.
FIG. 24 is a diagram illustrating a write processing flow of the standby computer according to the fourth embodiment.

When a write request arrives from the active computer to the standby computer, in the first and second embodiments, the memory file 302 is immediately updated and the log data is stored. On the other hand, in the present embodiment, before performing the update process on the memory file 302, the data update unit refers to the log data for recovery created by the previous recovery process flow from the first log entry, The log entry of the recovery log data overlapping the position is searched for, and the reflection of the log entry is performed simultaneously with the current write processing.

The data updating unit 301 retrieves the recovery log data one by one from the new log entry toward the old log entry (S18002), and checks whether the log data overlaps with the position and size of the write request from the active computer. (S18003).

When a duplicate log entry is found, the log entries up to that entry are updated in the memory file 302 in order from the oldest log entry (S18004), and the log entry updated from the recovery log data is updated. Remove.

After that, the write request from the active computer is reflected in the memory file (S18005) and stored in the log area (S18006).

As described above, by omitting the process of reflecting the log data at the time of state transition, it becomes possible to transition the operation state of the standby computer at a higher speed.

Embodiment 5 In the present embodiment, when the log of the writing process to the file on the active computer is stored, the rewritten contents are not recorded in the log, but are stored in the same location.
A description will be given of a high-speed transition method in which the size of the log is reduced by having a mechanism for storing the rewritten content in the log for the first time when the second write processing is performed.

This embodiment is different from the first embodiment in the writing process. FIG. 19 shows a write processing data flow in the fifth embodiment.

When the data update unit 202 receives the write request, it notifies the log acquisition unit 204 of its position and size.

The log acquisition unit 204 examines the entries on the log data 205 in order from the newest log entry to the oldest log entry, and checks whether there is any overlap with the notified position and size.

If there is an overlap, data is read from the memory file based on the position and size described in the log entry, and is stored in the log data as update data.

Finally, only the position and size are recorded as entries in the log data.

Thereafter, the memory file is updated to the contents of the write request.

A specific example of the change of the memory file and the log file according to the present embodiment will be described. FIG. 20 is a diagram showing a specific example of changes in a memory file and a log file according to the fifth embodiment.

There are a memory file 203 whose contents at position 512 are ABC, and log data 205 for which no log entry has been recorded yet.

When the write request (1) is received, since there is no log entry for the log data, only the position and size of the write request are recorded as entries in the log data.

Thereafter, the contents of the memory file are changed to zzz as requested by the write.

Subsequently, when a write request (2) is received,
Since there is a log entry overlapping the write request (2) in the log data, the contents of the position 512 and size 3 are stored in the log data from the memory file, and the position and size of the write request (2) are recorded in the log data. , The memory file is changed to yyy according to the write request.

According to this process, the log size can be reduced because it is not necessary to record the last rewritten content in the log while retaining the same content as in the first embodiment.

Embodiment 6 FIG. In the present embodiment, the size of the log file on the active computer is reduced by providing a mechanism for changing a plurality of log entries into one log entry based on the write position and its size. A high-speed transition method characterized by the following will be described.

The log according to the first embodiment is stored in the log data every time a write is made to the memory file. Therefore, when a write process is performed a plurality of times at the same location, a plurality of log entries for the same location are stored in the log data. Will be saved in the data.

In this embodiment, the log size is reduced by leaving only the latest log entry and deleting the old log entry.

For this purpose, the following processing is performed. Note that all the log entries included in the log data are processed in order from the newest log entry to the oldest log entry. FIG. 21 is a diagram illustrating a flow of a log size reduction process according to the sixth embodiment.

Checking whether the address of the log entry already reflected is duplicated or whether there is a duplicated part is as follows.
This is performed by constructing the list structure while sequentially comparing the completed entries with a double linked list structure.

FIG. 22 is a conceptual diagram of the link list processing according to the sixth embodiment. For example, as a specific example, when there is a link list of completed entries as described below, a process for processing a log entry having a position 1536 and a size of 1024 will be described.

Position 153 from top of double linked list
6 is searched until there is an entry having a position exceeding the position 1536. as a result,
It can be seen that the middle entry is location 1024, size 1024, and includes location 1536.

Next, the end position is obtained by adding the size to the position (1536 + 1024 = 2560). In order to search for an entry including the end position, the process is repeated until an entry having a position exceeding the end position appears. In this case, since the next entry is 2560 or more, no entry is included.

Next, a position not yet written in the memory file is determined between the entry including the position 1536 and the end position 2560. Then, it is found that it is necessary to add 512 bytes from 2048 (the end position of the middle entry) to the end position 2560 to the change data.

Thereafter, the selected entry and the entry being processed are combined into one, and a process of linking to the corresponding position is performed.

Here, position 1024, size 1024
By combining the entry with the position 1536 and the size 1024 into one entry, the entry becomes the position 1024 and the size 1536, and the number of log data entries is reduced by one.

By appropriately performing this processing at a timing at which data can be compressed, the size of the log can be compressed, and the amount of data transferred to the standby computer can be reduced.

Embodiment 7 FIG. In the present embodiment, the standby system is shifted to the operation state without reading data from the auxiliary storage device, and when the file access process actually occurs,
A high-speed transition method having a mechanism for reading data into a memory file for the first time will be described.

This embodiment is different from the second embodiment in that
The recovery processing method and the write processing performed when the standby computer transitions from the stopped state to the normal (standby) state are different. FIG. 23 is a diagram showing a standby computer recovery data flow in the seventh embodiment. FIG. 24 shows Embodiment 7
FIG. 7 is a diagram showing a standby computer recovery processing flow in the embodiment.

In the recovery processing, when the data from the active computer is stored in the recovery log data, the processing ends.

Next, the write processing of the standby computer will be described. FIG. 25 is a diagram showing a standby computer write processing data flow in the seventh embodiment.

When a write request arrives from the active computer, the data update unit 301 checks whether data has been read into the memory file 302 by referring to the memory file read flag 310.

After that, when the memory file has not been read, the storage file reading unit 308 is requested to read the memory file.

The memory file reading unit 308 reads data from the storage file 501 and writes the data into the memory file. Then, memory file read flag 3
10 is set to reading completed.

Subsequently, the contents of the memory file 302 read from the storage file are updated with reference to the recovery log data. Subsequent processing is the same as in the second embodiment.

With the configuration described above, the state transition can be performed at high speed by omitting the process of reflecting the saved file to the memory file at the time of the state transition.

Embodiment 8 FIG. In this embodiment, when the log size becomes larger than the designated size, a new log collection is stopped, and a mechanism for changing to a method of saving only the changed position of the memory file is provided. The method will be described.

The present embodiment is a modification of the first embodiment. In the active computer of the first embodiment, the standby computer 1
02 to the log data 205 or the memory file 20
3 is sent, but if the size of the log data 205 becomes large, it is judged that it is better to send the contents of the memory file 203.
In the present embodiment, the log acquisition unit 204 of the active computer 101 has a function of canceling the log collection itself when the size exceeds a certain size.

With the above configuration, the write processing of the execution system computer can be performed at high speed.

[0119]

As described above, according to the present invention, when the log data is small, the log data is reflected, and when the log data is large, the file itself is copied, so that the same memory file can be copied quickly. Therefore, the standby computer can transition the operation state at high speed.

Further, according to the present invention, since the data size to be read from the storage file to the memory file is reduced, the memory file can be identified at a higher speed, and as a result, the operation state of the standby computer can be changed at a high speed. be able to.

Further, according to the present invention, the log entry is not double reflected in the memory file, and the operation state of the standby computer can be changed more quickly.

Further, according to the present invention, it is possible to transition the operation state of the standby computer at a higher speed by omitting the log data reflection processing at the time of the state transition.

Further, according to the present invention, it is not necessary to record the last rewritten content in the log, so that the log size can be reduced.

Further, according to the present invention, the size of the log can be compressed, and the amount of data transferred to the standby computer can be reduced.

Further, according to the present invention, a state transition can be performed at a high speed by omitting a process of reflecting a saved file to a memory file at the time of a state transition.

Further, according to the present invention, the write processing of the execution system computer can be performed at high speed.

[Brief description of the drawings]

FIG. 1 is a diagram showing a state transition from the occurrence of an abnormality in a single computer to the subsequent transition to a normal state in both computers in a dual computer system.

FIG. 2 is a diagram showing a data flow in a data write process of an active computer and a standby computer when a dual system is in a state 1;

FIG. 3 is a diagram showing a data flow in writing data of an active computer and a standby computer when a dual system is in a state 2;

FIG. 4 is a diagram showing a write processing flow of an execution computer.

FIG. 5 is a diagram showing a memory file storage processing flow.

FIG. 6 is a diagram showing a write processing flow of a standby computer.

FIG. 7 is a diagram showing a data flow in a recovery process required for a standby system to transition from a stopped state to a normal (standby) state.

FIG. 8 is a diagram showing a processing flow of an executing computer when data is restored.

FIG. 9 is a diagram showing a recovery processing flow in the standby computer.

FIG. 10 is a diagram showing a recovery data flow of a standby computer in the second embodiment.

FIG. 11 is a diagram illustrating a flow of a recovery process in a standby computer according to the second embodiment;

FIG. 12 is a diagram illustrating an example of obtaining data that requires reading a storage file from update log data.

FIG. 13 is a diagram showing a log entry reflection processing flow in the first and second embodiments.

FIG. 14 is a diagram showing a log entry reflection processing flow according to the third embodiment.

FIG. 15 is a conceptual diagram of a link list process according to the third embodiment.

FIG. 16 is a diagram showing a standby computer recovery processing flow in the fourth embodiment.

FIG. 17 is a diagram showing a write processing data flow in the fourth embodiment.

FIG. 18 is a diagram showing a write processing flow of the standby computer in the fourth embodiment.

FIG. 19 is a diagram showing a write processing data flow in the fifth embodiment.

FIG. 20 is a diagram showing a specific example of changes in a memory file and a log file according to the fifth embodiment.

FIG. 21 is a diagram illustrating a flow of a log size reduction process according to the sixth embodiment.

FIG. 22 is a conceptual diagram of a link list process according to the sixth embodiment.

FIG. 23 is a diagram showing a standby computer recovery data flow in the seventh embodiment.

FIG. 24 is a diagram showing a standby computer recovery processing flow in the seventh embodiment.

FIG. 25 is a diagram showing a standby computer write processing data flow in the seventh embodiment.

[Explanation of symbols]

101 active computer, 102 standby computer, 10
3, 104 auxiliary storage device, 201 application, 202, 301 data update unit, 203, 302
Memory file, 204, 303 log acquisition unit, 20
5, 304 log data, 206, 305 memory file storage unit, 207 recovery data transmission unit, 306 recovery data reception unit, 307 recovery log data, 308 storage file reading unit, 309 memory file recovery unit, 310 memory file reading flag, 401,
501 Save file.

Claims (8)

[Claims] 1. A dual computer system comprising an active computer and a standby computer connected to each other and storing synchronized memory files, wherein the active computer stores a memory file used by the active computer. Executing computer-side memory file storage unit that stores the data, a data update unit that updates the memory file of the executing computer-side memory file storage unit, a log acquisition unit that acquires a log of the memory file update, and stores the acquired log. When recovering the memory file of the standby computer, the log file stored in the log file storage unit or the log file stored in the memory file storage unit of the active computer or the log stored in the log storage unit And a recovery data transmission unit for transmitting the selected recovery data, wherein the standby computer is A standby computer side memory file storage unit for storing a memory file to be used, a file storage unit for duplicating a memory file of the standby computer side memory file storage unit, and a storage file storage for storing the copied memory file as a storage file And when recovering the memory file of the standby computer, receiving the recovery data, and when the received recovery data is a memory file, storing the recovery data as a memory file in the memory file storage unit; When the recovery data is a log, a recovery data receiving unit that stores the received log in the recovery log data storage unit, and when the recovery data received by the recovery data receiving unit is a log, the storage file in the storage file storage unit is a memory file. File to be read into the standby computer memory file storage unit A reading unit, and a memory file restoring unit that updates the memory file read by the saved file reading unit into the memory file storage unit on the standby computer side based on a log in the log data storage unit for restoration. Heavy computer system. 2. The storage file reading unit determines a portion of the memory file that is not updated by the memory file recovery unit based on the log, and selects a portion of the storage file in the storage file storage unit that is not updated as the memory file. 2. The dual computer system according to claim 1, wherein the data is read. 3. A log acquisition unit acquires, for a plurality of updates to a memory file by a data update unit, a log entry having a log entry corresponding to each update, the log entry being able to grasp before and after the update timing. The memory file recovery unit, based on the new log entry of the sequential update timing, except for the already updated part,
2. The dual computer system according to claim 1, wherein the memory file is updated. 4. The standby computer further includes a write request receiving unit that receives a write request for a memory file of the standby computer. The memory file recovery unit determines whether the received write request is a log file of the memory file. 2. The dual computer system according to claim 1, wherein the memory file is updated based on the log when the write request is for a portion to be updated based on the log. 5. A log acquisition unit, when a data update unit performs a previous update and a subsequent update to a common part of a memory file, the log acquisition unit deletes the data recorded as the memory file at the time of the previous update. 2. The dual computer system according to claim 1, wherein the information is acquired as a part of a log at a later update. 6. A log obtaining unit obtains a log having a log entry corresponding to each update, the log entry being able to grasp before and after the update timing, for a plurality of updates to the memory file by the data update unit. The executing computer further includes, in the obtained log, a log entry corresponding to the previous update to the common part of the memory file;
And log entries corresponding to later updates,
Claims: A log compression unit for deleting data included in a log entry corresponding to a previous update and recorded in a common portion of a memory file at the time of the previous update from the log entry. Item 2. The dual computer system according to Item 1. 7. The dual storage device according to claim 2, wherein the storage file reading unit selects and reads a portion of the storage file in the storage file storage unit that has not been read by the storage file reading unit before. System computer system. 8. The dual computer system according to claim 1, wherein the log obtaining unit stops obtaining the log when the size of the log to be obtained exceeds a certain size.