JPH06202897A - Method and device for resume in network system - Google Patents

Abstract

PURPOSE:To improve the reliability of computer systems connected to the net work system. CONSTITUTION:The resume data of the main memory of an existent computer system are always stored in a resume data storage section 22 of a resume device 20. When the generation of fault in any existent computer system is detected by a fault detection section 21, the data on existent computer system where the fault is generated are read out from the resume data storage section 22 and sent to the standby computer system by a resume processing section 24. The standby computer system continues the processing of the existent computer system where the fault is generated based on the transmitted resume data.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a resume device and method in a network system.

[0002]

[Prior Art and its Problems] Conventionally, in a computer system connected to a network system,
The following are known to improve the reliability of computer systems.

One of them is to make a computer system perform the same processing in a multiplexed manner and take a majority decision of these processing results. However, according to this method, it is necessary to provide the computer system with a special configuration capable of performing multiplexed processing, which causes a problem of high cost.

The other one is to prepare a standby computer system in a standby state in preparation for a failure in the existing computer system. However, the standby computer system is only prepared in advance, and it is not possible to dynamically take over the processing of the faulty existing computer system in a short time. Further, there is a problem that the process to be taken over is limited to a process that is completed each time such as a transaction process.

Further, there is a computer system having a resume function when the power of the existing computer system is cut off. However, this computer system has a problem that it cannot cope with other failures only when the power is turned off.

[0006]

DISCLOSURE OF THE INVENTION A first aspect of the present invention aims to improve the reliability of a computer system connected to a network system.

A resume device in a network system according to a first aspect of the present invention is a device in a network system including a plurality of currently active computer systems and at least one standby computer system. Storage means for storing resume data constantly updated based on data provided from the predetermined computer system, necessary for taking over the processing performed by the predetermined computer system in the system; A failure of the computer system is detected, and when the failure is detected, a failure detection means for outputting a failure occurrence signal, and in response to the failure occurrence signal, the resume data stored in the storage means is transferred to the standby computer. .Means for transferring to system It is provided.

A resume method in a network system according to a first aspect of the present invention is a method in a network system including a plurality of current active computer systems and at least one standby computer system. The resume data, which is constantly updated based on the data provided from the predetermined computer system, necessary for taking over the processing performed by the predetermined computer system in the system, is stored in the storage means, and the predetermined data is stored in the storage means. When a failure of a computer system is detected and the failure is detected,
Outputs a fault occurrence signal and responds to the above fault occurrence signal by
The resume data stored in the storage means is transferred to the standby computer system.

According to the first aspect of the present invention, the resume data is stored in the storage means in preparation for the occurrence of a failure in a predetermined existing computer system. When a failure occurs, by transferring the resume data to the standby computer system, the standby computer system can continue the processing of a predetermined existing computer system.

Therefore, since a special computer system is not required by connecting the resume device to the existing computer system, the cost is reduced. Moreover, the processing of the existing computer system can be dynamically taken over in a short time. The process to be taken over is not limited to the process that is completed each time such as transaction process.

A second aspect of the invention is to improve the reliability of a computer system connected to a network system as in the first aspect.

According to a second aspect of the present invention, there is provided a resume method in a network system, wherein in a plurality of existing computer systems, at least one standby computer system, and the above existing computer system.
A file server that stores the context file of the system is provided so as to be communicable with each existing computer system, and the context file necessary for continuing the processing performed by each existing computer system is always described above. It is stored in a file server, and it is checked in the above file server whether or not any of the above existing computer systems has failed, and any one of the above current computer systems has failed. When a failure is detected, the context file of the computer system in which the failure has occurred is assigned to the standby computer system, and the context file of the current active computer system in which the failure has not occurred corresponds to the current owned computer system. Each transmitted to the system The standby computer system takes over the processing of the computer system in which the failure has occurred based on the received context file, and the computer system in which the failure has not occurred does The processing is restarted from the time of the file.

According to the second aspect of the present invention, the context file is stored in the file server in preparation for the occurrence of a failure in each existing computer system. When a failure occurs in any of the current owned computer systems, the context file of the failed current owned computer system is transferred to the standby computer system, so that the failure of the current owned computer system occurs. The processing is handed over to the standby computer system.

Therefore, the computer system does not need a special computer system, reliability can be improved by using an existing computer system, and a fault-tolerant network can be constructed. Since the context file of the existing computer system is stored in the file server,
Since the existing computer system can inherit the context file any number of times from the processing when the context file was stored in the file server, it can be used for network system analysis.

[0015]

[Explanation of Examples]

First Embodiment FIG. 1 shows the configuration of a network system according to an embodiment (first embodiment) of the first invention.

The network system includes a plurality of computer systems and a resume device 20 on the transmission line 10.
Are connected. In the preferred embodiment, four computer systems 11-14 are provided.

Of the computer systems 11-14, the computer systems 11-13 are currently active systems, and the computer system 14 is a standby system. Resume device
20 is an existing computer system that has a particularly important role in network systems and requires high reliability.
Connected to system 11. The resume device 20 can be connected not only to the computer system 11 but also to other existing computer systems 12 and 13. Also, 1
Resume device 20 with multiple existing computers
It can also be shared by the system.

The existing computer systems 11 to 13 are
Normal processing that is predetermined for each of them or that is instructed through the transmission path 10 is performed, and communication is also performed through the transmission path 10 as necessary. A highly reliable communication protocol, for example, tcp / ip is used for communication of the computer systems 11-14.

The standby computer system 14 is a substitute for the computer system 11 of the existing system to which the resume device 20 is connected, when a failure occurs. Therefore, the standby computer system 14 is in a standby state until a failure occurs in the existing computer system 11. When a failure occurs in the existing computer system 11, the standby computer system 14 takes over the processing of the computer system 11 that has been performed until then.

FIG. 2 is a block diagram showing the electrical construction of the resume device 20.

The resume device 20 includes a failure detection unit 21, a resume data storage unit 22, and a resume data communication unit 2.
2, and the resume processing unit 24.

The failure detection unit 21 is a computer of the existing system.
A fault in the existing computer system 11 is detected based on the alive message sent from the system 11. The alive message is a signal indicating the operating state of the existing computer system 11. When the failure detection unit 21 detects a failure in the existing computer system 11, it gives a failure detection signal to the resume processing unit 24.

The storage unit 22 stores the resume data of the existing computer system 11. In this embodiment, the resume data means all data stored in the main memory of the existing computer system 11 in which the standby computer system 14 should take over the processing. This allows the standby computer system 14 to continue the processing performed by the active computer system 11 as it is when a failure occurs in the active computer system 11.

The resume data needs to be updated when the data in the main memory of the existing computer system 11 changes. When the computer system 11 is started up, all the data in its main memory is transferred to the resume data storage unit 22 and stored therein. After that, only the changed data is changed from the computer system 11 to the storage unit 22 due to the change of the data in the main memory.
And is updated so that the data in the storage unit 22 is always the same as the data in the main memory. For example, the computer system 11 transfers the data in the main memory changed by the file access to the resume data storage unit 22 immediately after accessing the file (for example, external memory) in the system. Immediately after communicating with the computer system 12 or 13, the computer system 11 transfers the data in the main memory changed by this communication to the storage unit 22. The state of the main memory may be monitored at regular time intervals and the changed data may be sent to the storage unit 22 at regular time intervals. Data transfer to the right place in the program
The transfer time is shortened by inserting the command and compressing the data to be transferred.

The resume processing section 24 waits for the resume data stored in the storage section 22 based on the failure detection signal input when the failure detection section 21 detects a failure in the existing computer system 11. The resume data communication unit 23 is controlled so as to transfer to the standby computer system 14 in FIG.

The resume data communication unit 23 is a computer system which is in a standby state for the resume data stored in the storage unit 22 in response to a command from the resume processing unit 24 when a failure occurs in the existing computer system 11. To be transmitted via the transmission line 10.

Computer system in standby state
Upon receiving the resume data of the computer system 11 in which the failure has occurred from the resume device 20, the 14 takes over the processing performed by the existing computer system 11 based on the received resume data. The computer system 14 switches the address in the network system of the computer system 14 to the address of the network system owned by the existing computer system 11 until a failure occurs. When the address is switched, the computer system 14 notifies the existing computer systems 12 and 13 that the address has been switched. This allows computer system 14 to communicate with computer systems 12-13, which computer system 11 was doing, on behalf of computer system 11.

Second Embodiment FIG. 3 is a block diagram showing the configuration of a network system according to an embodiment (second embodiment) of the second invention.

The network system comprises a plurality of existing computers connected to each other by a transmission line 30.
Systems 31-33, standby computer system 34,
And a file server 35. The standby computer system 34 is normally in a standby state, and when any of the existing computer systems 31 to 33 fails, the current active computer system in which the failure occurs
Take over on behalf of any of the systems 31-33.

The file server 35 stores files (data in the main memory and, if necessary, external memory) of the existing computer systems 31 to 33. The file data of the existing computer systems 31 to 33 are updated according to the processing of the systems 31 to 33. The file server 35 stores not only the latest data of the files of the systems 31 to 33 but also the data before the update up to the data before the update plural times before.

The file server 35 sends a freeze signal to the existing computer systems 31 to 33 periodically or as needed. Upon receiving the freeze signal, the existing computer systems 31 to 34 transfer the file data of the computer system to the file server 35 as a context file. The file server 35 saves the context files transferred from the existing computer systems 31 to 33.

When the file server 35 detects a failure in any of the existing owned computer systems 31 to 33, the latest context file saved in another existing owned computer system in which no failure has occurred. Return the one corresponding to (the one transferred from each computer system). Upon receiving the latest context file, the existing computer system returns to the time when the received context file was transferred and restarts the processing based on the received context file. The context file of the faulty existing computer system is transferred to the standby computer system 34. Based on the received context file, the standby computer system 34 takes over the processing of the faulty current computer system from the time when the current computer computer transferred the context file.

The entire network system returns to the state at the time when the file server 35 sends the latest freeze signal.

4 and 5 are flow charts showing the processing procedure of the file server 35. FIG. 6 is a flow chart showing the processing procedure of the computer system, and FIG. 7 shows the details of a part of the processing (processing of step 61).

The existing computer systems 31 to 33 are processing their respective applications (FIG. 6, step 61, FIG. 7, step 71).

The file server 35 sends a freeze signal to the currently owned computer systems 31-33 (FIG. 4, step 41), and each currently owned computer system 31-33 receives the freeze signal from the file server. Then, (FIG. 7, step 72), the application processing that has been performed until then is stopped (FIG. 7, step 773). Each existing computer system 31-33 sends a freeze start signal to the file server.
35 (Fig. 7, step 74), freezes the kernel (Fig. 7, step 75), and performs the processing that was being performed in case of failure in another standby computer system (implementation). In the example, a context file is created for the standby computer system 34) to continue (FIG. 7, step 76).

The file server 35 updates the freezing start signal from each of the existing computer systems 31 to 33 within a predetermined time after transmitting the freeze signal (step 42 in FIG. 4), and then receives the latest context file. In preparation for saving, the file history processing is performed to save the already saved context file as a past one (step 43 in FIG. 4). When the file server 35 finishes the file history processing, it sends a history processing end signal to each of the existing computer systems 31 to 33 (FIG. 4, step).
44).

Each existing computer system 31-33
When the history processing end signal is received from the file server 35 (FIG. 7, step 77), the context file generated in step 76 is transferred to the file server 35 (FIG. 7, step 78). Each existing computer system 31
After completing the transfer of the context file, each of ~ 33 sends a transfer end signal to the file server 35 (Fig. 7, step 79). Each of the existing computer systems 31 to 33 returns to step 71 and moves to application processing.

The file server 35 stores the context file transferred from each of the existing computer systems 31 to 33 (FIG. 4, step 45). When the file server 35 receives the transfer end signal from the existing computer systems 31 to 33 within a predetermined time after transmitting the freeze end signal (step 41) (FIG. 4, step 46), it returns to step 41 and freeze signal again. Is output. Needless to say, the freeze signal may be output after a certain waiting time after receiving the transfer end signal.

In step 42 of FIG. 4, if the file server 35 does not receive the freeze start signal from the current owned computer system within the predetermined time, the file server 35 normally operates for the current owned computer system which has not received the signal. An inquiry is made as to whether or not it is operating, and the response to this inquiry determines whether or not a failure has occurred in the computer system (step 47 in FIG. 4). When it is confirmed that the present owned computer system inquired by the file server 35 is operating normally, it waits for a freeze start signal transmitted from the present owned computer system (FIG. 4, step 48).

In step 46 of FIG. 4, if the file server 35 does not receive the transfer end signal from the current owned computer system within the predetermined time, it normally operates for the current owned computer system which has not received the signal. An inquiry is made as to whether or not it is operating, and the response to this inquiry determines whether or not there is a failure in the computer system (step 49 in FIG. 4). When it is confirmed that the current owned computer system inquired by the file server 35 is operating normally, it waits for a freeze start signal transmitted from this owned computer system (FIG. 4, step 50).

The file server 35 proceeds to step 47 described above.
Alternatively, it is assumed that, at 49, it is detected that a failure has occurred in, for example, the current owned computer system 31 in any of the currently owned computer systems. The file server 35 sends a fault occurrence signal indicating that a fault has occurred in the currently owned computer system 31 to the currently owned computer systems 32-34 (FIG. 5, step 51).

Upon receipt of the fault occurrence signal (step 72 in FIG. 7), the existing computer systems 32 and 33 stop the normal processing performed by the computer system (FIG. 6, step 63) and start decompression. The signal is transmitted to the file server 35 (FIG. 6, step 64). When the standby computer system 34 receives the failure occurrence signal, it sends a decompression start signal to the file server 35.

When the file server 35 receives the decompression start signal from each of the existing computer systems 32 and 33 (step 52 in FIG. 5), it stores the latest context information of these computer systems 32 and 33. The files are returned to each (step 53 in FIG. 5,). When a decompression start signal is received from the standby computer system 34, the faulty existing computer system 31
The latest context file of the above is transmitted to the standby computer system 34.

Existing computer system 32 and
When receiving the context file of its own from the file server 35, the 33 performs a process (decompression of the context file) of returning to the state at the time when the context file was generated (FIG. 6, step 65). The existing computer systems 32 and 33 send a decompression end signal to the file server (FIG. 6, step 66).

When the standby computer system 34 receives the context file of the current owned computer system 31 in which the failure has occurred, the standby computer system 34 starts the current owned computer from the time when the context file is generated based on this context file. The processing for taking over the processing of the system 31 (decompression of the context file) is performed (FIG. 6, step 65). At this time, the standby computer system 34 switches the address in the network system to the address of the existing computer system 31. The standby computer system 34 sends a decompression end signal to the file server (FIG. 6, step 66).

When the file server 35 receives the decompression end signal from each computer system 32-34 (FIG. 5, step 54), it sends a start signal to the computer system 32-34 (FIG. 5, step 55).

Existing computer systems 32 and 33
Receives the start signal from the file server 35 (step 67 in FIG. 6,), restarts the processing.

When the standby computer system 34 receives the start signal from the file server 35 (step 67 in FIG. 6), the standby computer system 34 takes over the process performed by the faulty existing computer system 31.

The network system will start processing from the time when the file server 35 generates the latest context file.

[Brief description of drawings]

FIG. 1 shows a network system according to a first embodiment.

FIG. 2 is a block diagram showing an electrical configuration of a resume device.

FIG. 3 shows a network system of a second embodiment.

FIG. 4 is a flow chart showing a processing procedure of a file server.

FIG. 5 is a flow chart showing a processing procedure of a file server.

FIG. 6 is a flow chart showing a processing procedure of a computer system.

FIG. 7 is a flow chart showing a processing procedure of a computer system.

[Explanation of symbols]

10, 30 Transmission line 11, 12, 13, 31, 32, 33 Current computer system 14, 34 Standby computer system 20 Resume device 21 Failure detection unit 22 Resume data storage unit 23 Resume data communication unit 24 Resume Processor 35 File server

Claims (3)

[Claims] 1. A device in a network system comprising a plurality of existing computer systems and at least one standby computer system,
Storage means for storing resume data which is constantly updated based on the data given from the predetermined computer system necessary for taking over the processing performed by the predetermined computer system among the existing computer systems Detecting a fault of the predetermined computer system, and detecting a fault, outputting a fault occurrence signal, fault response means, and responsive to the fault occurrence signal, the resume data stored in the storage means. A resume device in a network system comprising means for transferring to the standby computer system. 2. A method in a network system comprising a plurality of existing computer systems and at least one standby computer system,
Storing resume data, which is constantly updated based on the data given from the predetermined computer system, required to take over the processing performed by the predetermined computer system among the existing computer systems, in the storage means. Then, when a failure of the predetermined computer system is detected and a failure is detected, a failure occurrence signal is output, and in response to the failure occurrence signal, the resume data stored in the storage means is transferred to the standby system. Resume method in network system to transfer to computer system. 3. A plurality of currently existing computer systems, wherein at least one standby computer system and a file server for storing a context file of the current existing computer system are provided in each of the currently owned computer systems. A context file necessary for continuing the processing performed by each of the existing active computer systems is stored in the file server at all times so as to be communicable with the system. The file server checks whether or not a failure has occurred, and when it is detected that a failure has occurred in any of the currently existing computer systems, the context file of the failed computer system is detected. The above standby computer system Each of the context files of the existing owned computer system in which no failure has occurred is transmitted to the corresponding existing owned computer system, and the standby computer system receives the above-mentioned failure based on the received context file. A resume method in a network system in which the computer system that has taken over takes over, and the computer system that has not had the above-mentioned fault restarts the process from the time of the context file based on the received context file.