JP4476190B2 - Multi-computer system - Google Patents

Description

 The present invention relates to a multi-computer system that synchronizes internal data in a computer during parallel recovery from a system failure.

Control computer systems that require high reliability, such as railway operation management systems, are not only active computers that perform processing, but also multi-systems that have standby computers that take over processing in the event of a failure in the active computer. Often used as a system.
As in Patent Document 1, in a conventional multiplex system, when one computer is stopped due to a failure or the like and then the computer is started, time information and various information are obtained from the other active system computer after the start. By receiving data, parallel recovery was performed to synchronize the active and standby computers.

Japanese Patent Laid-Open No. 6-311112 (pages 3 to 4, FIG. 1)

In the conventional multiplex system of Patent Document 1, when a computer fails and is recovered by restart, synchronization is achieved by receiving more data from the normal operating system via the network. . This amount of data is large, and there is a problem that the load is heavy on the active computer.
Further, in order to achieve reliable synchronization, data transmission is performed after suppressing functions corresponding to data to be transmitted. Maintaining such a function of the active computer for a long time has a problem of adversely affecting the system.

 The present invention has been made to solve the above-described problems, and is capable of quickly recovering to a multi-system configuration without transmitting / receiving data via a network when recovering from a failure. The purpose is to obtain a multi-computer system with high accuracy.

In the multi-system computer system according to the present invention, in the multi-system computer system in which active and standby computers constitute a multi-system via a network, each computer is an operating system for executing an application, and this operating system is A microkernel comprising a microkernel that operates independently and manages a communication management program that communicates with other computers via a network, and a memory having a shared memory area that is managed by the microkernel and accessed by the operating system It is shared memory area, the store data necessary when the operating system of its own computer to recover from a failure, while the operating system is stopped, if the data is updated as required when recovering Receives the updated data from another computer, it is to update the stored data.

As described above, according to the present invention, in a multi-system computer system in which active and standby computers constitute a multi-system via a network, each computer is an operating system that executes an application, and is independent of this operating system. A microkernel that manages a communication management program that communicates with other computers via a network, and a memory that has a shared memory area that is managed by the microkernel and that is accessed by the operating system. In the shared memory area, save the data required when the computer's operating system recovers from a failure, and if the data required for recovery is updated while the operating system is stopped, calculator It receives et updated data, since updating the stored data, during recovery from a failure, by using the data of the shared memory area, without the transmission and reception of data via the network, fast multiplexing system Can recover to configuration.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a multi-computer system according to Embodiment 1 of the present invention.
In FIG. 1, the multi-computer system is a dual system composed of two control computers. However, the control computer may be composed of three or more.
In FIG. 1, control computers 10 and 11 operate as an active computer and a standby computer, respectively. Each of the control computers 10 and 11 includes a network card 50, a main memory 20, a central processing unit (hereinafter referred to as CPU) 60, and a DIO (Digital I / O) card 110, which are connected by a bus. The In addition, a hard disk device, an input / output device, or the like may be connected.
The network card 50 is connected to an Ethernet (registered trademark) network 90, and the Ethernet (registered trademark) network 90 is also connected to other computers. Via this network 90, the control computer 10 and the control computer 11 communicate, and also communicate with other computers. The computers of both systems are connected using the DIO contact 120.

When both the control computer 10 and the control computer 11 are in a normal state, an OS (operating system) 80, a microkernel 40, an application 70, and communication management are stored in the main memory 20 of the control computer 10 that is an active system. An application 100 that operates under the microkernel 40 such as a program is loaded.
Similarly, the OS 80, the microkernel 40, the application 70, and the application 100 of the microkernel 40 are loaded into the main memory 20 of the control computer 11 that is a standby system. These programs are executed by both the control computers 10 and 11. In the control computer 11, the application 70 may not be executed.
The application 70 is a program that performs processing that is the purpose of the corresponding multisystem. The main memory 20 stores data 30 necessary for recovery by the microkernel 40.

Next, the operation will be described.
The microkernel 40 is independent from the OS kernel, and is positioned below the OS 80. The microkernel 40 monitors the operating status of the OS 80 and manages programs that require real-time performance, such as a communication management program. The processing time of the CPU 60 is preferentially assigned to the microkernel 40, and the remaining time is assigned to the OS 80.
The main memory 20 is allocated to an area managed and used by the microkernel 40 and an area managed and used by the OS 80. When there are other devices such as PCI (Peripheral Component Interconnect), the microkernel 40 assigns whether to manage by the microkernel 40 or the OS 80. The DIO card 110 is managed by the microkernel 40 and used for mutual communication with the microkernel 40 of another computer. The microkernel 40 also manages the network card 50.
The communication management program is a program that transmits data necessary for the application 70 to the OS 80 among the received data received from other computers via the network 90. For this communication, bus communication or virtual Ethernet (registered trademark) communication is used.

An area managed by the microkernel 40 of the main memory 20 can be accessed from the OS 80. That is, it is a shared memory area shared by the microkernel 40 and the OS 80. This shared memory area is recognized by the OS 80 in the same manner as a normal RAM disk and can be directly accessed. Since this shared memory area is managed by the microkernel 40, the contents are not lost even if the OS 80 is stopped while the microkernel 40 is running.
In this shared memory area, various data 30 necessary for recovery, such as those received from other computers at the time of recovery in the conventional system, are stored. Since the microkernel 40 manages the communication management program as described above, the microkernel 40 can receive messages from other computers and other devices even when the OS 80 is stopped. When the state changes while the OS 80 is stopped and the data necessary for recovery is updated, the microkernel 40 receives the data through the network 90 and updates the data 30.

Next, a recovery operation from a failure in the multi-computer system according to the first embodiment will be described.
The control computer 10 and the control computer 11 are connected through the DIO card 110, and the microkernel 40 monitors other computers by communicating with the microkernel 40 of another computer via the DIO card 110. That is, the microkernel 40 constantly monitors the operation state of each other's computers via the DIO card 110. When the control computer 11 determines that an abnormality has occurred in the control computer 10 by this mutual communication, it makes itself an active computer.
Further, the microkernel 40 of each of the control computers 10 and 11 requests the OS 80 to send a survival message at regular intervals. When an abnormality occurs in the OS 80 and the survival message cannot be received, the microkernel 40 detects it and restarts the OS 80.
The microkernel 40 holds data 30 necessary for recovery in a shared memory area, and uses this data 30 to recover to a multi-system configuration.
Using the network card 50 managed by the microkernel 40, while the OS 80 is stopped, various data are received from the microkernel 40 and other devices of the active control computer 11 and this data is updated and used. In some cases, a multi-system configuration may be restored.

  According to the first embodiment, in this way, a memory area managed by the microkernel is provided separately from the OS, the OS can access this area, and data necessary for recovery from the failure is stored in this area. By saving and using this data during failure recovery, there is an effect of shortening the recovery time from an abnormal stop.

Embodiment 2. FIG.
In the first embodiment, the recovery time can be shortened by storing the data 30 necessary for recovery in the shared memory area managed by the microkernel 40. In the second embodiment, however, the shared memory area further includes The operation status of the own computer is saved.
FIG. 2 is a block diagram showing a multi-computer system according to the second embodiment of the present invention.
In FIG. 2, 10, 11, 20, 30, 40, 50, 60, 70, 80, 90, 100 are the same as those in FIG. In FIG. 2, the control computers 10 and 11 do not have the DIO card 110. That is, as in the first embodiment, mutual monitoring of computers is not performed using DIO contacts.

Next, the operation will be described.
The microkernel 40 of the control computer 10 in FIG. 2 transmits an operation status confirmation message to the application 70 managed by the OS 80 at regular intervals. In the application 70 that has received the message, the microkernel 40 manages information on the operation status of whether the self-control computer is operating as an active system, a standby system, or a recovery operation. Is stored in a shared memory area with the OS 80 to perform
On the other hand, the microkernel 40 of the control computer 11 similarly transmits an operation status confirmation message to the application 70 managed by the OS 80, obtains the operation status of the own control computer, and stores it in the shared memory area.
If there is no access to this information for a certain time or more, the microkernel 40 determines that the OS 80 of the self-control computer is stopped, and changes this information to a stopped state.

The microkernels 40 of the control computers 10 and 11 can perform monitoring of other systems by mutually communicating information on the operation status through the network 90 and monitoring each other.
In the first embodiment, monitoring is performed using DIO contact information, but in the second embodiment, by using the present invention, it is possible to monitor other systems at a lower cost than in the first embodiment. become.

  According to the second embodiment, the microkernel of each control computer confirms the operation status with respect to the application managed by the OS, and indicates whether the self-control computer is an active system, a standby system, or being recovered. Other systems can be monitored by storing information in a shared memory area and exchanging this information with each other.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram showing a multiplex system computer system according to Embodiment 1 of the present invention; It is a block diagram which shows the multiplex system computer system by Embodiment 2 of this invention.

Explanation of symbols

10 control computer A system, 11 control computer B system, 20 main memory,
30 data required for recovery, 40 microkernels,
50 network card, 60 CPU, 70 OS application,
80 OS, 90 network,
100 applications on the microkernel, 110 DIO cards,
120 DIO contacts.

Claims (3)

In a multi-system computer system in which active and standby computers constitute a multi-system via a network, each of the computers operates an operating system that executes an application, operates independently of the operating system, and passes through the network. A microkernel that manages a communication management program that communicates with other computers, and a memory that is managed by the microkernel and that has a shared memory area that is accessed by the operating system. The microkernel is stored in the shared memory area. , saves the necessary data when the operating system of its own computer to recover from a failure, while the operating system is stopped, when the data is updated as required when the recovery is said other computer Multiple system computer system receives Luo the updated data, and updates the data described above stored.   2. The computer according to claim 1, wherein each of the computers includes a DIO card, and the microkernel monitors the other computer by intercommunication with a microkernel of another computer via the DIO card. Multi-computer system.   In the shared memory, the operation status of the own computer is stored by the application according to the confirmation of the operation status from the microkernel at a fixed interval. 2. The multi-computer system according to claim 1, wherein said other computer is monitored by mutually communicating the operation status of the own computer in the shared memory area with the kernel.

Images